beta_solutions-v2

3 Habits To Communicate Better - In Slack, Teams and Email

Thu, 28 Aug 2025 03:31:27 GMT

A practical guide for engineers, project managers, and technical teams.

Poor communication is one of the most common and avoidable reasons projects fail.

The Project Management Institute found that ineffective communication is the main cause of project failure about one-third of the time[1].

It doesn’t matter how skilled your engineers are or how strong your idea is. If you can’t communicate clearly, execution suffers.

At Beta Solutions, we’ve found that a few small habits can dramatically improve project outcomes. These changes are simple to implement and deliver results fast.

While the examples here are for written communication (e.g Slack, Teams, or email), the principles apply to other forms too.

Habit 1: Address Your Messages to a Single Person

It’s tempting to drop a message into a group chat and hope the right person picks it up. The problem is that when responsibility isn’t clear, it’s often ignored. Tagging the whole team does not make something urgent - It makes it no one’s job.

❌ Bad Example

When you send a group message without naming a single person responsible, you create uncertainty. Everyone sees it, but may assume someone else will handle it. In many cases, no one does or it gets picked up far later than it should have been.

Why this is bad:

Creates diffusion of responsibility, also known as the “bystander effect” (in chat form)
Slows decisions and handovers
Leaves the sender wondering if anyone has even seen the message
Can lead to frustration or blame that could have been avoided

Instead of leaving ownership up in the air, you can make your message more actionable by making one person clearly responsible while keeping others in the loop for awareness. Let’s look at how this could have been done.

✅ Good Example

When you tag a single person and clearly state what you need, there is no doubt who owns the task. Others can still be tagged for visibility, but they know they are not expected to act. This same principle applies in email - for example, starting with “Hi Jason (cc: Morten & Jono)” makes it clear Jason is the one being addressed, while noting others have been copied for awareness.

Why this is good:

One clear owner knows they must act
Others tagged understand they are for awareness only
Speeds up response times

Key takeaway: Always hand the “baton” to one person. Others can still be tagged for visibility.

Habit 2: Number multi-part questions

Long, unstructured messages with multiple questions are a recipe for miscommunication. Even with the best intentions, people tend to answer what is easiest or most recent and skip the rest. Numbering questions makes it easy for the recipient to respond to each one, and for you to check nothing has been overlooked.

❌ Bad Example

When multiple questions are buried in one sentence or paragraph, people skim and latch onto the part that stands out to them. This leads to partial answers, confusion about what “yes” or “no” refers to, and more follow-up messages just to clarify.

Why this is bad:

Replies may skip points without anyone noticing
“Yes” or “No” answers become ambiguous
Makes it harder to revisit decisions later in the thread
Creates extra back-and-forth just to clarify what was meant

The fix is simple. Give each question its own number. This creates a structure that is easy to respond to and impossible to misinterpret. Let’s see what that looks like in practice.

✅ Good Example

Numbering your questions forces clarity. It helps the recipient reply in a structured way and makes follow-up changes quick and precise.

Why this is good:

Keeps questions and answers aligned, making it easy to match reply to question
Makes follow-up fast - for example, “3. Can we make that 10:15am” instantly tells the recipient exactly which point to adjust without re-reading the whole conversation.
Improves accuracy when multiple people are involved in a reply
Works especially well for asynchronous communication where hours or days pass between messages

Key takeaway : If you ask multiple questions, make it easy for the other person to give you multiple clear answers.

Habit 3: Respond Even If You Don’t Have the Answer

This might sound counterintuitive, but if you are unable to provide the full answer in a reasonable timeframe, respond anyway.

Silence creates uncertainty. If someone is waiting on you, even a quick acknowledgement shows you have seen their message and are working on it. You do not always need the final answer right away.

❌ Bad Example

People often have the right intention by waiting to reply, but this can create unnecessary delays. Silence to the person asking can feel frustrating and create uncertainty. A quick acknowledgement closes the loop and reassures the sender you’re on it.

✅ Good example

Why it works:

This may not be the answer the sender ideally wants - in a perfect world, you’d just get it done straight away - but that’s not always doable. This sets clear expectations and keeps everyone aligned.

Note: It goes without saying - if you tell someone you’ll get back to them by a certain time, you need to keep your word and follow through.

TLDR:

When sending a message get into the habit of:

Tagging one person - Assigns the batton of responsibility to someone.
Numbering your points - Brings clarity & readability to your message.

When replying to a message, get into the habit of:

Responding promptly, even with partial info - Closes the loop quickly.
Using numbers to structure your reply- Brings clarity & readability to your message.
Check your messages regularly to avoid delays.

Found this post useful?

Subscribe to our newsletter for more practical tips on engineering communication, project execution, and product development. Subscribe here .

References

[1] Project Management Institute. (2017). A Guide to the Project Management Body of Knowledge (PMBOK® Guide) (6th ed.). Newtown Square, PA: Project Management Institute.

How to Effectively Run After-Action Reviews (AARs)

Tue, 20 May 2025 23:15:21 GMT

Like many high-performing organizations, we borrow best practices from industries that operate in demanding environments. Medicine, aerospace, military operations, and Formula One racing all rely heavily on structured reviews to improve their processes and ultimately their outcomes.

Inspired particularly by the U.S. Navy SEALs' method of After Action Reviews (AARs) performed after every mission, we've adopted a similar approach to critically assess our projects. While we didn’t invent the concept, we've embraced its simplicity and effectiveness.

What is an AAR?

An After Action Review (AAR) is a structured meeting where the whole team comes together to review a recently completed project. It involves examining key events and actions during the project to clearly understand what happened, why it happened, and what improvements can be made for next time. Rather than complicating things with unnecessary paperwork or lengthy meetings, we intentionally keep AARs concise, focused, and actionable.

Why Does It Matter?

At Beta Solutions, the value of AARs is deeply embedded in our culture. It's not just another meeting—it's an opportunity to be genuinely transparent. AARs encourage team members to openly reflect on their successes and mistakes. While these conversations can sometimes be uncomfortable, we believe that embracing vulnerability builds trust within the team. This mutual trust, in turn, makes it easier for individuals to honestly own their mistakes and genuinely commit to improvement.

We recognize that continuous improvement doesn't happen by chance. It requires deliberate reflection and an environment built on trust and open dialogue. Our AARs help ensure that every project contributes tangible learnings for future ones.

How To Conduct AARs

While there are various ways to conduct AARs and the structure would likely differ industry to industry, here's how we've structured the process.

Timing

It's important to run AAR sessions promptly after project completion. Immediate reflection ensures the events and details remain fresh in everyone’s minds, making insights more accurate and actionable.

We schedule an AAR meeting on the last Thursday of every month. This consistent scheduling ensures all team members can plan around the meeting, and it matches the frequency and volume of projects we typically handle.

Led by the Project Manager

Our AAR meetings are usually led by the Project Manager (PM), who prepares and presents a concise presentation summarizing the project's critical points. These meeting are attended by everyone, from the CEO to Project Managers, Engineers, Administrators, Sales, and Marketing. Having company-wide attendance ensures the team is aligned about our overall performance and vision.

The Presentation

It's important for this slideshow not to get bogged down in technical details—it should clearly communicate the most important information in the most concise way possible. No fluff! The presentation centres around only 5 key sections:

Project overview: A brief description of what the client wanted, highlighting key objectives and expected outcomes. This section ensures team members, including those not directly involved in the project, clearly understand the project's goals and context.
Key metrics : Project data—in our case this will likely include initial estimations, actual time spent per task, project budgets, and any variations that occurred during the project.
What went well : The PM highlights the successes and effective practices from the project, explaining why certain methods or actions led to positive outcomes.
What didn't go well : Here, the PM outlines the challenges or shortcomings faced during the project, providing context and analysis to help the team understand underlying causes.
New standards : Based on the above observations, the PM suggests new standard operating procedures (SOPs) to ensure these mistakes or poor outcomes are not repeated.

The slideshow aligns the team by clearly summarising essential points, ensuring everyone is informed and ready to engage openly.

Team Participation and Open Dialogue

Importantly, these meetings explicitly avoid blame or personal criticism. However, individuals are encouraged to take personal responsibility for mistakes or poor decisions. This approach ensures accountability, prevents repeated issues, and further strengthens trust among team members.

Open dialogue allows team members to express their thoughts and perspectives freely. This involvement significantly improves the likelihood of team members buying into new SOPs. Identifying new standards is only beneficial if the people responsible for implementing them fully support these changes.

Our AARs last approximately 30 minutes. Keeping meetings short respects everyone's time and maintains high engagement throughout the session. This concise approach leads to clearer, more practical outcomes.

Common Challenges and How To Address Them

Implementing AARs can come with some typical challenges:

Resistance to openness : Initially, team members may hesitate to openly discuss mistakes or poor outcomes. At Beta Solutions, we address this by setting clear expectations that AARs are about improvement, not blame, fostering an environment where vulnerability is viewed positively.
Time Management : Without structure, meetings could become lengthy or unfocused. Our strict 30-minute limit and clear agenda ensure meetings remain concise and productive.
Engagement : Meetings can sometimes feel dull, reducing participation and honest feedback. To counter this, we create a positive atmosphere with coffee and snacks—making it something the team can look forward to.

Key Takeaways for Implementing AARs in Your Team

Schedule AARs promptly after project completion.
Maintain consistent and predictable scheduling.
Keep meetings focused and concise.
Prepare a concise presentation consisting of:
A project overview
Key metrics
What went well
What didn’t go well
New standards
Foster an environment of trust and openness by emphasising accountability without blame.

References:

Banner photo: https://x.com/PFF1/status/502750634689724416/photo/3

Modern Electronic Hardware Design: An overview

Wed, 26 Jun 2024 03:30:19 GMT

Electronic hardware is found in a wide range of products, from consumer gadgets to industrial equipment. The market success of these products often depends on how well their hardware performs. However, designing hardware is inherently complex— involving multiple stages with unique challenges and requirements. At Beta Solutions, with over 16 years of experience, we have developed a process that helps manage these complexities. This blog aims to provide a high-level overview of some of this process, highlighting the first three distinct phases. While developing these products successfully entails hardware, mechanical and firmware development, this blog will focus specifically on electronic hardware design.

The Beta Solutions Hardware Design Flow

The following process has been refined over time and is now used for the majority of our product development projects. However, it's worth mentioning that some projects demand a more “agile” methodology due to their unique needs and constraints. Agile hardware design will be a topic we tackle in a future blog.

While these phases appear linear on paper (known as a Waterfall model), keep in mind that in reality, each stage involves multiple iterations, and some stages may run concurrently.

Phase 0 - Pre-engagement Evaluation

In this initial phase, we hear about our clients' initial requirements, discuss how Beta Solutions can assist with their development, and discuss what the R&D journey might look like. We also address aspects like intellectual property, sales projections, and provide ballpark estimates for the possible investment and timeframes involved.

The project will formally kick off once the prospective client has approved both our Letter of Engagement and our Phase 1 proposal.

Phase 1 - Discovery & Specification

The objective of Phase 1 is to scope the project work, providing a strong foundation for the remainder of the product development process.

Core Outcomes/Tasks:

Understand : Gaining insight is crucial for informed design decisions. Focus areas include:
Customer: Gain insights into the client's needs, objectives, and pain points.
Market: Analyze the industry landscape to identify risks, challenges, and opportunities.
End User: Understand the target demographics and user behaviours to ensure the product meets user expectations and requirements.
Compliance : Evaluate what regulatory compliance standards the project must meet - including electromagnetic compatibility (EMC) requirements.
Risk Analysis : Assess potential project risks and ways to mitigate them.
Hardware Features : Define what hardware features are needed in this product.
Electrical Specifications: Develop reasonably comprehensive specifications that outline the design requirements and goals.
Architecture : Develop a high-level system architecture outlining the interconnections between major features and subsystems.
Select Key Components : Identify and choose specific components that will form the core of the product.
Production Estimates : Provide a ballpark estimate for manufacturing costs. These estimates are based on initial assumptions and are just that, an estimate.

All of the above are defined in a “Project Discovery Report.” This report summarises the findings from Phase 1, confirming that there are sufficient specifications and a solid foundation to advance to the next phase.

When Is Phase 1 Finished?
Phase 1 is completed when the client approves the Project Discovery & Specifications Report. This approval confirms that the project's foundation is solid and that all critical specifications have been identified.

Subsequently, a proposal outlining the remaining work to be undertaken and the estimated investment required will be presented to the client. The move to Phase 2 is initiated once the client approves this new proposal.

Phase 2 - Detailed Design & Testing

The objective of Phase 2 is to meticulously refine and engineer the product into reality. This is where the bulk of the detailed work is undertaken.

Phase 2 of hardware design can be broadly split into three main sections:

Schematic Design.
PCB Placement & Routing.
Prototyping & Testing.

Schematic Design

What is a Schematic?

A Schematic is a detailed graphical representation of the product's electronic circuitry. It serves as a blueprint, showing which components are interconnected to which.

What makes Schematic Design Complex?

Schematic design is inherently complex due to the need for a deep understanding of every component. This complexity is further compounded by the necessity to consider how each component interacts with all the other connected components.

Core Outcomes/Tasks:

NB: While the tasks listed below cover some of the key aspects and outcomes essential to the schematic design phase, they are by no means exhaustive.

Component Selection: Finalize the selection of each component, ensuring they meet the design requirements and constraints. This task involves reviewing datasheets to understand the properties and capabilities of each component. Each component will have tradeoffs to consider including:
Performance.
Size.
Cost /Availability.
Component Interconnection: Connect the selected components in the schematic.
Documentation : A finished schematic will be populated with copious notes from the hardware engineer for several reasons:
Design calculations.
Key design decisions.
Instructions for firmware engineers.
Circuit Simulation: Perform circuit simulations to validate the design before physical prototypes are made. This involves using software tools like LT Spice to simulate the behaviour of the circuit, allowing for the identification and resolution of issues in a virtual environment.
Peer Reviews: Throughout the schematic design process, the design is reviewed multiple times by multiple people, including senior hardware engineers and firmware engineers working on the firmware aspects of the project. Additionally, an architecture review is conducted to ensure that the overall system design aligns with the initial specifications in phase 1.

During this phase, we continue to collaborate with the client to ensure that the design remains on track and aligned with the client's requirements.

PCB Placement & Routing

What is a PCB?
A Printed Circuit Board (PCB) is the physical board that hosts the components and connections illustrated in the schematic.

What is PCB Placement and Routing?

PCB Placement & Routing involves arranging the electronic components on the physical board and creating the conductive pathways (traces) that connect these components.

What Makes PCB Placement & Routing Complex?

PCB Placement & Routing is inherently complex due to multiple factors:

Electromechanical Considerations : Collaborating with mechanical engineers (internal or external) to ensure that the initial layout of key components fits within the product’s physical constraints.
Signal Integrity : Ensuring that signals are transmitted without interference or degradation. This is particularly important for any high-frequency signals.
Thermal Management : Managing heat dissipation to prevent overheating.
Manufacturability : Depending on where the PCBs are intended to be manufactured, they will need to meet specific design rules and constraints set by the manufacturing facility. These rules can include minimum trace widths, spacing, number of layers, and other parameters that ensure the boards can be produced using the equipment and processes available at the facility.
Electromagnetic Compatibility (EMC) Considerations : This involves both minimizing the emissions generated by the circuit and enhancing the design’s immunity to external electromagnetic influences. If you would like to learn more about EMC you can read our article - Electromagnetic Compatibility (EMC) Compliance - Answers to Frequently Asked Questions .

Core Outcomes/Tasks :

Again, we must stress that this is not an exhaustive list, and just touches on some of the tasks and outcomes in PCB Placement
& Routing.

Define PCB Specifications
Dimensions : Identify the optimal size and shape for the board, in collaboration with mechanical engineers, to ensure it fits within the product’s physical constraints.
Aesthetics: Change the colour of the PCB or brand logos can be added purely for looks.
Design Rules : Establish the manufacturing constraints, adhering to the design rules mandated by the production facility.
Stack Up : Decide on the number of copper layers, total board thickness, and the purpose and order of each layer. This will affect factors such as signal integrity, power distribution, and thermal performance.
Component Placement : Strategically place the components on the PCB, considering factors such as electrical performance, thermal management, and mechanical constraints.
Trace Routing : Create pathways (traces) that connect the components as per the schematic design. This involves:
Signal Path Optimization.
Avoiding Interference.
Layer Management.
Add Strategic Test Points : Test points are specific locations on a PCB designated for testing purposes. They are usually small pads or vias that allow engineers to probe the circuit with test equipment to measure electrical signals, verify functionality, and diagnose issues. Our test rigs we design, in phase 4, use these to catch manufacturing defects.
Peer Reviews : This is another time for peer reviews, involving collaboration with both mechanical and senior hardware engineers.

An integral part of Phase 2 is the effective handoff to our firmware team. Although firmware engineers have been involved since the beginning of the project, a thorough brief on the finalized hardware design is essential for seamless integration and development.

Click here to read more about how we do firmware development - https://www.betasolutions.co.nz/an-introduction-to-firmware .

Prototyping & Testing

Once the schematics and PCBs are finalized, we can begin procuring prototypes. Typically, we budget for 2-3 spins, although more may be necessary for complex products.

The objective of Prototyping & Testing is to verify that the hardware design functions as intended and meets all specified requirements.

The first prototype spin provides the firmware team with their first opportunity to test their functionality on actual hardware, which often leads to valuable insights and learnings. This is why prototype runs are so important; despite thorough research and risk mitigation in Phase 1, due to the inherent complexity, it is highly likely that issues will arise during the prototyping phase. This should be expected.

By addressing these issues early through rigorous testing and validation, we ensure that the final product is reliable, functional, and meets all design criteria.

What is Tested & How?

NB: Depending on the project, only some of the following tests may be undertaken.

Visual Inspection : Conduct a visual examination of the prototype to identify any obvious manufacturing defects.
Power Supply Verification : Validate the power supply subsystem to ensure it delivers the correct voltages and currents without undue noise or instability.
Thermals : Assess the thermal performance of the prototype by measuring key components and overall board temperatures under various operating conditions. Boards are run under a thermal camera looking for hot spots.
Circuit Subsystems : Any parts of the circuit that can be tested without firmware will be.
Electrical Extremes : Test all electronics over their full working range.
EMC Pre-compliance Testing :
TEM Cell : Used to measure approximate far-field radiated emissions from the PCB over a wide frequency spectrum by placing the PCB inside it. More convenient than far-field testing setups since it can all fit inside the EMC shielding tent.
Near field probes : Used to measure specific frequencies at locations around the board to identify which areas are contributing to emissions.
Far-field testing: measures the emissions using an antenna at a specific distance from the device e.g. 3 meters. The emissions levels are compared with limits from a compliance standard e.g EN55032.

The firmware development team is responsible for developing and integrating the software that controls the hardware. Firmware development has been running concurrently with hardware design and will likely continue throughout the entire process. Typically, the majority of the hardware's functionality requires firmware to be fully tested. The firmware engineers will work closely with the hardware team to verify and ensure all components function as intended. The detailed process and challenges of firmware development will be covered in more depth in a future blog.

Client Testing

After verifying most of the hardware, the client is expected to conduct thorough testing themselves. This is crucial as they can test the product in its intended use case. This often involves testing the hardware on actual equipment and environments where it will be used, as well as involving beta testers who can provide real-world feedback and identify any potential issues that may not have been evident during initial testing phases. To expedite this process, the product is often provided to the client before the firmware is fully implemented, allowing them to test a partially featured product. This ensures earlier detection of issues and provides ample time to address any long-term concerns.

What Tools Are Used?

Not every tool is necessary for every PCB project, but each has its specific applications and advantages. At Beta Solutions, we utilize a variety of tools including:

Oscilloscope
Multimeter
Logic Analyser
Signal analyser/ Network analyser
Signal Generator
IR Camera

Once all the hardware has been verified we can move into Phase 3 (Regulatory Compliance) or Phase 4 (Production Readiness) depending on the product. This move is generally marked by the formal approval of the customer acceptance testing results defined in phase 1. If all criteria are met, the product is deemed ready for stringent regulatory compliance checks and/or pre-production preparations.

Overview of Phases 3, 4, and 5

These phases will be explored in greater detail in future blog posts, covering specific challenges and best practices.

In Phase 3 - Regulatory Compliance , we ensure the product meets all required standards through rigorous testing and documentation to obtain compliance certificates. At times, this can be a considerable undertaking, often involving multiple iterations and reports to satisfy regulatory requirements.

Transitioning to Phase 4 - Production Readiness involves optimizing the design for large-scale manufacturing, conducting pilot production runs, and finalizing production documentation and pricing. This phase can be a significant investment depending on the setup costs involved. This is sometimes known as Non-recurring engineering (NRE).

Finally, Phase 5 - Volume Production transitions the design into mass manufacturing while focusing on quality assurance, ongoing support, and lifecycle management. For some customers, this might be 50 boards, for others it could be 100,000.

Why Choose Beta Solutions

Designing modern electronic hardware is a complex, multi-faceted endeavour that requires meticulous planning, deep expertise, and rigorous testing. At Beta Solutions, our time-tested process ensures that the complexities at each stage are managed effectively. This approach helps mitigate risks, meet regulatory standards, and ultimately bring robust, quality products to market.

As we have touched on in this blog, this process can be daunting, but with the right partner, it becomes manageable and highly rewarding.

If you're considering outsourcing your R&D to optimize your resources and focus on your core competencies, Beta Solutions are your ideal partners. See how you can get access to a dedicated team of experts here .

An Introduction to Firmware

Sun, 17 Mar 2024 23:15:37 GMT

As a firmware engineer, I often find myself in a classic conundrum when asked about my profession. Responses like, "Oh, so you're a software engineer?" or "How is that different from hardware?" are all too familiar. It's challenging to explain what firmware engineering entails since it resides in a niche space that bridges hardware and software. However, consider this astonishing fact: High-end cars like the Mercedes-Benz S-Class contain over 10 million lines of code , all of which are processed by up to 100 microprocessors networked throughout the car. This immense amount of firmware-based coding powers everything from engine management systems to advanced safety features.

However, the influence of firmware extends far beyond the automotive industry. The vast majority of electronic devices you interact with will have firmware in them, from household appliances to personal gadgets and industrial machinery. This central role in both mundane and complex technologies should give you a sense of how widespread firmware is in our daily lives.

In this article, I aim to clarify some basic questions about what firmware is, its relationship between hardware and software, and how it's developed.

What is Firmware?

When we talk about firmware, we are referring to the code that runs on a microcontroller inside an electronic device. Microcontrollers are compact integrated circuits that include a processor, memory, and input/output peripherals all on a single chip. The firmware programmed onto the microcontroller dictates how the device functions and behaves. Often firmware code is written in the C or C++ programming language, but these days there are a few more viable options.

How does Firmware relate to Hardware?

The microcontroller the firmware runs on is soldered onto a printed circuit board (PCB), that connects it to other hardware components like power supplies, sensors and actuators. The firmware controls and reacts to these components and makes decisions enabling the device to perform specific tasks.

Take an air conditioning unit as an example, the firmware continuously reads data from the hardware, a temperature sensor, to monitor the ambient temperature. When the temperature exceeds a predefined threshold, the firmware instructs the compressor (more hardware) to start. This cools the room. Once the temperature reaches the desired level, the firmware signals the compressor to stop again. This is a very simplistic example. For those seeking a deeper dive into our work, take a look at the project we completed with Rinnai here .

How is Firmware different from Software?

The line between firmware and software can sometimes appear blurry, but some distinctions set them apart. Firmware typically operates at a lower level than traditional software and directly communicates with the hardware it is programmed for. Unlike generic software applications that can run on various hardware platforms, firmware is tailored for specific devices, often referred to as ‘dedicated hardware’.

One crucial aspect that sets firmware apart is its real-time responsiveness. Firmware often needs to react swiftly to events with precision timing requirements, making it critical for time-sensitive applications. In contrast, software is more concerned with high-level decision-making and doesn't prioritise direct hardware interaction.

Complexity in firmware development also arises due to the limited availability of libraries. Unlike software development where programmers can leverage a multitude of libraries and frameworks, firmware engineers often have to write code from scratch, due to the specialised nature of the hardware they are working with.

Moreover, firmware often needs to exhibit a higher degree of robustness. If the firmware encounters an error or a fault,
it may need to autonomously reboot the device or trigger specific routines to recover from the issue.
This is essential in ensuring the continuous and reliable operation of the device, especially in critical systems where downtime is not an option.

How is Firmware developed?

The development of firmware typically follows the hardware design phase, as hardware specifications and requirements shape the firmware's capabilities. However, there is some overlap in the development process - firmware engineers may need the hardware to change to ensure certain functions are possible so that the device meets specifications.

Firmware engineers will typically use development software that is recommended by the microcontroller manufacturer. These provide a comprehensive suite of tools such as compilers, debuggers, and simulators tailored to the specific microcontroller architecture, simplifying the coding process and optimising performance.

Additionally, firmware developers often leverage libraries provided by the microcontroller manufacturer or third-party sources. These libraries contain pre-written code snippets and functions that expedite development, by offering ready-to-use solutions for common tasks, such as interfacing with hardware peripherals or implementing communication protocols.

Want to learn more about Firmware?

If you want to learn more about firmware check out some of the previous blogs:

One year developing Firmware with AI : Our experience with integrating AI into our workflows.
Test-driven development : A powerful method for writing modern-day firmware.

What next?

Looking to outsource your firmware R&D? Whether it's upgrading existing products or venturing into new firmware development projects, our expert team is here to support you. Start by contacting us for a free consultation.

One year developing Firmware with AI

Wed, 13 Mar 2024 04:42:19 GMT

Introduction

In our ongoing effort to continuous improvement, we’ve embraced the emergence of Artificial Intelligence and integrated it into our workflow. The adoption of such technologies has improved the quality of our code while also freeing developers to focus on higher level decision making throughout a project. By trialing a number of AI tools, we’ve seen improvements in our development practices, albeit not without encountering a set of challenges.

This piece aims to share how AI has improved our firmware and what we’ve found useful, what AI doesn’t do well and how we’ve mitigated this, and what we hope the next year will bring in terms of Artificial Intelligence with regards to embedded systems firmware.

Although there are a lot of AI tools out there, this article focuses in on our experience with using GitHub Co-pilot

How has AI changed the way we develop firmware?

GitHub Copilot is an AI-powered code completion tool that integrates into our code editor (Visual Studio Code). It assists developers with writing code by generating suggestions for small to medium chunks of code by leveraging a generative AI model trained on public code repositories. Copilot predicts and generates code suggestions as developers type. You may have noticed tools in other fields implementing similar features to help predict user actions. Gmail for example are developing a tool to help you draft emails which you can read about here .

GitHub Copilot can accurately anticipate the developer's next steps based on the context of the code being written. This predictive capability enables software and firmware engineers to focus on higher-level design aspects and complex problem-solving, rather than getting bogged down in repetitive coding tasks. Take this scenario:

A developers needs to write a driver for a TMP451 temperature sensor they are using in a weather station they are developing. They write the function below which reads the temperature back over a communication bus.

They test the function and if all went well they move on to the next function which is an API to get the humidity. Before AI assistance, the developer would now have to write a very similar function to get the humidity, but with small and important details changed.

With Co-pilot however, it sees how the get temperature function was made and makes a prediction on what the humidity function will be.

The AI suggestion has copied the logic and style from the "get temperature" function but changed the important parts so it is now reading back the humidity. We reviewed this code and it is indeed correct!

This example highlights some important features of developing with AI:

Consistently styled code - The code it suggests is written in the same style as the rest of the file. This doesn’t necessarily affect the performance of the code but it does make it easier for developers to read and thus improves the code’s maintainability.
No human errors - When developers copy and paste it’s easy to miss or forget to change parts and often leads to code that looks correct but isn’t. This could be reworded to: "This leads to hard-to-diagnose bugs which have a high chance of not being detected until far further down the development pathway.
Saved time - It obviously saved time by writing the whole function in seconds, but an underestimated factor is it took less mental energy of the developer to do it. This allows developers to stay focused on the high level architecture and not get bogged down in the details. Of course the details are very important, but it takes a lot less energy to review the function than to write it.

It’s important to mention that even when writing the original temperature function, Co-pilot was making suggestions here too. Co-pilot saw the temperature was calculated above so there is a high probability that we would want to "print" (log) this reading. The handy thing here is it gives a nicely formatted readable message.

Another thing that AI does really well is its documentation. GitHub Co-pilot has an inbuilt command specifically to force it to suggest documentation. You can see below that it’s generated a doxygen style comment for our get temperature function.

Limitations and Tips

While the integration of AI tools into firmware development has brought significant advancements, it has not been without its set of challenges. These limitations, ranging from issues of code relevance to the intricacies of managing AI-generated suggestions, have necessitated a thoughtful approach to leveraging AI technology effectively.

Where does AI struggle?

By the previous section you might be thinking AI tools could just write the whole program by itself. Fortunately for software developers, it’s not THAT good…. yet.

Unlike human developers who possess an innate contextual understanding and have project-specific knowledge, AI tools can sometimes offer solutions that are not relevant to the specific problem at hand and for really specific/niche problems, sometimes it just makes stuff ups!

Here is an example where the developer wants to use the law of sines to calculate the ground distance and GitHub Co-pilot just makes up an equation. This points out the importance of code review to add an additional layer of scrutiny, ensuring that AI tools aren’t slowing down and making the development process worse.

AI tools need to understand the problem space to generate relevant solutions.This often involves detailed input from developers, which can be time-consuming and seem like a waste of time. However, refining the process of inputting context can markedly improve the quality of their outputs, making them more applicable and useful.

You can see the developer here has given the AI contextual information in the form of a comment explaining the steps to correctly initalise the sensor. Without this the AI cannot really give an accurate suggestion. As you can see its done a pretty good job now that it does have the information.

In the specific case of Github Copilot, it also grabs contextual information from open tabs in the IDE. This gives the AI context beyond other code written in the file. Other tools have specific places to input contextual information. For example Open AI’s GPT playground has a system message for this very reason.

Looking Towards the Future of AI in Firmware Development

It's clear that the landscape of AI is continuously changing, with new advancements on the horizon that promise to further revolutionise our development processes. Looking ahead, we anticipate several key trends and developments that could shape the future of firmware development with AI.

Specific fine tuned AI Models - While generic AI models are undoubtedly valuable and continue to improve, there is a growing desire for more specialised models tailored for specific tasks, such as embedded system development. Many existing AI tools have been primarily trained on languages like Python and JavaScript, highlighting the need for targeted training on languages and frameworks commonly used in firmware development. By focusing on developing and training AI models that are optimised for the intricacies of embedded systems, we can further improve their contextual understanding of firmware development.

User Interface Improvements - While the advancements in AI tools for firmware development have been significant, there is room for improvement in their user interfaces. Tools like GitHub Copilot, while valuable in functionality, can sometimes be unpredictable and their suggestions can end up just being distracting. Similarly, interfaces like the GPT playground have been noted for their complexity and lack of user-friendliness. It is understandable that these tools prioritize functionality over sleek user interfaces, given the complexity of their underlying AI models. However, moving forward, there is an opportunity to enhance the user experience by making AI tools more intuitive and responsive to user needs.

Higher Level Descion making - Expanding the scope of AI capabilities in firmware development has shown promise in enhancing lower-level coding tasks. However, when it comes to higher-level decision-making processes such as specification gathering, firmware architecture design, and code reviews, the utility of AI tools has not yet been fully realized. Moving forward, there is a growing anticipation for AI to play a more significant role in these critical areas. By leveraging AI for tasks such as automating specification gathering, optimizing firmware architecture, and automating code review processes, we aim to unlock new efficiencies and insights that can elevate the overall quality and innovation in our development projects. As AI technologies evolve, we look forward to harnessing their potential to drive advancements in these higher-level aspects of firmware development.

Conclusion

Finding the balance between leveraging AI-generated assistance and relying on human expertise is crucial. Developers need to discern when to accept AI suggestions and when to proceed based on their professional judgment and understanding of the project. This delicate balance underscores the increasing importance of establishing comprehensive project specifications at the outset. Doing so lays a good foundation for the collaboration between AI tools and human developers. Clear project requirements early on, allows developers to harness AI technology effectively, enhance productivity, and elevate the overall quality of firmware.

If you have a project that involves firmware development, feel free to reach out to us to discuss the potential benefits we can offer with this AI-integrated approach.

The Art of the Launch: How Real-Time Diagnostics Can Secure Product Success

Tue, 12 Dec 2023 19:42:15 GMT

How to Nail Your Products Release.

Product reliability issues can escalate into crises that exact a heavy toll on customer satisfaction and brand reputation. No product is 100% perfect and no doubt issues will arise, but how do you best capture these issues quickly before they get a chance to snowball? Take this hypothetical scenario for example.

Example Scenario:
Navigating customer issues the outdated way.

Customer Complaint:

Jane recently purchased a smart thermostat from your company and is now facing a persistent issue: her device keeps resetting to a default temperature, undoing her personalized settings. This problem leads to fluctuating indoor temperatures and undermining the energy efficiency benefits Jane expected. Feeling the sting of high utility bills and the annoyance of a non-cooperative device, she contacts your customer support for help.

Initial Troubleshooting Challenges:

With no advanced diagnostic tools at their disposal, the customer support team relies on traditional troubleshooting methods. They query Jane for details surrounding each reset incident. This approach depends on Jane's memory and descriptive abilities and fails to provide the technical granularity needed for proper diagnosis.

Difficulty in Pinpointing the Problem:

The engineering team is left in the dark, without error logs or system snapshots to shed light on the issue. They make educated guesses on the cause of the resets, but without solid data, their efforts could lead to misdiagnoses and misdirected resources. There's also uncertainty whether this is a widespread firmware issue or a hardware fault unique to Jane's thermostat.

Inefficient Development and Deployment:

The guesswork continues as the firmware team hastily crafts a patch based on speculation. Without facilities for remote updates, Jane must endure a complicated manual process to apply the fix—a procedure far from user-friendly and fraught with risks. Moreover, since the engineers cannot replicate the issue, they've effectively turned Jane into an involuntary beta tester.

No Controlled Rollout or Direct Feedback:
The lack of a structured deployment system means all new factory devices receive the untested firmware while existing customers are left with faulty units. As more customers report similar problems, a manageable situation quickly spirals out of control. A full product recall becomes the last resort, a costly and brand-damaging move.

The Impact on Brand Reputation:

The updates are finally implemented across all units, but the inability to confirm their effectiveness breeds more uncertainty. Customers, including Jane, have faced inconvenience and disappointment, prompting public expressions of dissatisfaction and skepticism towards your product and, by extension, your company as a whole.

This hypothetical scenario paints a stark picture of the pitfalls that await a product that doesn’t have the necessary tools to handle such issues. Keep reading to find out about one of the tools we use.

Introduction

Electronic development, or any product development for that matter, can be an elaborate and nuanced endeavor. Success is marked not only by the product launch but also by its sustained performance and reliability in the field. Post-launch, project managers often face the formidable task of managing field issues that, if left unchecked, can lead to detrimental effects on the user experience and the brand's reputation. Moreover, the financial implications associated with recalls, extended support, and compensatory actions further accentuate the need for vigilant post-launch management.

This article aims to introduce you to the idea of real-time diagnostics and hopefully enlighten you to the significance they play in mitigating post-launch hazards, explore a tool that Beta Solutions has utilized in previous projects, known as Memfault, followed by a revisit to the above scenario, leading to a much happier Jane.

The Value of Real-time Diagnostics:

As mentioned above, maintaining the operational integrity and user satisfaction of a new electronic product post-launch demands vigilant monitoring and rapid response to emerging issues. Real-time diagnostics play a pivotal role in this maintenance process, transforming the traditional reactive methods of problem-solving into a proactive, strategic asset for any organisation.

What Do We Mean When We Say Real-time Diagnostics?

Real-time diagnostics refers to the continuous monitoring and analysis of a product's performance data as they function in their operational environment. This process detects, logs, and addresses issues immediately as they occur, without delay.
The system provides instant access to device data, error codes, and device metrics that would typically only be available if the device was physically connected to a pc.
With real-time diagnostics, companies have a live feed of their product's health, allowing them to maintain high standards of performance and reliability.

What Is the Proactive Approach?

The proactive approach in diagnostics anticipates potential issues before they become user-facing problems, as opposed to the reactive approach that responds only after a problem has been reported.
Proactive diagnostics is characterised by the prevention of faults, predictive maintenance, and the optimization based on ongoing data analysis.
By implementing a proactive strategy, companies can ensure greater product uptime, a better customer experience, and a more robust understanding of their product’s real-world usage.

What are the Benefits of Real-Time Data?

Real-time data analysis supports continuous improvement of the product by identifying usage trends and customer preferences.
It enables data-driven decision-making for feature updates, hardware revisions, and customer support initiatives.
Real-time data provides an invaluable feedback loop for development teams, resulting in more agile and responsive product evolution.

What's Memfault?

Memfault is a state-of-the-art Real-Time Diagnostic Tool that we, at Beta Solutions, utilize to ensure our partners products perform at their best. It's effectively a comprehensive health monitoring system for electronic devices, offering a suite of powerful tools that work in unison to detect, diagnose, report on, and resolve issues as they occur. Here are some key points that make Memfault great:

1. Monitoring:

Memfault continuously tracks how a device is functioning in real-time. It collects information on how the device is performing, such as its temperature, battery life, and whether all parts are operating correctly. It's akin to a 24/7 surveillance system that's always watching to ensure everything is working as it should.

2. Diagnosing:

When something goes wrong, Memfault is ready. It gathers all the relevant information so the engineering team has everything its needs to identify the problem. It can spot errors, breakdowns, and other issues using the data it collects.

3. Reporting:

Once an issue is detected, Memfault sends an alert. This could be akin to a warning light on your car's dashboard that tells you when it’s time for an oil change. It notifies the engineers or the product team about the issue so they can start working on a solution.

4. Updating:

Memfault can send out updates or fixes directly to the device, similar to how a computer receives software updates. These updates can correct errors, introduce new features, or improve existing ones, all without the user having to take the device to a service center.

5. Analyzing:

Beyond fixing immediate problems, Memfault collects data over time to find bigger trends and patterns. It helps companies understand how their devices are used and how they can be improved in future designs.

Memfault's Backend

Upon logging into your Memfault account via their website (which you can access anywhere with an internet connection), you'll find yourself at the command center for all your devices and the data they generate.

The dashboard is your primary interface which showcases and overview of your fleet and the features mentioned above. A project manager might allocate merely two minutes at the beginning of their workday to visit this dashboard, but this is enough time to obtain a snapshot of their devices' status and address any issues that need to be handled. You can see all your devices on the devices page where you can search and filter based on things like what software version or which cohort they have been assigned to.

Cohorts serve to categorize devices into organized groups based on intended function or product stage. An example would be the grouping of devices into a "Test" cohort, which encompasses units designated for internal validation and pilot testing. These devices are used for latest firmware updates. Conversely, a cohort named "Production" might be established to encompass all devices post-manufacture but pending distribution to end-users. This systematic arrangement via cohorts effectively maintains order and structure within the device fleet.

All issues can be seen from an issues page where you again can filter and track issues based on a range of parameters such as:

Occurrence time
Software version
Cohort
Hardware version

Clicking into a specific issue gives you all the information an engineer would need to diagnose the Problem. It reveals:

Which devices have had this issue
The times the issue has occurred
Device logs
The coredump of the device (This is essentially what exactly was the device doing when the issue occurred)

This capability is nothing short of remarkable. It's akin to having the device right in front of you the very moment the problem arises, which significantly speeds up the diagnostic process. Additionally, you're equipped with a complete history of the issue each time it has occurred.

Metrics: Devices check in every hour and report to memfault any key metrics you want to know about them. Things like:

Battery level
Number of time the device was turned on
Cellular signal (RSSI)
Number of times a feature was used on a device
Temperature

All this information is graphed in a timeseries plot so you can see how any devices metrics have changed over time.

Alerts can be setup to notify you or the engineering team of potential problems. This helps catch issues before the customer does. An example of one being setup to alert everyone on the team if a device stays below 10% battery for more than 2 hours.

Firmware management: Here you can manage different firmware versions. With the ability to select which cohort should have what firmware. You can also stage firmware rollouts by choosing to update only 10% of your device fleet to mitigate the risk of introducing more issues. This is amazing to beagle to test new features and being able to compare them to previous versions at the same time.

With an understanding of the challenges encountered in traditional device management, as illustrated through Jane's experience with her smart thermostat, let's examine how the scenario could unfold differently. By integrating real-time diagnostics tools into the process, we witness a much different outcome.

Example Scenario:
Navigating Customer Issues With Real Time Diagnostic Tools

Customer Complaint:

Imagine a customer, Jane, who has recently purchased a smart thermostat from your company. She contacts customer support to report that her thermostat sporadically resets to a default temperature, negating her preset schedules and resulting in discomfort and increased energy costs.

Initial Assessment with Real time diagnostic tools:

The customer support team acknowledges Jane's issue and tells her to not worry, they’ll look into it. They no longer have to rely on Jane’s memory and descriptive ability they can just log into the platform and search for Jane's device ID. Upon finding the device, they review the recent logs and notice several entries that correspond to the times Jane reported the thermostat resets.

Identifying the Root Cause:

The logs indicate a recurring error code that precedes each reset event. Additionally, the support team retrieves a Core Dump that was automatically recorded during the last incident. Customer support passes this infomation on to the engineering team. Good news, they are already aware of this as were alerted of Jane’s issue at the time it occurred. Using remote diagnostics tools, they had reviewed the core dump and error logs and determined that a firmware bug caused the thermostat’s operating system to crash under a specific set of circumstances, leading to a full system reset.

Developing a Firmware Fix:

With the issue identified, they can now recreate it on their devices. They create a new firmware version that addresses the issue and thoroughly test it in a controlled environment to ensure that it resolves the problem without introducing new issues. No more relying on customers to test the firmware.

Rolling Out a Segmented Update:

Once satisfied with the stability of the new firmware, the team decides to roll out the update using the Segmented Rollout feature.They start with a small percentage of devices that exhibit similar issues to Jane's thermostat. By monitoring these devices closely after the update, they can confirm that the fix works and that no new issues have arisen.

Broad Deployment and Resolution:

With positive results from the limited rollout, the team proceeds with a full fleet update. Jane's thermostat, along with all others affected, automatically receives the update. The device applies the update during a time that minimizes disruption for the user.

Customer Follow-Up and Confirmation:

After the update is deployed, customer support reaches out to Jane to confirm the resolution of her issue. They leverage real time diagnostic tools to ensure that her device has received and is running the new firmware version. Jane confirms that her thermostat has maintained the set temperatures and that the resets have ceased. She is pleased with the proactive approach and the timely fix, reinforcing her trust in your company's brand.

Closing the Feedback Loop:

The engineering team uses this experience and the data collected through Metrics to improve their development process. They plan to utilize Key User Metrics and Device Health indicators more effectively to prevent similar issues in the future. The proactive, data-driven approach offered by real time diagnostic tools ensures continuous enhancement of the product and customer satisfaction.

Conclusion

In summary, the ability to responded to real time issues efficiently and effectively has a huge impact on a products success. Not only do you keep you devices out in the customers hand working reliably, you also gather useful information about how your product is being used to aid in future iterations.

If the concept of Real Time Diagnostic Tools resonates with you, or if you have a product that could benefit from it, do not hesitate to get in touch with Beta Solutions. Our team has experience using the platform and is ready to bring it into your product strategy, ensuring you stay ahead in the dynamic world of electronics.

How Legendary All Blacks Utilise Real-time Player Tracking Technology

Fri, 20 Oct 2023 00:26:00 GMT

Decoding the Tech Behind Rugby World Cup Player Analytics

Overview

The Rugby World Cup is pushing portable, durable, athlete data measuring wearables into the limelight. This year biometric wearables company STATSports is serving as the go-to team for the high-level player tracking needs of the legendary All Blacks [2]. What are these wearables? How do they work? What are they being used for? From the type of sensors employed, their safety profile, power efficiency, and robustness, to the real-time retrieval of data and the categorizing of the collected metrics, this post will unwrap the technicalities of sports wearable trackers. This post will unwrap the technicalities of sports wearable trackers, from the type of sensors, employed, their safety profile, power efficiency, and robustness, to the real-time retrieval of data and the categorising of the collected metrics.

The Heart of Sports Tracker: Sensors

Performance tracking wearables leverage several types of sensors for collecting data. Fundamentally, they include GPS units, accelerometers, gyroscopes, and magnetometers — the last three collectively referred to as inertial measurement units (IMUs).

1. **GPS units** measure the players position with a typical accuracy of 1-5m. Multiple constellations can be used to improve accuracy, along with augmented wide-band positional beacons for best-in-class positioning [3].

2. **Accelerometers** measure linear acceleration, which is valuable for tracking sudden changes in speed and direction. They are much more precise than GPS units for this, but suffer from positional accuracy losses over time.

3. **Gyroscopes** pick up the rate of rotation around an axis, making them highly effective for measuring movements like spins or turns.

4. **Magnetometers** sense the strength and direction of magnetic fields. While their use in day-to-day sporting activities is relatively minimal, they can provide valuable insights in niche cases.

In evaluating player safety, some trackers incorporate more specialised sensors like shock sensors. These units measure impact force, which can be critical information in collision sports like rugby.

There are a large number of companies providing wearable trackers for athletes, but they generally all contain some mix of the above sensors as part of their product offering. Sensor fusion algorithms may be used to combine the data and mitigate the weaknesses of each sensor type.

The Wearable Factor: Safety, Comfort, and Durability

For wearables like sports trackers, ergonomic design is paramount. The devices must be compact and lightweight enough not to inhibit the wearer's movements or pose a safety risk in event of a fall or accident. Ensuring they do not pose a safety hazard involves designing them without any sharp corners and securing them firmly to avoid them loosening during intense movements. The World Rugby Organisation has strict regulations on the size and shape of wearable devices allowed at the Rugby World Cup [4].

Being on the wearer for extended periods, these devices are exposed to varying elements and must therefore be robust. This durability is achieved through ingress protection (IP) rating systems that ensure resistance to external factors like dust and water.

Battery Life: Making Energy Efficiency a Priority

Given their intended use during prolonged sports events or intense training sessions, sport trackers must manage energy efficiently. Optimizing the device's power management demands a delicate balance, ensuring battery longevity without compromising the frequency or precision of data collection. This must be enclosed in a small product that when worn must just about “disappear”. when playing rugby The device also needs a system for timely alerts of low battery levels to avoid interruptions in data collection.

Live Data Retrieval: Ensuring Timely Feedback

For real-ti me decision making in a dynamic sports scenario, the ability to retrieve data live from the device is crucial. This feature typically requires wireless communication protocols, such as Bluetooth, cellular or Wi-Fi. These technologies allow the device to communicate instantaneously with a computer, smartphone, or another receiving unit, facilitating on-the-fly strategic adjustments based on the received data.

What Metrics and How?

The precise metrics that a sports tracker collects depend on the specific sport and what elements are most valuable to monitor. Commonly tracked metrics include speed, acceleration profiles[5], distance covered, heart rate, impacts[5], and skin temperature.

More specialised metrics, like collision force or hyperextension in sports like rugby, may also be tracked to assess safety risks.

Advanced algorithms analyze these collected metrics, converting raw sensor data into valuable insights. Data from different sensors can be combined to produce more complex metrics - for example, data from an accelerometer and a gyroscope can be used together to evaluate a player's agility.

"One of the metrics we look at is high MET, and essentially, what we’re trying to understand is the acceleration/deceleration volume, or intensity."
– Nic Gill, NZ All Blacks , Head of Strength and Conditioning [1].

Hitting the Finish Line

The interplay of intricate technology, precise implementation, and nuanced understanding of individual sports requirements culminates in the impressive potential of sports wearable trackers. Observing their incorporation into big-league games like the Rugby World Cup is just a snippet of the possible applications.

As specialists in electronics product design , we relish the challenges and rewards that come with designing devices that accommodate such a collection of demands. With our wealth of experience, we can be valuable partners for anyone seeking to design their unique wearable tracking solution for other applications. We welcome you to contact us to discuss further.

We're excited to observe and participate in the continual evolution of the wearable tech space. The potential applications and benefits are profound and extend far beyond sport, in to areas like health monitoring and workplace safety. Watch this space for more insights into tech developments.

References

[1] STATSports (2023, Apr 20). STATSports X NZ All Blacks - Behind The Science. YouTube video. Retrieved 2023-10-17, from https://www.youtube.com/watch?v=DikSJQ1O9DY

[2] STATSports (2023, Sep 25). STATSports Signs Up Five New International Rugby Union Teams In The Rugby World Cup. Press release. Retrieved 2023-10-18, from https://pro.statsports.com/statsports-sign-5-rugby-world-cup-teams/

[3] STATSports. Apex Pro Series . Product page. Retrieved 2023-10-18, from https://pro.statsports.com/apex-pro-series/ .

[4] World Rugby. Equipment Specifications - Player Monitoring Devices . Ret rieved 2023-10-18, from https://www.world.rugby/the-game/facilities-equipment/equipment/specifications/pmds .

[5]: PitcheroGPS (2022). GPS Rugby Trackers . Product page. Retrieved 2023-10-24, from https://www.pitcherogps.com/pages/rugby-gps-tracker .

[6]: Bionics Inc. Titan Sports . Product page. Retrieved 2023-10-24, from https://www.titansensor.com/ .

Advancing the future with today’s innovative Satellite solutions

Thu, 09 Mar 2023 21:54:43 GMT

With over half a million New Zealand businesses connected to some form of IoT and approximately 13.1 billion connected devices world-wide (42 billion installed and unconnected), it is not surprising that some form of global coverage will be a necessity in the future [1][2]. This is even more evident when we consider the inherent limitations of our more traditional terrestrial networks. Although we have relied on them to support most IoT applications, the fact remains that only 10% of earth's landmass is served by cellular towers (or 63% of the global population in 2021) making connectivity extremely difficult in remote areas [3][4].

Fortunately, the advent of cheaper and more abundant access to orbit, along with a new wave of miniature satellite technology, has led to an explosion of satellite launches on the scale of multiple mega-constellations [5][6]. This has allowed the number of active satellites to more than double in just a few years from 2017 to 2021 (fig. 1), and has been slowly but surely bringing more and more land-mass online [6].

Currently the largest use-space contender for satellite mega-constellation technologies are telecommunications providers such as Starlink , OneWeb or Amazon’s project Kuiper . These companies are seeking to bring high-speed data and broadband to isolated areas around the globe [7].

So what does this mean for more simple IoT device functionalities, such as assembly line monitors, weather sensors or tectonic plate movement detectors? Today only 2% of global satellite coverage is represented by IoT connectivity revenues [8] with the remainder providing a number of other essential services such as navigation, telephone calls, weather climate modeling and ever increasing broadband capabilities [6]. At this stage in time, utilizing a service such as Starlink for these applications would be outrageously expensive.

This provides an opportunity for lBesser known companies, such as Inmarsat , Iridium , ORBCOMM , Globalstar , who account for over 80% of the IoT satellite market, to provide these essential services, and and are expected to hold approximately 20% of the global market by 2026 [8]. Revenues from these companies are expected to grow 14 times faster than revenues from traditional satellite use cases.

Technologies for Satellite IoT

There are three types of satellite technologies that support IoT, known as LEO, MEO and GEO (low, medium and geostationary earth orbit).

Despite the differences in orbital patterns, all three types of satellite networks generally transmit data the same way. To begin, an IoT device must transmit data to a satellite, and in some cases, it will also receive data/commands from the satellite. Upon receiving this data, the satellite transmits it to a ground station, which processes and forwards it via the internet to the end user, who is then able to organize and use it to manage their IoT project successfully.

LEO Satellites

LEO satellites take up the bulk majority of IoT satellites in orbit today for a number of reasons:

Low power : They are better suited for low power applications due to their relatively low deployment height (~200 - 2000 km). A shorter distance to earth means less power is required to transmit and receive a signal to earth. This also reduces propagation signal loss [9].
Simple: Signal strength and power demands of transceivers are low, which simplifies the design process and reduces the complexity of components. As a result, revolutions have occurred in the industry such as the development of 'cube-sats', which allow for the mass production of commercial parts for cost-effective IOT satellite solutions. Cube-sats are currently the industry's most preferred and popular option [8].
Low-cost: As the satellites are closer to earth, less resources are required to place them in orbit, thus enabling cheaper and faster deployment. This low orbit allows LEO satellites to make one orbit around the earth in 90 minutes. When set up in constellations, transmitted signals will always be picked up by one of the satellites in the orbital ‘mesh’.

Today, many would-be LEO satellite swarms for IoT are yet to be deployed, with only two thirds of operators having launched any form of satellite(s) into orbit. Overall there is demand for low cost satellite connectivity, however the economics can be challenging for these companies as satellite constellations tend to have high fixed costs and low variable costs. This makes it difficult for operators to repay initial costs and creates a tricky situation for investors. It is currently expected that many LEO satellite operators will cut their losses and decide not to launch. This has recently been the case for a few prospective operators such as Hiber, who have decided to rely on larger companies [9].

GEO Satellites

GEO or Geo-Stationary Equatorial Orbit satellites fit into a special category of applications because of the way they orbit the Earth. The clues to exactly how their orbit is different to other satellites can be found in the name ‘geo-stationary’, which means stationary orbit. This implies that to a ground observer, a GEO satellite appears stationary and does not move through the sky and around the planet, as a MEO or LEO would. Instead, GEO satellites are launched into orbit with a velocity that matches the spin of the earth in order to constantly remain above a set position on the Earth's surface [10].

Due to the Earth’s low speed of rotation (1670 km/h) [11], this type of configuration means that GEO satellites must be positioned much further away from Earth (~36,000km) [12]. This helps these relatively slow moving satellites to maintain orbit and also provides a large coverage area over the entirety of Earth’s disc. These factors make them ideal for applications that require little data interruption such as weather monitoring, television broadcasting and low-speed data communications [13].

IoT applications utilizing GEO satellites are uncommon because of a combination of factors surrounding their extremely high earth orbit such as:

High latency (~250ms).
Powerful signal amplification requirements.
Satellite dish and signal line-of-sight required (i.e. must be clear of buildings or trees).
High service costs due to more complex and longlife satellites.
High launch costs.

One advantage of GEO satellites is that they typically hold large solar panels and powerful antennas. This means that they can be ideal for more critical IoT networks when factors such as ultra reliability, uninterrupted data links or large amounts of data transfer are required [14].

MEO Satellites

Medium earth orbit satellites (MEO) orbit somewhere in between GEO and LEO satellites from around 10,000km to 20,000km above the Earth. In comparison, MEO satellites are relatively rare and are almost exclusively used for GPS/Maritime navigation, communications and other similar services.

MEO satellites are typically considered a compromise between LEO and GEO satellites because of their 'medium' earth orbit, which offers low latency (LEO), a greater coverage area (GEO), and continuous coverage (GEO). For full geographic coverage, MEO constellations are still needed, but far fewer than LEO constellations [10].

With all this in mind, it may seem like MEO satellites could be a perfect compromise for IoT applications however, MEO satellite applications for IoT are relatively uncommon for many similar reasons as GEO satellites. These reasons range from high launch costs, high serviceability and larger, more complex satellites being required to transmit the greater distance [15].

What are Micro Satellites?

As LEO satellites are already inherently cheap to launch and manufacture, companies have turned to ‘shrunken’ satellite technologies referred to as anywhere between ‘micro’ and ‘femto’ satellites to keep the cost of satellite IoT services as low as possible.

The weight of a large satellite is typically around 1,000 kg, whereas the weight of a microsatellite can range anywhere between 10 and 100 kg. Even smaller satellites are known as Nano (1 - 10kg), Pico (0.1-1kg) and Femto (>0.1kg) [17]

In this field, Swarm Space operates a swarm of the smallest global coverage IoT satellites in orbit today (just 11 x 11 x 2.8cm per device, see figure 4). Each SwarmSpace satellite weighs approximately 400 grams, making it a 'pico' satellite. In addition to their small size, they are also incredibly optimized for IoT with the company often able to beat competitor pricing (e.g. ORBCOMM or Iridium ) by a tenth or even a hundredth.

Currently a Swarm data plan for a single IoT device costs as little as $5 USD per month and provides 750 data packets per month of up to 192 bytes per packet and 60 downlink (2-way) data packets [18]. Granted this is an extremely small amount of information for modern day standards, it is more than enough for simple IoT applications such as taking weather measurements or transmitting location data.

Swarm Space has managed to fill an extremely trendy market as of late. Their success combined with rising product applications and the low cost of sending these tiny satellites to space has created a worldwide movement. Several companies are now actively manufacturing these tiny satellites, such as Pumpkin Space Systems , EnduroSat , or AAC Clyde Space , and are likely to reap significant gains in the future.

Since all satellites are capable of transmitting and receiving such data, selecting the type of satellite network to run from ultimately falls upon the specific IoT application and depends on a number of factors such as service reliability, data importance, budget, device location and more.

Conclusion

As IoT applications become more prevalent in our everyday lives, satellite IoT is a rapidly growing industry. According to industry experts, the satellite IoT market is currently crowded and competitive with much of the competition having still not launched any form of tangible product. It has even been referred to as 'overheating' before even having a chance to grow and it is likely that only a few competitors will survive. With this in mind, the way of the future looks to be with LEO Satellite technologies and providers.

The key advantages of LEO Satellites means that these Satellites are used for a variety of applications. Their use is expected to continue growing in the coming years. The increasing number of LEO satellites in orbit is leading to new and innovative applications, which are closely linked to Global Positioning Systems (GPS) These Satellites have revolutionized the way we live and work, making it possible to providing mapping and tracking in real-time e.g. the location of ships and aircraft, vehicle navigation while traveling, and improve weather forecasting. The use of GPS and satellites has become an essential aspect of modern life and has far-reaching applications in fields such as agriculture, transportation, and military operations.

Here at Beta Solutions we have experience in designing electronic solutions that integrate GPS functionality to provide tracking, mapping or location based services. Get in touch if you have a project that may require these capabilities.

References:

[1] (2021). INTERNET OF THINGS . Spark. https://www.spark.co.nz/iot/home/?cq_src=google_ads&cq_cmp=11350525164&cq_term=iot&cq_plac=&cq_net=g&cq_plt=g&gclid=CjwKCAjw7p6aBhBiEiwA83fGujYrUcYPVmvns6Y9vRWNkDufbPuAexLZETExB5SNO2kbsqW-AFqRbxoChkwQAvD_BwE&gclsrc=aw.ds

[2] G, N. (2022). How Many IoT Devices Are There in 2022? [All You Need To Know] . TechJury. https://techjury.net/blog/how-many-iot-devices-are-there/#gref

[3] The World’s Problem & Lynk’s Solution . Lynk World. https://lynk.world/our-technology

[4] Percentage of global population accessing the internet from 2005 to 2021, by market maturity . Statista. https://www.statista.com/statistics/209096/share-of-internet-users-in-the-total-world-population-since-2006/

[5] (2018). How Small Satellites Are Providing Low-Cost Access to Space . IEEE. https://site.ieee.org/sb-uol/how-small-satellites-are-providing-low-cost-access-to-space/

[6] (2021). The Impact of Mega-Constellations on Astronomy – Zooniverse Project . Staffblogs. https://staffblogs.le.ac.uk/physicsastronomy/2021/08/05/the-impact-of-mega-constellations-on-astronomy-zooniverse-project/

[7] (2022). How Starlink Satellites Could Affect IoT . Link-Labs. https://www.link-labs.com/blog/how-starlink-satellites-could-affect-iot

[8] Pasqua, E. (2022). Satellite IoT connectivity: Three key developments to drive the market size beyond $1 billion . IoT Analytics. https://iot-analytics.com/satellite-iot-connectivity/
[9] Hatton, M. (2022, March). The LEO satellite space will be overcrowded, at least for IoT. transformainsights.com. https://transformainsights.com/blog/leo-satellite-iot-overcrowded#:~:text=The%20IoT%20focused%20LEO%20roll,surface%2C%20but%20are%20constantly%20moving.

[10] (2022). Types of satellite networks making IoT solutions a reality . IoTm2mCouncil. https://www.iotm2mcouncil.org/iot-library/articles/smart-industries/types-of-satellite-networks-making-iot-solutions-a-reality/#:~:text=Types%20of%20satellite%20connectivity%20available,orbit%20(MEO)%20and%20geostationary .

[11] (n.d.). Challenge 10 – Shadow Speed and Earth’s Rotation . Eclipse2017.Nasa.Govt. https://eclipse2017.nasa.gov/shadow-speed-and-earths-rotation#:~:text=Earth%20rotates%20once%20every%2023h,h%20%3D%201%2C674%20km%2Fhr .

[12] (n.d.). GEO, MEO, and LEO . SatelliteToday. https://www.satellitetoday.com/content-collection/ses-hub-geo-meo-and-leo/

[13] Kohrs, R. (2010). Geostationary Satellites . NOAA. https://sos.noaa.gov/catalog/datasets/geostationary-satellites/#:~:text=Geostationary%20satellites%20are%20a%20key,on%20Earth%20all%20the%20time .

[14] Kostina, O. (2022). Types of Satellite Networks Used in IoT Solutions . IoT For All. https://www.iotforall.com/types-of-satellite-networks-used-in-iot-solutions#:~:text=GEO%2DSatellites&text=Typically%2C%20a%20GEO%2Dsatellite%20network,be%20sent%20up%20and%20down

[15] Heukelman, C. (2018, June 26). LEO vs. MEO vs. GEO Satellites: What's the Difference? symmetryelectronics. Retrieved from https://www.symmetryelectronics.com/blog/leo-vs-meo-vs-geo-satellites-what-s-the-difference-symmetry-blog/
[16] Blackman, J. (2019, November). What is LEO, and how will LEO satellites transform the IoT sector? rcrwireless.com. https://www.rcrwireless.com/20191104/fundamentals/what-is-leo-and-how-will-leo-satellites-transform-iot

[17] Geospatial World. (2016, May 26). How are small satellites changing the earth observation industry? [Video]. YouTube. https://www.youtube.com/watch?v=vd4oecayaMk

[18] (n.d.). SWARM M138 MODEM. SwarmSpace. https://swarm.space/product/swarm-m138-modem/#popup1

Unprecedented NZ Motorway Radar

Thu, 04 Aug 2022 01:49:48 GMT

The Transmission Gully Motorway has a ‘first of its kind in NZ’ radar system [1].This system enables the operations team to respond to events immediately, 24/7, and uses radar sensors to monitor the motorway at all times for incidents which require a response. The system has 100% coverage of the 28 km motorway, which could quite possibly make Transmission Gully one of the safest stretches of motorway in New Zealand [2].

These radar sensors, together with conventional monitoring systems like CCTV cameras, form an effective Intelligent Transport System (ITS) to ensure the Transmission Gully motorway meets the highest safety standards.

CCTV cameras, mounted at key points, can be directed by operators to look in the direction of a radar detected event to assess the degree to which the Incident Response Team must react and/or co-ordinate with emergency services.

Systems like this are used widely in Europe but are new to New Zealand. They’re a proven technology and we’re excited to see them integrated on more NZ motorways.

How does an ITS Radar System work?

The process by which an ITS system responds to an event is as follows [3]:

The radar system detects an anomaly or a condition which exceeds a pre-set threshold
The system alerts an operation team who resides in a control room. (This control room is manned 24/7/365 to ensure that events can always be responded to effectively).
An operator then acknowledges the event in the system to mark it as ‘under investigation’
The operator would then direct a CCTV camera at the segment of road for which an alert has been generated to assess the nature of the alert.
Lastly, the operator can decide to trigger the Incident Response Team resolve the issue and/or coordinate with emergency services.

The radar system can detect dangerous events like:

crashed vehicles
stopped vehicles
vehicles traveling in the wrong direction
obstacles on the road (E.g. Like a box that has fallen off a truck or a fallen tree)
animals
pedestrians, or
any object larger than ~20cm that’s not meant to be there.

The ITS uses a system of individual radar sensors at regular intervals along the side of the motorway, for the entirety of the 28 km length. This ensures there are no dead spots.

Each sensor is mounted on a ~4m pole above the road surface so its view is less likely to be obstructed by vehicles traveling in proximity to it. The specific sensor that was used on this project is the CTS350-X from NavTech Radar who are based in the UK.

Housed within the white radome is a continuously rotating radar transceiver antenna assembly.

The circular element with concentric marks in the middle of the image is a lens used to collimate the projected radar RF emissions to achieve a beam width of just 1.8°. This ensures the resolution of the sensor is sufficient to sense and discern small objects (>20cm) from the background.

The operating frequency range of the radar transceiver is 76 - 77 GHz. This falls under the mmWave classification as the wavelength is just ~4mm. This is a very similar frequency to the mmWave radar sensor Beta Solutions developed for RadarIQ Ltd , where that sensor operates in the 60Ghz range.

Unlike the Navtech sensor, the ‘RadarIQ-M1 Sensor’ is a solid state sensor with no moving parts. It achieves this with an array of RX/TX antennas to form an synthetic aperture to capture depth and velocity data in 2.5 spatial dimensions. The horizontal field of view is 110 degrees while the vertical field of view is just 15 degrees. Hence 2.5 spatial dimensions (2.5D) rather than 3D.

The Navtech sensor achieves 2D sensing using a single radar transceiver (RX/TX) which rotates 360° 4 times per second. NavTech Radar also supplies a ITS software solution which they call ClearWay™. This video demonstrates how their system works to improve motorway safety - Clearway Stopped Vehicle Detection

Conclusion

Only time will tell as to how effective the Clearway system is in improving safety on the Transmission Gully motorway, but we can be sure that with 100% coverage of the 28 km motorway, t he motorway traffic and activity is being intensely monitored. We look forward to seeing if the road safety statistics in the years to come reflect the claims made by the road's designers.

We would love to talk to you about any potential new radar product development ideas you may have or any other R&D electronics project you may want to pursue. Please do get in touch with us!

References:

Global Chip Shortage - How Beta Solutions has adapted

Wed, 27 Jul 2022 01:17:23 GMT

Introduction

Since 2008 Beta Solutions Ltd has been privileged to provide electronic design and production solutions for our clients. During this time (like many businesses) there have always been challenges* and opportunities, and learning to adapt has been essential.

Our most recent significant challenge has been the global chip shortage. If you haven’t already, refer to our previous blog “ Global Chip Shortage - Why We Are In This Crisis ”, to learn more about what this crisis is and the causes behind it.

As Andy Grove once said of business: “There are two options: adapt or die.” Needless to say - we have chosen to adapt(!) and this blog provides answers to the following questions:

How have we adapted to the challenges faced by the global chip shortage?
What are the implications for our clients?
How long will these adaptations last?

* For example, commencing business during the GFC in 2008!

How have we adapted to the challenges faced by the global chip shortage?

Design Process - Traditional approach pre-shortage

Research and Development (R&D) can be a complex process, and the more Research that is required - the more convoluted this process will be. (This is because Research by definition contains an element of uncertainty).

Suffice to say that the R&D process does not always follow a nice water-fall project flow - where the tidy completion of one stage then leads to the commencement of the next phase.

That said, in very simplified terms , our hardware design process would generally follow these following phases:

R&D - Architecture and Specification
The project is scoped out, and a “block level” preliminary design and specification is formed.
“Key components” are identified and a general risk assessment is performed.

An example of a key component would be a motor driver for a project that contains one or more motors. Most projects also include a microcontroller, which is also categorized as a key component.
R&D - Design Schematics
This is where our engineers take the block-level design and transform this into an actual meaningful logical circuit design in CAD.
Key components are (mostly) locked-in and all the other components are selected.
R&D - Design PCB
Whereas the previous phase is concerned with establishing the logical function of the electronics, the PCB design phase is where the actual circuit board is designed in CAD.
R&D - Prototyping and Testing
Components are ordered along with the bare PCBs - enough for a few prototypes to be made. We build enough boards for hardware and firmware verification

R&D - Design Iterations (If required)
For all but the most simple PCBs, it is reasonably common practice to make some optimisations to the design after testing and prior to going into production.
Production
The R&D phase is complete and we move into production, where the Bill Of Materials (BOM) is generated and production files are finalised. All components are then procured, and the PCB is fabricated!

NB: This entire process from Phases 1-6, could (usually) take anywhere from 6 - 24 months depending on the complexity of the project.

Look to our blog for upcoming articles about our R&D process in more detail.

Design Process - New (Adapted) approach post-shortage

Downside of Traditional Approach

While our traditional approach has worked well in the past, due to the global chip shortage, the problem is that the components that are selected during the architecture and design schematics stages could well be out-of-stock by the time the PCB is ready for production.

For a circuit board to work as intended, all intended components must be present. If a single component is unavailable, the production of an entire PCB (sometimes consisting of several hundred components) will be stalled.

Some electronic components have manufacturer lead times of 26 - 72 weeks (!).

New Adapted Approach

Clearly it is not acceptable to unexpectedly have to wait for this length of time before receiving their products to sell, so we have adapted our process to include the following three changes:

Ordering in Advance
We now aim to procure production component stock immediately after the Design Schematics Phase. This comes with an element of risk, especially considering the prototyping phase has not been completed at this stage and components may be required to change.

However, this is a risk that many of our clients are willing to accept, given that the alternative is to simply “roll the dice” and see if components will be available at production time.
Additional consideration in component selection
Traditionally components are selected based on function, price, availability and physical properties such as size and/or temperature range. Selection criteria has grown now to include drop-in-replacement alternatives, generics, and stocking history.

For example, it is often now more attractive to consider more highly priced components with a better supply chain, over cheapest parts that have excessively long lead times.
Design for alternatives
Sometimes there are component alternatives that are a drop-in-replacement (components that have a different part number, or even made by a different manufacturer, that will fit into exactly the same space on the circuit board) . However, while these can be common for simple components (resistors, capacitors etc), for more complex components, this can be difficult or not possible.

Therefore, a board can be designed that works with one of two possible components depending on stock at the time of production or future production runs. These can be selectable by placing certain supporting components or not.

Overview: Traditional Vs Adapted Design Phases

What are the implications for our clients?

Andy Grove’s quote again (“There are two options: adapt or die”) can be taken quite literally in a business sense. If our customers do not have stock to sell - and therefore the ability to trade - there is every possibility they will have to permanently close their doors.

Despite the Global Chip Shortage, we have been able to change the way we undertake product development - which largely mitigates the impacts (especially the excessively long leadimes) on our customers.

That said, there are some implications that our clients should be aware of, these being:

Cash Flow
R&D is an investment and as such there is a capital outlay required. Moreover, not only does the client have to fund the R&D process, they also have to fund Production. Together and R&D and Production form the “Total Project Capital Outlay”.

Regarding the two different approaches (Traditional Vs Adapted) - the Total Project Capital Outlay doesn’t significantly change. However, what does change is when the Capital is expended - which is known as Cash Flow.

Previously, Production Costs were incurred at the time of manufacturing - in essence, at the end of the project). Now a sizable portion of these costs have shifted and are incurred much earlier in the design process, meaning our clients need to have such funds in place.
Implications on Risk

As previously mentioned, the lowest risk approach is to wait until the R&D phase is ABSOLUTELY COMPLETE, prior to ordering Production Stock.

Ordering Production Stock prior to the completion of the R&D phase does always carry some risk - as any changes to the design may mean that pre-purchased stock may no longer be required for production. In this instance the customer would have incurred a “sunk cost”.

However, this is a calculated risk that many of our customers are willing to accept - given the extremely long lead time alternative!

[2] There are likely some increased overheads with the Adapted Approach - due to (for example) additional freight for the components. Also, if components have been ordered, and then later deemed to be non-needed, these components could be a sunk cost.

How long will these adaptations last?

Analysts predict an easing in the second half of 2022 as demand-supply gaps decrease. Semiconductor and component research analyst William Li said just recently that “coupled with wafer production expansion and continuous supplier diversification, we have witnessed significant improvement in the component supply situation, at least in the first quarter.”

For the moment though, the global tech supply chain remains uncertain. Manufactuaring the chips used to make our electronics and consumer goods, still depends on the stability of our world. As we still manage a global pandemic, implications of lockdowns with ongoing Covid waves, and the effects of the war between Russia and Ukraine and the supply of neon*, we are not on the home straight as yet.

*Neon is a colourless and odourless gas, and plays a vital role in the manufacturing of the tech that we use everyday.

Conclusion

Here at Beta Solutions the Global Chip Shortage has resulted in us changing the way we execute our R&D and Production projects. Our new (adapted) approach has meant that we have largely mitigated the most deleterious effects of the crisis - enabling our customers to still meet their target production schedules.

Our new (adapted) process remains for the foreseeable future - at least for another few years - to benefit our clients, both in the way we undertake product development, and how we largely reduce the impacts of long lead times and risk to our customers.

We would love to talk to about any potential new R&D electronics projects, or concerns you may have about the implications of the Semiconductor shortage to your business. Please do get in touch with us!

References:

Analysts Predict End is Near for Global Chip Shortage https://www.tomshardware.com/news/analyst-predicts-end-of-chip-shortage
The neon shortage is a bad sign https://www.vox.com/recode/22983468/neon-shortage-chips-semiconductors-russia-ukraine

Global Chip Shortage - Why are we in this Crisis?

Thu, 24 Feb 2022 02:02:50 GMT

What is the chip shortage?

The global chip shortage of 2020 - 2022 is an on-going problem affecting many industries and businesses. Those affected may not even be in manufacturing, but could be anyone who relies on modern technological devices.

Certain products are getting harder to source and/or the price of such goods is rising[1]. This has been felt severely in the automotive industry, with examples such as Nissan making half a million fewer vehicles[2]. The computer gaming industry has been similarly affected with gaming consoles such as XBox and Playstation, and PC graphics cards having long waiting lists and premium prices[3]. Hidden behind these shortages and price increases is an underlying problem - the inability of the supply of integrated circuits, also known as chips, to keep up with demand.

Where do chips come from?

Integrated circuits (ICs or simply chips) are in almost every product we use today. They are in our fridges, microwave ovens, washing machines, cars, headphones, TVs, phones, watches, toys, eftpos machines, lab equipment, air conditioning, and so many more devices.

Chips are cheap to make en-masse, but setting up the facilities to make them can take several years and billions of dollars. There is a limited number of these facilities, known as fabrication-plants or ‘fabs’, around the world[6].

A lot of manufacturing of electronic goods (as well as non-electronic goods) were, or still are, produced by a production model of lean manufacturing, also known as just-in-time(JIT) manufacturing. JIT boils down to ordering components of a product just as they are needed to produce the item. This reduces waste from stock-piling components and the costs of storing these; it does however require a robust supply chain[7].

So, what happened before, during and after 2020 to cause the global chip shortage?

A perfect storm of problems coincided in a short enough time-frame to cause a singular problem which we now label as “The global chip shortage”.

Geopolitical tensions were rising between the United States and China.
Severe weather events at key fabrication-plants and serious fires at others, caused major production issues.
The global pandemic which we will refer to as Covid-19 for the purposes of this article, which increased demand almost over night.
Shipping and local deliveries slowing down the supply chain from raw material to finished product, which has also played a role in this shortage.

Let’s look at these four points in detail:

1. Trade Wars

The precursor to the Global Chip Shortage was the rise in tensions between the United States and China.

Then U.S. President Donald Trump initiated a trade-war with China as far back as 2018, which included a 25% tariff on Chinese imports, including semiconductors (IC chips) and their raw materials[15]. Now only 5% of chip imports to the U.S. are from China, but imports from other countries have not increased enough to offset the supply loss.

In another move in the trade war, in 2019, the same presidency banned the sale of U.S. semiconductors to the Chinese company Huawei. This caused some U.S. manufacturers to scale back production as their primary client was suddenly shut to them. China boosted production and began hoarding the stock. This further disrupted supply chains causing a long-term flow on effect, which is still being felt.

The Biden White House has not yet repealed the Trump tariffs. They have made negotiations with other countries to boost trade of semiconductor commodities, as well as measures to drastically increase American production capability.

In conjunction with this, a Japan - Korea trade war in 2019 started with Japan placing restrictions on raw materials that were used to create chips. Korean manufacturers such as SK Hynix and Samsung were impacted, which added to the destabilisation of the global supply chain.

2. Disasters

Taiwan drought

Taiwan, home to TSMC (Taiwan Semiconductor Manufacturing Company Ltd), which is one of the largest chip fab’s in the world, suffered a drought over the entire year of 2020. Taiwan usually relies on seasonal wet seasons to refill it’s reservoirs of water, but they did not get a reprieve until May of 2021 [10].

TSMC uses 156 million litres of purified water per day[9]. Much of this is recycled, and the company did not report any downturn in normal production levels. However it did prevent them from scaling up production to cope with the new heightened demand.

Texas Snowstorm

In February of 2021, the state of Texas suffered extreme cold weather for an unprecedented long time. Texas is home to many of the United States chip fabs, such as Samsung, Texas Instruments, NXP, Infineon, and others.

The cold weather and snow caused power-grid problems as well as staffing safety issues. Plants either shut their doors (on short notice) or went into emergency operating mode [13].

Fires

Nittobo suffered a factory fire in Fukishima, Japan in 2020. They are the world leaders in printed circuit board (PCB) substrate material[15]. This caused supply issues downstream in manufacturing for 6 months.

A large three-day fire raged through a Asahi Kasei Microsystems (AKM) factory in Japan in Oct of 2020. This high-end audio chip manufacturer disaster caused major supply problems for other audio manufacturers such as Solid State Logic, TASCAM, miniDSP, Merging Technologies, SPL of Germany, RME, and others [16].

A fire at a Renesas Japan (chip) fabrication plant, in March of 2021, caused a significant drop in capacity from that company. Production was fully restored by the end of that June [11].

As recently as 2nd of January 2022, a fire at a plant in Berlin, Germany, owned by ASML Holding has a lot of people worried. This plant manufactured the lithography machines which are sold to chip fabs. They are the largest supplier of these machines, and the only source for extreme ultraviolet (EUV) lithography machines [12].

3. The Pandemic

In late 2019, the Virus SARS-COV2 began causing the disease known as Covid-19, and reached pandemic status in early 2020. A lot of the global population was encouraged to stay at home to prevent further spread of the virus.This drove demand for work-at-home products such as laptops, webcams, new phones, etc, sky-high.

Also in hot-demand were entertainment devices such as PC graphics cards and gaming consoles. The PC graphics cards were also being competed for by crypto-miners causing even more demand [18]. Some chip fab’s changed their tooling to accommodate more for the consumer market to meet this demand, which then impacted the automotive industry again [17].

An unprecedented rise in e-commerce happened almost over night as public spaces were closed due to the pandemic. This put a strain on shipping and last-mile delivery, not to mention port-workers and truckers either staying home or getting sick.

4. Shipping and Supply Chain

The global supply chain of everything, including products involved in chip making, is primarily made up of shipping and ports. Ships account for 80% of global trade. The system is a complex industry of logistics, and pre-pandemic running at near capacity for efficiency.

When demand for trade goods did not slacken after the Christmas peak of 2019, strain was put on the system. Then the pandemic became global, causing shipping and dock workers to stay home. The increase in e-retail put even more strain on the already strained industry [14].

What is the plan to solve the crisis?

New Fabs

Intel announced in March of 2021 that they were building two new fabrication plants in Arizona, USA, to the tune of $20 billion US dollars [20].

TMSC is also looking to expand into the US, tentatively at first. They are hiring US engineers and executives to do a year’s training in Taiwan before returning to work at the new facilities in Arizona [21].

Samsung will start construction of a new fab in Texas in the first half of 2022. It expects it to be operational in the second half of 2024 [22].

Texas Instruments plans to expand its current operations with 4 new fabs at a cost of $30 billion USD [22].

There are other fabs being built around the world, such as singapore and the EU, as the supply chain has shown its problems.

Most of these new fabs are expected to come online and begin production sometime in 2023 or 2024, and whilst we can be encouraged at the positive impact these will bring to reducing the crisis in the long-term, , there is no short-term relief to the current problem

Predictions from experts

Pat Gelsinger, CEO of Intel - “The global chip shortage is set to last until 2023. Demand continues to soar amid the coronavirus pandemic even as semiconductor manufacturers rush to expand production.”

Lisa Su, CEO of AMD - “The global chip shortage will become less severe in the second half of 2022. We’ve always gone through cycles of ups and downs, where demand has exceeded supply, or vice versa,this time, it’s different.”

Gina Raimondo, Commerce Secretary for the Biden administration - “We aren’t even close to being out of the woods as it relates to the supply problems with semiconductors. There is a significant, persistent mismatch in supply and demand for chips. The semiconductor supply chain remains fragile.”

Conclusion

Problems with the supply of semiconductor chips, both in terms of maxed out production, lean manufacturing and supply chain fragility, were prevalent well before the global pandemic sent demand out-of-control.

A global work-force that is negotiating working-from-home, sickness and self-isolation has all had a significant impact on supply chains and halted many ships, ports and distribution stations. Couple this with severe events such as droughts and fires, and the pressure on the supply end magnifies.

As the Trade-war between China and the US continues, with no new policies on either side to resolve this, the underlying politics of the supply chain problems remain unresolved as well. [23]. The private sector is betting on geographical diversification, such as companies primarily based in East Asia building new fabs in the U.S. or the E.U., or vica versa, as a solution to alleviate the political situation[19].

The chip shortage has certainly presented many challenges for businesses locally and globally, and here at Beta Solutions we are no exception. What has resulted out of this though, are many opportunities to problem solve and explore new ways to do things to alleviate delays in delivery and price increases as a result of this global shortage.

Keep an eye out for part 2 of Global Chip Shortage - New ways of doing things, where we will share some of ways of working with you.

References:

The Chip Shortage Is Driving Up Tech Prices—Starting With TVs https://www.wired.com/story/chip-shortage-electronics-prices-tvs-displays/
Nissan to make half a million fewer cars in 2021 due to chip shortag https://www.cnbc.com/2021/05/13/nissan-to-make-half-a-million-fewer-cars-in-2021-due-to-chip-shortage.html
Inside the GPU Shortage: Why You Still Can't Buy a Graphics Card https://www.pcmag.com/news/inside-the-gpu-shortage-why-you-still-cant-buy-a-graphics-card
PBS Why is There a Global Chip Shortage? https://www.youtube.com/watch?v=YEtu9JbDih0
WSJ Why the Global Chip Shortage Is Hard to Overcome https://www.youtube.com/watch?v=FP_g-as29x0
List of semiconductor fabrication plants https://en.wikipedia.org/wiki/List_of_semiconductor_fabrication_plants
Coronavirus pandemic exposes fatal flaws of the 'just-in-time' economy https://www.abc.net.au/news/2020-05-02/coronavirus-pandemic-exposes-just-in-time-economy/12206776
Secretive Giant TSMC’s $100 Billion Plan To Fix The Chip Shortage https://www.youtube.com/watch?v=GU87SH5e0eI
Why Taiwan’s drought means you can’t have a new smart TV https://www.smh.com.au/world/asia/why-taiwan-s-drought-means-you-can-t-have-a-new-smart-tv-20210524-p57uq1.htm l
Parched Taiwan prays for rain as Sun Moon Lake is hit by drought https://www.theguardian.com/environment/2021/may/09/parched-taiwan-prays-for-rain-as-sun-moon-lake-is-hit-by-drought
UPDATE 10 - Notice Regarding the Semiconductor Manufacturing Factory (Naka Factory) Fire: Production Capacity Recovery Status https://www.renesas.com/us/en/about/press-room/update-10-notice-regarding-semiconductor-manufacturing-factory-naka-factory-fire-production-capacity
Fire at vital tech factory could worsen global computer chip shortage https://www.newscientist.com/article/2303316-fire-at-vital-tech-factory-could-worsen-global-computer-chip-shortage/
Texas winter storm blackouts hit chip production https://www.ft.com/content/ec2f93ad-d23c-4ff4-867a-59385d1cc8a6
14 You shopped like never before. The supply chain couldn't handle it https://www.cnet.com/news/features/you-shopped-like-never-before-the-supply-chain-couldnt-handle-it/
PCB Maker Nittobo Suffers Major Fire That Could Drive Hardware Price Hike https://hardwaresfera.com/en/noticias/hardware/incencio-nittobo-pcb-hardware /
AKM Factory Fire—A Pro-Audio Industry Disaster https://www.prosoundnetwork.com/business/akm-factory-fire-shakes-up-pro-audio-industry
Understanding the global chip shortage, a big crisis involving tiny components https://www.popsci.com/technology/global-chip-shortage/#:~:text=What%20is%20the%20chip%20shortage,rippled%20up%20the%20supply%20chain .
Why the Chip Shortage Hasn’t Been Fixed Yet https://slate.com/technology/2021/11/chip-shortage-semiconductors-demand-pandemic.html
The Global Chip Shortage: A Timeline of Unfortunate Events https://info.fusionww.com/blog/the-global-chip-shortage-a-timeline-of-unfortunate-events
Intel is spending $20 billion to build two new chip plants in Arizona https://www.cnbc.com/2021/03/23/intel-is-spending-20-billion-to-build-two-new-chip-plants-in-arizona.html
Taiwan’s TSMC begins hiring push for $12 billion Arizona facility https://www.taiwannews.com.tw/en/news/4085851
Texas To Get Multiple New Fabs as Samsung and TI to Spend $47 Billion on New Facilities https://www.anandtech.com/show/17086/texas-to-get-multiple-new-fabs-as-samsung-and-ti-to-spend-47-billion-on- new - facilities#:~:text=Samsung%20will%20start%20construction%20of,up%20all%20the%20necessary%20equipment .
The supply-chain disaster that is eating Christmas is being driven by a Biden-Xi conflict that many are overlooking https://www.businessinsider.com.au/supply-shortages-semiconductor-chips-crisis-us-china-trade-war-biden-2021-10?r=US&IR= T

5 Technologies to Watch in 2021

Mon, 03 May 2021 23:56:03 GMT

To say that we are passionate about technology at Beta Solutions is an understatement!

We keenly follow new technology developments and often have lively debates on the merits of such technologies over our morning coffee. The most fun part however is thinking up ways how these new technologies could take our clients’ products to the next level. That being said, we are also very aware of the value of tried and trusted technologies and the risks of incorporating leading edge technologies in product development projects. To this end, we sometimes do internal research projects first to gain knowledge and experience on a new technology that we think may be of particular value to many clients.

Below are 5 intriguing technologies that we would like to highlight for you.

1. Lithium-metal Batteries

Has the limitations of electric vehicles, namely their cost and the fact that you can only drive a few hundred kilometers before you have to stop for a lengthy recharge, put you off purchasing one?

QuantumScape ¹ is claiming to be redefining the frontier of battery technology with their solid-state lithium-metal battery. Their early test results are said to have shown that the battery could boost the range of electric vehicles by up to 80%.

A solid-state lithium-metal battery is a battery that replaces the polymer separator used in conventional lithium-ion batteries with a solid-state separator. The replacement of the separator enables the carbon or silicon anode used in conventional lithium-ion batteries to be replaced with a lithium-metal anode. The lithium metal anode is more energy dense than conventional anodes, allowing the battery to store a greater amount of energy in the same volume. Some solid-state designs use excess lithium to form the anode, but the QuantumScape design is ‘anode-free’ design in that the battery is manufactured anode free in a discharged state, and the anode forms in situ on the first charge. - QuantumScape

According to QuantumScape ¹ , relative to a conventional lithium-ion battery, solid-state lithium-metal battery technology has the potential to:

Increase the cell energy density - by eliminating the carbon or carbon-silicon anode;
Reduce charge time - by eliminating the charge bottleneck resulting from the need to have lithium diffuse into the carbon particles in conventional lithium-ion cell;
Prolong life - by eliminating capacity fade that results from the unwanted chemical side reaction between the carbon and liquid electrolyte in conventional lithium-ion cells;
Improve safety - by eliminating the combustible organic porous separator and organic anolyte material in conventional cells; and
Lower cost - by eliminating the anode materials and manufacturing costs.

The battery is still just a prototype, but QuantumScape has a deal with Volkswagen who claims it will be selling EVs with this new battery type by 2025 ² .

The success of lithium-metal batteries could finally make EVs attractive to millions of consumers.

2. Green Hydrogen

Imagine livin g in a futuristic city with flying taxis, robots for servants, and that has almost no pollution. Well, this city, NEOM, in northwest Saudi Arabia on the Red Sea is already underway - being built from the ground up as a living laboratory. ³ Green Hydrogen is set to fuel this Smart City, from a green hydrogen facility that US-based company Air Products and Chemicals has been developing in joint venture for the past four years. ⁴

Green Hydrogen is the production of hydrogen from the electrolysis of water. The hydrogen produced is carbon-free because the energy source to power the electrolysis is from wind and solar.

The mass production of green hydrogen is said to be enabled by ⁵ :

Innovations in electrolysis technologies that improved system efficiencies to nearly 90%,
Decreasing costs of renewable energy sources - more low-cost solar photovoltaic and wind power plants globally,
Increased economies of scale with yearly additions to electrolysis capacity and increased average project sizes.

Figure 1: A schematic overview of a hydrogen based energy economy. ⁶

The benefits of lower-cost green hydrogen will could be advantageous to advancing environmental sustainability goals as well as various industries. For example ⁵ :

Heat
Carbon emissions can be reduced in residential and commercial heating by blending hydrogen with natural gas,
Countries with low renewable energy potential can reduce their carbon footprint by importing green hydrogen to produce electricity.
Transport fuel
Electric vehicles’ power can be provided by either:
Charging the EV’s battery by plugging into the electricity network, or
Converting hydrogen in the EV’s tank to electricity through fuel cells
Which could be more cost-effective than hydrocarbons and battery-powered EVs in countries with high electricity prices, and
Many studies have shown that fuel-cell electric vehicles are more cost-effective than battery-powered ones for long-haul heavy duty transportation.
The technology for zero-emission hydrogen fuel cells in transportation already exists with hydrogen powered busses, trains, trucks, cars, aeroplanes, and even forklifts. ⁷
Urban air mobility
Flying cars, air taxis, and advanced drones would need to run on hydrogen fuel because of the specific power and energy density it offers. (As NASA knew over 50 years ago, carbon fuels are too heavy and batteries die too quickly to get small vehicles off the ground.) ⁷
Uber Elevate and many other air taxi developers are turning to lithium-ion batteries, hydrogen fuel cells, or a combination of these to get zero-emission urban air mobility vehicles off the ground. ⁷

Figure 2: Hydrogen Fuel Cell (Retrieved from AZO Cleantech )

3. Hyper-accurate Positioning

As discussed in our previous Blogs, New Zealand Trialling Centimetre Level GNSS and Precision GPS for NZ... on it's way , the latest positioning technologies have accuracies within a few centimeters or millimeters.

These super-accurate technologies are opening up new possibilities for various industries, including delivery robots, self-driving vehicles, rescue operations, and even landslide warnings.

Among other reasons - this is being made possible by:

Advancements in ground-based “base stations” infrastructure, which provide precise calibration data - essentially accounting for (and correcting) the inherent variations of the satellite radio signals as they pass through the ionosphere and troposphere.
New and exciting technologies (such as the Precise Point Positioning (PPP) ) are enabling sub 10 cm accuracy with just a single receiver.
China’s global navigation system, BeiDou (or Big Dipper), which was completed in June 2020, provides positioning accuracy to anyone in the world of 1.5 to 2 meters. ⁸
And GPS got an upgrade with four new satellites for GPS III launched in November 2020 and more expected by 2023. ⁸

4. 5G Networks

The future of 5G is something we are excited about because its promised faster data transmission combined with increased bandwidth will be a powerful catalyst for innovation.

In our Blog, 5G for IOT? , we discussed the current release’s shortcomings from an IoT perspective. However Release 17, which is not scheduled to be frozen until at least early 2022, holds many possibilities, including ⁹ :

Uploading and downloading data from remote locations with the kind of speed most people associate with Wi-Fi,
Sharp improvements in the way applications can be written, deployed and interacted with by mobile users,
Advancements in data-intensive applications and IoT.

5. Multi-skilled AI

Although there has been vast advancements in artificial intelligence recently, it can be said that robots and AI remain “dumb” in many ways - especially when solving new problems or navigating unfamiliar environments ² . This is because they lack the human ability to apply learnt knowledge in a specific context to new situations. For example, DeepMind’s game-playing algorithm AlphaGo can beat the world’s best Go masters, but it cannot apply the learnt strategy beyond the realms of the board ¹⁰ . So although deep-learning algorithms are good at detecting and learning patterns, they cannot “understand” and adapt instinctively to a changing world.

An interesting new methodology to improve the skills of AI is to expand its senses. What if you can create a new multi-modal intelligent system by combining the advanced capabilities of speech technology, vision technology and natural language processing? Would AI with all these abilities be able to “talk” about what it sees and hears - gaining human-like intelligence? Would AI with multiple senses be able to achieve more flexible intelligence and show a degree of common sense?

This may just be possible due to the latest advances made in language-processing algorithms like OpenAI’s GPT-3 which combines with computer vision to create bimodal visual-language AI models. Last September, the researchers at the Allen Institute for Artificial Intelligence, AI2, created a model that can generate an image from text - showing the ability for algorithms to associate words with visual information. And not long after that in November a model was developed by researchers at the University of North Carolina which can incorporate images into existing language models. Using these ideas, OpenAI extended their GPT-3, releasing two visual-language models in early 2021. One of the models can link objects in an image to words that describe them and the second model can generate images based on a combination of concepts it has learnt. For example, if you instruct it to “create a painting of a surfer sitting on a beach with his surfboard at sunset”, it can create many combinations of such a painting based on what it has learnt before about the different items ¹⁰ .

These more sophisticated multi-modal systems could arguably be less easy to trick, be more reliable as it may fail less in unfamiliar surroundings, and thus be safer. This opens up may possibilities for the likes of more advanced robotic assistants and self driving cars, to name a few. It will be interesting to see how this development evolves.

Figure 3. Multi-skilled AI (Selman Design - retrieved from MIT Technology Review )

Conclusion

Some of the technologies above already have proven results, like precision GPS, whilst others still have some way to go. All-in-all, it will be interesting to see where they end up and what solutions they will unlock.

We hope that introducing you to some of these interesting technologies have sparked some ideas of how your product could perhaps be taken to the next level or how some consumer problem may be solved. Feel free to contact us if you would just like to discuss the possibilities over a coffee or if you would like for us to assist you in your next electronic product development project.

References:

Referenced from FAQs - QuantumScape
Referenced from 10 Breakthrough Technologies 2021
Referenced from NEOM: an accelerator of human progress
Referenced from The Green Hydrogen Revolution Is Now Underway
Referenced from The dawn of green hydrogen
Retrieved from https://www.researchgate.net/figure/Schematic-overview-of-a-hydrogen-based-energy-economy_fig1_328037427
Referenced from Council Post: It's Time For Elon Musk To Admit The Significance Of Hydrogen Fuel Cells
Referenced from 10 Breakthrough Technologies 2021
Referenced from Council Post: Technology Trends That Will Lead The Way In 2021
Referenced from Multi-skilled AI
Green H2 Image created using images retrieved from Pixabay
Multi-skilled AI image by Stefan Keller from Pixabay
Hyper-accurate Positioning image retrieved from Financial Times .
5G Networks Image by mohamed Hassan from Pixabay
Banner image background retrieved from Image by Gerd Altmann from Pixabay

Things To Know About Lightning Strikes & Electronics Design

Thu, 29 Apr 2021 00:50:33 GMT

Introduction

In the real world, electronics can be subject to all sorts of nasty electromagnetic events. These 'transient' events could damage or destroy devices unless careful consideration has been given in the design stage to integrating protections. Large transient energies caused by lightning strikes (aka surges), electrostatic discharge (aka ESD), Electrical Fast Transients (EFT), and other electromagnetic phenomenon including can all damage electronic devices. These energies can be introduced through ports or anywhere a user can physically touch the electronic device.

Devices which have been damaged by environmental transients are often difficult to diagnose in the field. The device may appear to have stopped working or it may exhibit erratic behaviour.

The measure of a device's robustness to transients is it’s Electro-Magnetic Compatibility (EMC). This is a measure of how a device interacts with it's electromagnetic environment, and with other equipment. Most markets legally require that your device be EMC tested in a lab to prove it is compliant with their EMC requirements before it can be sold. Failing this testing is expensive as this requires redesign and retesting late in the project.

Many EMC problems are not simple or obvious, so they must be considered at the start of the product design. Leaving these considerations to the end of the design cycle can lead to overruns in the engineering budget and schedule.

See our previous article on compliance for more general EMC information - Electromagnetic Compatibility (EMC) Compliance - Answers to Frequently Asked Questions .

This article specifically addresses how lightning induced transients (surges) can damage your device even if there was no ‘direct’ strike.

How Lightning Damages Electronic Devices

Direct Strikes

The most obvious way in which lightning destroys electronic devices is, of course, by a direct strike. These will almost vaporize anything and as such there aren’t many practical ways that a device could be protected from it.

The image below shows a household electrical distribution box which has been destroyed by a direct strike.

Figure 1: A household electrical distribution box which has been destroyed by a direct lightning strike.

Indirect Strikes

Thankfully direct Strikes are very uncommon but there are also indirect effects from nearby lightning strikes which can still wreak havoc. It is actually possible to protect against the damage these indirect effects can cause as their energy is magnitudes lower than a direct strike.

These indirect effects are much more common than direct strikes because the effective radius of a lightning strike is so much larger than the small area the lightning directly strikes. A strike could indirectly damage electronics even hundreds of meters away.

The image below shows a product’s circuit board/PCB which had components partially vaporized by indirect lightning effects. There’s a damaged shield bonding capacitor (these are normally rated for 1-2kV) and a damaged ethernet transceiver IC. The pressure from the chip being vaporised internally was enough to blow off part of the encapsulation compound!

Figure 2: An example of a product which had components partially vaporized by indirect lightning effects.

The effective radius of a strike can be hundreds of meters because of the incredible power that they contain. The average lightning strike contains, on average, approximately 1 billion joules of energy. This sounds like a lot of energy but, to give this some perspective, this is ‘only’ enough energy to boil ~24 cups of coffee. Lightning is so destructive because this modest amount of energy is delivered in just milliseconds or even microseconds. 1 billion joules over 1 millisecond is 1 trillion watts of power .

Even just a small portion of this energy coupled into a device is enough to cause damage.

There are two mechanisms that allow energy from a strike to couple into a device indirectly. These are:

Near and far field electromagnetic coupling
Ground potential gradient coupling

Near-Field and Far-field Electromagnetic Coupling

Figure 3: Graphic illustration of electromagnetic pulses propagating outwards from a lightning strike.

The incredible voltages and currents of lightning strikes produces electromagnetic pulses (EMPs) which propagate outwards (at the speed of light) in electromagnetic waves. The magnitude of these waves can be so high that they couple or induce significant current and energy into nearby conductors (e.g. power lines or communication cables). The coupled current can be on the order of hundreds to thousands of amps.

There are two simple factors which gauge how much energy is transferred:

The proximity to the lightning strike (EMP source) - The intensity of the EMP pulse adheres to the inverse square law with respect to distance from the strike. Intensity = 1/(Distance^2). I.E. the intensity of the EMP quickly drops off the further away the strike is.
The length of the cable run - Generally the amount of energy transferred is a function of how long the cable run is that is exposed to the EMP. As a rule of thumb, cable runs longer than 10 meters may be susceptible [1].

It’s important to note that buildings and the materials they’re made from do little to impede or attenuate EMPs. Cable runs are almost as susceptible inside as they are outside. Dangerous energy can couple in regardless.

Ground Potential Gradient Coupling

This coupling mechanism works through the interaction of lightning strikes and the earth (e.g. dirt, soil, etc.) around impact points.

Soil has a finite resistance and therefore the earth is literally a giant resistor. The current from the strike on earth travels outwards from that point. This current produces voltage potentials in the earth's surface as it travels out that then drops off in magnitude as the current spreads out and the current density reduces. Ohm's law states that where there is current flowing through a resistive material that there must also be a voltage potential. For example - the voltage potential between two points on the earth, in proximity to a strike, could be tens of thousands of voltages different due to the charge flowing through the ground.

Systems which have multiple connections to earth (in different locations) can be susceptible to energy coupling through those grounding points from ground potential gradients induced by lightning currents.

One example of a common system which can be susceptible to this coupling mechanism is an ethernet cable that spans from one building to another building (e.g. the network of a university campus). Each of these buildings will have their own ground rod dug into the earth which the neutral and earth conductors of the building's wiring system are tied to. Therefore the electrical devices in that building are roughly at the same potential as the building’s earth rod. During the instance of a lighting strike in close proximity, these two buildings could be at potentials thousands or tens of thousands of volts different. This might not otherwise have been an issue if it wasn’t for the ethernet cable that connects the two buildings together. This cable would have equipment at either end which are at very different voltages. Ethernet is normally designed to be robust to these events through galvanic isolation, but it will still fail if the voltage potential is great enough to breakdown the isolation.

The image below illustrates this coupling mechanism.

Figure 4: Graphic illustration of ground potential gradient coupling.

Some ways that electronic products can be designed to be robust to transients

There are various components that designers can choose to protect their electronic products from lightning surges. The exact method implemented and which protection components are selected depends on the project and technical constraints.

Here are several categories of transient protection components which a designer can pick and choose from.

Transient Voltage Suppression (TVS) Components

TVSs are semiconductor devices designed to provide protection against transients by limiting voltage. They are designed to operate in the Avalanche mode and essentially clamp to a set voltage threshold.

TVS diode - TVS’s can be used to protect high speed data lines as they can have low capacitance. They’re also extremely fast so can react and protect for transients on the order of nanoseconds.

Zener Diode - Low cost and have a stable steady state voltage when reverse biased.

Metal Oxide Varistor (MOV) - Can absorb high energy transients and operate in ~1nS but have a high capacitance.

Crowbar Components

Crowbar devices conduct when the voltage over them reaches a threshold which causes them to trigger to an on-state. In this state the crowbar devices limits the voltage all the way down to approximately zero. Almost like a metal crowbar was placed across the conductors (hence the name)!

Gas Discharge Tube (GDT) - is a sealed device containing a special gas mixture trapped between two electrodes, which conducts electric current after becoming ionized by a high voltage spike. These devices are often rated to shunt thousands of amps. GDTs are much slower than TVSs as they take ~1uS to engage. This makes them less useful for protecting from ESD but they’re great for absorbing surge transients.

PCB Spark Gap - A transient protection feature can be designed into the PCB itself. Making this a 'zero' cost solution for diverting transients safely to ground. Normally a PCB spark gap consists of two copper traces separated by a small gap with no solder mask. High voltage transients can jump this gap but it does not conduct during normal operation. The main drawback is that the designer does not have much control on performance and the performance can degrade with time and environment.

Thyristor - A crowbar type thyristor is triggered into a low impedance state when the voltage across its terminals reach a certain voltage threshold. It then resets back into a high-impedance state once the transient current subsides and reduces below a threshold.

Protection by Galvanic Isolation

This is just a fancy way of saying that there is an ‘air-gap’ in the device or port which prevents a transient surge from being able to conduct to/from earth.

Galvanic isolation protects devices because transients' currents are prevented from ever entering the electronic device. Electricity can’t directly jump the gap without significant potential.

Figure 5: Graphic illustration of an isolation component preventing current from flowing during a ground potential transient.

Here are a few examples of common components which designers use to galvanically isolate power or data ports:

Transformer - Whereby power and/or data is sent over galvanic isolation by being converted to magnetic field flux by a coil, which is then passed through the transformers core and then converted back to electricity by another coil.

Optocoupler - Whereby data is sent over galvanic isolation by being converted to photons by a light emitting diode (LED) and then converted back to an electrical signal by a transistor. There are optocouplers that can be used to transfer power but this is more of a niche application.

Other Transient Protection Devices

Transient Blocking Unit (TBU) - circuit protection devices that block transients using a current disconnecting mechanism rather than diverting or shunting the surge to ground. These devices are triggered by high currents and work to limit current going into a device. TBUs are often used in conjunction with a component which can divert or shunt the transient current to ground.

Example RS485 Communication Bus Protection

The design below demonstrates how a designer might go about protecting a RS485 transceiver which could be attached to hundreds of meters of cable. The solution is robust to high high energy surge, EFT, and ESD transients.

Figure 6: Image of RS485 protection components on a Bourns development board. [2]

Figure 7: Symbolic representation of the above PCB design. [3]

The circuit design was designed with three different transient protection components:

Gas Discharge Tube (GDT)
This is used to crowbar/divert the <5kA currents from any induced surges on the RS485 bus.
Though it is not fast enough to suppress ESD or EFT transients.
It requires 150V to trigger.
Transient Blocking Unit (TBU)
This is used to help ensure the GDT can be triggered so it clamps during a surge.
These components trigger and limit the current to just 100mA in ~1uS.
They can limit current up to a voltage of 650V. Though the GDT will trigger before this.
TVS Diodes
These are used to suppress nanosecond transient like ESD and EFT events.
They also limit the voltage on the RS485 bus during a surge by diverting the 100mA let through by the TBU to ground.

RS485 transceivers are inherently tolerant to -7/+12V inputs, hence the choice of these TVS values are just above these values.

These parts all work in unison to make the RS485 bus incredibly robust to almost any transient which in turn improves reliability.

Conclusion

A designer can only hope to protect their electronic product until they have gained a fundamental understanding of the principles by which transients can be generated and damage their designs.

Lightning strikes can damage devices even if they do not experience a direct strike. The potential damage radius of a lightning strike can be hundreds of meters.

Designing for lightning and other high-voltage electrical transients can seem like a daunting task at first, but there are common, conventional, and proven methods for protecting electronic circuits from these events.

At Beta Solutions, our experienced electronics hardware engineers have the skill and proven track record to use the best suited design methodology to protect our clients' products from the transients they may ever experience. You can get in touch with us to discuss any idea you have in mind, or a problem you may need solved, via our contact page or give us a call .

References

Electronics and the America's Cup

Thu, 11 Feb 2021 00:49:51 GMT

“When you push off the dock, your life is in the hands of computers – and the guy piloting the boat out of the water, of course. All the systems on the boat, other than the winches rattling around, rely on computers…”

- Freddy Carr of Ineos Team UK ¹

Introduction

With much anticipation (and no-doubt “relief” considering Covid 19) the 36th America’s Cup is well underway! It has been exciting to see the boats of the three challengers (Luna Rossa Prada Pirelli, Ineos Team UK and NYYC American Magic) battling it out for the Prada Cup - with NYYC American Magic already sent packing. The winner of the Prada cup will compete against (the defender) Emirates Team New Zealand - with the aim to win the oldest international sporting trophy and sailing’s ultimate prize - the America's Cup.

Even if you are not a sailing fan, it’s difficult to not be taken aback by the sight of these 75 foot hi-tech mono-hull foiling boats - flying above the water at speeds topping 50knots (93 km/h). At those speeds (often more than 3 times the actual wind speed) the forces these AC75 boats generate are so large (tonnes) that it is understandable that the sailing crew need some help from some onboard electronics and actuators.

This blog takes a look at the role that electronics plays in the modern America's Cup era and how they are ultimately making these boats go faster than ever before.

Caption 1: More than ever before - electronics and control systems are playing a pivotal role in the America’s Cup. ²

Background

Electronics is an umbrella term which can extend to a whole manner of different “tech” used within the America’s Cup. While it’s safe to say the use of electronics has been increasing for decades, it is nigh impossible to know exactly what new electronics a team has implemented during a specific campaign - as it is often highly confidential and shrouded in secrecy due to the competitive advantage it brings.

Nevertheless, we know that since the 1980s ³ , boat designers have been increasingly using computers to help with CAD design - including copious amounts of hydrodynamic and aerodynamic simulations.

On board the boats, it has also long been the case that teams will use computers and visual displays to assist the sailing crew with various functions - such as:

Communications - enabling the crew to be in constant communication.
Navigation - Assistance with navigation, including monitoring course boundaries, ley lines and boat speed. (With this data, the boat's Velocity Made Good (VGM), can be calculated - which is one of the most important metrics in sailing as a higher VGM will get you to the next mark sooner).
Race start - Assisting the crew to ace the start, by optimising the “time on distance”. (In other words informing the crew when to “gun it” for the start line - not wanting to be too late nor too early).

34th America's Cup - A New Era

“It was August 2012 when the sailing world was turned upside down by a 72 foot catamaran flying in the Hauraki Gulf. Emirates Team New Zealand had brought the foils to the America’s Cup and changed the face of top-level yacht racing forever. Since then the increase in performance for America’s Cup boats has been greater than at any point in the 170-year history of the event.”

- americascup.com ⁴

The hydrofoils that Team NZ introduced to the America’s Cup work in a similar way to aircraft wings, in that they create lift. As boat speed increases the hydrofoil correspondingly creates more lift, until the lift force exceeds (and then balances) the weight of the boat, lifting the hull out of the water and dramatically reducing drag in the process. Lower drag means the boat can travel faster for the same amount of energy.

Caption 2: Hydrodynamic Force. ⁵ Hydrofoils essentially minimise drag, by lifting the “draggy” boat hull out of the water. Less drag equals more speed for the same energy.

Despite Team NZ inventing foiling, they were ultimately defeated in the 2013 America's Cup by Oracle Team USA by a score of 9 to 8, after Oracle won eight consecutive races from Race 12 onwards. (NB: As a New Zealander - it still hurts to recall!)

35th America's Cup - Smaller, Cheaper, Faster

The 2017 America's cup saw the boat sizes reduce in size (to 50 foot catamarans) in order to save costs. The challenger, Emirates Team New Zealand, won by a score of 7 to 1 over the defender, Oracle Team USA.

Despite the boats being smaller, not only were the AC50s significantly faster than the AC72s, but they “flew” on their foils far earlier in the wind range, at 7-8 knots true ⁶ . Not only could the AC50s foil downwind, they could now consistently foil upwind too.

The AC50s packed in more electronics than ever before to help control the boats to maintain their optimum zone. Indeed there was much fascination when the media spotted Team New Zealand wing trimmer Glen Ashby with an electronics device - dubbed his “Xbox console".

While the other race yachts in this year's America's Cup employed a traditional hydraulic winch system to control the wing, Ashby is adjusting the twist and camber in the wing via a toggle on his little handset. It gives Ashby the ability to make micro adjustments to the wing - sometimes just a case of a few millimetres - swiftly and accurately. ⁷

Caption 3: Glen Ashby with this so-called electronic X-Box controller, was used on Team New Zealand’s AC50, to help trim the sails ⁸ .

36th America's Cup - Foiling Monohulls

“Go downstairs on Britannia 1 and it feels like you could be in a space shuttle getting blasted off to the moon. The amount of computer programming and logic that goes into sailing the boats is mind blowing”.

- David Carr, INEOS Team UK ⁹

Caption 4: Ineos Team UK, flying above the water in their AC75. ¹⁰

The AC75 Class Rule

With Team New Zealand having won (and dominating) the 35th America's Cup in Bermuda, the privilege fell to them (in conjunction with the Challenger of Record - Luna Rossa) to define the Class Rule for the following campaign.

Rather than staying with catamarans - as per the previous three A/C campaigns - they chose to revert to the more traditional mono hulls. Audaciously however, these mono-hulls would not have a traditional keel, but would rely on electronically controlled hydrofoils to maintain stability and fly the boat above the waterline.

Highlights of the AC75 Class Rule include:
The hull can be any shape its designers conceive provided it fits into a theoretical defined box of the length, beam and draft.
Strict limitations on the number of components that can be built including hulls, masts, rudders, foils, and sails, thus encouraging teams to do more R&D in simulation and subsequently less physical construction and testing ¹¹
Supplied foil arms and cant system to save design time and construction costs ¹¹
Supplied rigging ¹¹
One design mast tube ¹¹
Maximum weight of 6520 kg - not including sails or crew.
(NB: The America's Cup boats between 1992 and 2007 boats weighed a maximum weight of 24 tonnes - of which the keel weighed a mighty 19 tonnes! While necessary, the keel also introduced tremendous drag, slowing the boats down).

Caption 5: The AC75 Class Rule, defines what aspects of the boat the designers have control over. The AC75 does away with the tremendously heavy keels of the past, and instead relies on Hydrofoils to keep the boat stable. ¹²

Ever since the AC75 rule was published on the 29th of March 2018, designers have been flat-out devising - what they believe to be - the optimum configuration of aerodynamics, hydrodynamics, control systems and hydraulics. These areas of the boat design are all inextricably linked - with electronics playing a significant role in this intricate dance of “trade offs”.

For example, a designer must choose how inherently stable to make their boat. Not stable enough and it will be tossed about too much - slowing it down; too stable and the boat will not be responsive enough to make rapid tacks and jibes - again slowing it down.

In modern fighter-jets design (eg: the F16, F22 and F35 aircraft), the trend has been to design the airframe inherently unstable and let the electronics and control systems keep the craft stable - known as fly by wire. Similarities can be drawn with the modern AC boats - where the hydrofoils can be designed with negative inherent stability. Overall boat stability can still be achieved however through the use of the electronic control systems constantly moving the control surfaces toward equilibrium. This design approach relies on considerable lines of computer code(!), but it does ultimately allow the boat to achieve the holy grail of both stability and high maneuverability. ¹³

“You’re very aware that you are now sailing a boat that wholly relies on computer code. But I guess when you’re sitting in an airplane at 40,000 feet, it’s exactly the same”

- Freddy Carr of Ineos Team UK ¹⁴

Foil and Cant System Control

Central to the AC75s ability to foil, is some brand new technology called the cant system [The word ‘cant’ being a nautical term synonymous with ‘tilting’]. As the boat swaps tacks, the cant system is activated, placing one hydrofoil in the water, and lifting the other one out, where its weight becomes ballast.

The system is a 48 V, 5 kWh ¹⁵ battery-driven, hydraulic power-unit that supplies the energy to lift and lower the immensely strong - and heavy (1385 kg each) cant arms and foils.

The [cant system] design is essentially an electronically controlled hydraulic system which controls the 40 tonne hydraulic cylinders which lift and lower the foil arm...

We have been working hard to effectively trim the PLC to minimise the oscillations through the arm which if exist when sailing will vibrate through the boat…”

- Emirates Team New Zealand Project Manager Peter ‘Brush’ Thomas ¹⁶

Caption 5: The foil cant system is an electronically controlled hydraulic system which raises/lowers the foil arms into the water. ¹⁷

Like the AC50s the crew are allowed precise control over rudder rake - which controls the pitch of the boat. A sailor can simply set a command that they want the foil to be at “X” position and a transducer can provide direct feedback on the actual position to the electronics which in turn drives the foil to that position.

“In [2013 America’s Cup Campaign] San Francisco, we were pressing a button that said open the valve for .03 of a second and see what happens. The problem is that if you do that at 20 knots, it’s a different result than at 40 knots, so the boats were hard to sail because of that. We are [now] in the modern world of control systems…” ¹⁸

- Nick Holroyd, lead designer and technical director with SoftBank Team Japan

Unlike the previous catamarans however, the AC75s have computer-adjustable flaps on the foils (like those on an airplane wing), run by a complex array of electronics and hydraulics. These have been custom-designed by each syndicate and these systems will be software tuned to deliver various response speeds and increments, depending on wind and wave conditions. The flaps move constantly to keep the boat trimmed in smooth flight.

Caption 6: New to the 36th America’s Cup is the trailing edge flaps. The electronically controlled flaps move constantly to keep the boat trimmed in smooth flight. ¹⁹

What About The Crew?

Sitting alongside all the impressive technology on the AC75s, is the impressive world-class sailing crew. After all, it is still a sporting contest - albeit sometimes a polarising one.

There are those who believe that the America’s Cup has gone too far with technology and that sailing should return to a more simple form. At the other end of the spectrum, there are those who absolutely believe in the relentless adoption of technology - to the point where the boats can perhaps sail themselves.

Like Formula One, it can be difficult to strike the right balance between technology and sport. For now the AC75 crew of eleven - with the aid of electronics and wearable technologies - still have plenty of work to do, perhaps more so than ever before.

Generally speaking the crew of an AC75 will resemble something similar to the below configuration:

The Helmsman x1
Has control of the wheel and various things, but his primary role is to steer the boat around the course.
The Flight Controller x1
The Flight Controller - a relatively new role in the history of America’s Cup sailing - monitors an array of electronic data and controls the foil cant, foil flap and rudder rake - to ensure the boat is “flying” in the optimum zone.
The Trimmer x1
Primarily controls the shape of the mast and the twin-skinned mainsail - to ensure the wing is providing maximum power.
The Grinders x8 (one of whom may also serve as a Tactician)
While essentially everything below the waterline is powered by battery, everything above the waterline is still powered by the humans (aka Grinders) - who can output an insane 300 W average power ²⁰ and much more in peaks. They essentially pump oil in order to accumulate enough hydraulic pressure to enable the move and trim the sails..

“We have power meters on our handles and our live heart rate at all times, and … [we are] very aware of where we can sit in terms of heart rate zones, and what that means in terms of the power going into the handles.”

- Freddy Carr of Ineos Team UK ²¹

Final Note

The team at Beta Solutions will be following the progress of the Americas Cup, and naturally (being based in New Zealand) we will be wishing Emirates Team NZ all the best to defend the cup!

We appreciate the R&D complexity and admire the vision of the rule makers to put forth such an audacious boat concept, pushing the boundaries of what is possible with electronics and technology.

If you perhaps have an electronic product idea, be it something that is pushing the boundaries of what is possible with electronics and technology or something uncomplicated, please do get in touch - we’d love to chat and discuss your options.

References:

Quote retrieved from https://www.yachtingworld.com/americas-cup/ineos-team-uk-grinder-david-freddie-carr-interview-ac75-sailing
Retrieved from https://www.youtube.com/watch?v=VQUl_hf6yo8
Retrieved from https://www.latimes.com/archives/la-xpm-1986-12-14-sp-3142-story.html
Retrieved from https://www.americascup.com/the-technology
Hydrodynamic Force, retrieved from https://en.wikipedia.org/wiki/Aerodynamic_force
Retrieved from https://www.yachtingworld.com/americas-cup/americas-cup-news/podcast-secrets-behind-race-fastest-americas-cup-foils-107154
Retrieved from nzherald.co.nz/sport/americas-cup-ashbys-flying-on-wingsail-and-his-special-xbox-console-toy/W77SRQWL3CNQTVAY2ROEY7YSPQ/
Retrieved from https://www.yachtingworld.com/americas-cup/design-in-detail-exactly-what-made-emirates-team-new-zealand-so-fast-109101
Retrieved from https://www.yachtingworld.com/americas-cup/ineos-team-uk-grinder-david-freddie-carr-interview-ac75-sailing-127246
Ineos Team image from https://www.yachtingworld.com/americas-cup/ineos-team-uk-grinder-david-freddie-carr-interview-ac75-sailing-127246
Retrieved from https://www.americascup.com/en/official/the-class-rule
Retrieved from https://www.sailmagazine.com/racing/auckland-analysis-of-the-36th-americas-cup
Retrieved from https://www.sailingworld.com/tech-behind-americas-cup/
Retrieved from https://www.yachtingworld.com/americas-cup/ineos-team-uk-grinder-david-freddie-carr-interview-ac75-sailing-127246
Retrieved from https://www.sailingscuttlebutt.com/2019/12/11/americas-cup-how-techie-are-you/#:~:text=It%20is%20300W%20per%20grinder,but%20much%20more%20in%20peaks .
Retrieved from https://emirates-team-new-zealand.americascup.com/en/news/349_A-look-inside-the-AC75-foil-cant-system.html
Cation 5 image ref https://www.youtube.com/watch?v=VQUl_hf6yo8
Retrieved from https://www.sailingworld.com/tech-behind-americas-cup/
Retrieved from https://www.youtube.com/watch?v=VQUl_hf6yo8
Retrieved from https://www.sailingscuttlebutt.com/2019/12/11/americas-cup-how-techie-are-you/#:~:text=It%20is%20300W%20per%20grinder,but%20much%20more%20in%20peaks .
Retrieved from https://www.yachtingworld.com/americas-cup/ineos-team-uk-grinder-david-freddie-carr-interview-ac75-sailing-127246
Banner images (first and last images) retrieved from https://www.americascup.com/en/news

Key Trends in Deployment of Visual Artificial Intelligence

Mon, 09 Nov 2020 19:00:02 GMT

Author: Aaron Fulton, Systems Developer

One of our team members recently attended the Edge AI and Vision Alliance Summit to learn about the recent developments and trends in Visual Artificial Intelligence (AI). We thought we would share some of these learnings with you.

Computer vision has come a long way in the last decade. Traditionally hand-designed algorithms were used for vision applications, but advances in Neural Network (NN) algorithms and the availability of large data sets to train them, have shifted the balance towards AI-based vision solutions.

Some of the key trends in Embedded vision:

Algorithms

Neural networks are becoming dominant. In 2016, roughly 38% of vision applications used neural networks whereas, in 2019, over 80% of vision applications used them.
The consolidation of Neural Network models is a key trend. Training a high-accuracy model from scratch takes a massive amount of data and effort, however, in recent years, the trend is towards using pre-trained 'off the shelf' models much more. These off-the-shelf models can be used to speed up development by:
Using them as out-of-the-box.
Refining them to focus on a specific application using additional training data (transfer learning).
Using them to detect ‘features’ (eg. edges and shapes) from images then using other classification techniques for the full classification.
TensorFlow is still the most popular framework for developing NNs ¹ , however, there is beginning to be a lot more variation in frameworks used.

Industry Trends

Typically neural networks models have been very large, computationally expensive and power-hungry to compute, however, with the rise of Industry 4.0, the key application for AI-based vision technologies are now best deployed 'at the edge' (near the source of the information).
Computer vision at the edge is important from a performance, latency, bandwidth, privacy, reliability, and economic perspective.
The use of 3D perception in vision-based systems is also increasing rapidly. 3D perception involves using time-of-flight cameras or lidar-based systems to add depth information.

Hardware

Dozens of specialised, price-optimised, chipsets are now available which are capable of running optimised versions of state-of-the-art NNs at several frames per second in only a few milliwatts of power. The innovation of specialised chipsets is a trend which will continue for some time.
While CPUs ( central processing units) and GPUs ( graphics processing units) remain the dominant method of currently deploying Neural Networks, the trend is towards using dedicated deep learning processors and digital signal processors.

Conclusion

The use of AI based vision sensing will continue to be a trend well into the future. There is a lot of innovation happening in this space at the moment, driven by Industry 4.0 trends. Running complex artificial intelligence algorithms is now very possible in embedded devices.

At Beta Solutions, our experienced embedded electronics engineers have the skill and proven track record to use the best suited firmware design methodology for our clients' products. You can get in touch with us to discuss any idea you have in mind via our contact page or give us a call .

References:

bit.ly/DeveloperSurveyWhitePaper2020

Supporting Information:

Edge AI and Vision Alliance
bit.ly/VisionIndustryMap
Banner Image retrieved from https://online.stanford.edu/courses/cs231n-convolutional-neural-networks-visual-recognition
Image in blog by https://pixabay.com/?utm_source=link-attribution&utm_medium=referral&utm_campaign=image&utm_content=4791810 "Pixabay>

Artificial Neural Networks on the Edge

Thu, 24 Sep 2020 01:11:08 GMT

Author: Dr Matthew van der Werff, CTO and Aaron Fulton, Systems Developer

Introduction

As Artificial Intelligence (AI) becomes more prolific, we at Beta Solutions, are increasingly investigating how these technological advancements can unlock new and unique features for our clients' projects.

This article takes an overview of Artificial Intelligence and explores it’s fundamental building blocks - namely the Neural Network. Finally, we look at the suitability and considerations of operating Neural Networks on the Edge.

Don’t know what the ‘Edge’ or a ‘Neural Network’ is? Then read on …

Background of Neural Networks

You have probably heard the terms, Artificial Neural Networks (ANN), Convolutional Neural Network (CNN), Artificial Intelligence, or Machine Learning. These are synonymous terms, all based upon the concept of training a computer algorithm to operate similar to the way your brain learns. Artificial Neural Networks are now beginning to be used everywhere - from algorithms that recognise the words you say (e.g: “Hey Siri”), to Tesla's driverless cars, to which ads to show you, to some quite Orwellian types of camera surveillance that are in use in some countries. You may be forgiven to think that neural nets are a fairly recent development but in actual fact research in neural nets has been going on for a long time, back to the late 1940s ^[1] with the term ‘Artificial Intelligence’ (AI) coined by John MacCarthy who, in 1959, co-founded MIT’s Artificial Intelligence Lab ^[2] .

In regards to AI research, there have been a number of periods of boom and bust over the years. One of the reasons for the limited use of AI back then was the limited processing power that was available. Indeed, AI networks require a huge amount of processing power to train them, and a moderate amount of computing power to operate. Now, with large cloud farms of parallel computing and arrays of GPUs easily available, practical large AI networks can be trained in a relatively short amount of time - not to mention some amazing advances in ANN training algorithm development. Also, given that your mobile phones have more and more processing power, it is relatively simple for them to run smaller neural networks to process things such as voice. Where AI was once a buzz word for researchers and science fiction for writers alike, now AI is just another tool on the belt for engineers and programmers to use for solving classification problems such as voice, image analysis, and other problems that can be labelled.

However, there are some caveats to what neural networks can be used for. As much as I enjoy a good sci-fiction book (such as Issac Asimov’s 1950’s Robot series, where Robots ‘General AI’ are able to self learn and apply their processing to many different problems) we are still very much in the ‘Narrow AI’ era where AI can solve very specific ‘narrow’ problems such as image identification or speech recognition. With a ‘Narrow AI’ a large amount of training data is required to classify your problem and your solution is only ever as good as your data .

How does Artificial Neural Network work?

An Artificial Neural Network takes inspiration from how our brains work. A child sees a cat and her Dad tells her it is a cat - which trains the child that this ‘is a cat'; the child then sees a dog and her Mum tells her it is a dog - which trains the child that it is 'not a cat’. See enough cats and the child becomes quite proficient at identifying cats and ‘not cats’!

A simplified Neural Network is shown in the diagram below. It involves the input of data (such as camera data), which in turn is presented to a number of hidden weighted nodes (lots of nodes and layers), then after considerable maths (multiplication and addition), a weighted value is outputted - which can be used to identify the object of interest.

Figure 1: Artificial Neural Network ^[3]

The Convolutional Neural Network (CNN) ^[19] also should be considered. These are typically used for image or speech processing. These allow for spatial recognition such as where ones eyes are in relation to ones nose. etc. - which can help significantly with image recognition.

Figure 2: Convolutional Neural Network ^[21]

The basic/simplified steps to train a neural network are:

Capture data,
Label data,
Train NN,
Run NN.

If you were trying to make an ANN that could identify cats, then you would need to capture data of a large number of images of cats that you as a human has labelled as a cat, and an equally large number of images that you know are not a cat. The ANN is then trained by an optimising algorithm (which takes a lot of processing power) that adjusts the weights of the hidden nodes until the output gives a high number if ‘it is a cat’ and a low number if it is ‘not a cat’. Now you can run the ANN which - when presented with a camera image of a cat - will give a probability that the image is a cat. But try and remember the results are only as good as your data!

NB: There is a lot more that could be said here, but this is the basic concept.

One main disadvantage of AI over traditional algorithm methods is that it is very much a black box design. You never know quite what is going on inside the model:

Data in → ‘Magic’ multiplication and addition → Data out!

Also note that an ANN will never be 100% correct, there will always be statistical outliers, much like we as humans get things wrong but we have the advantage of being able to problem solve or know that we do not know what it is. In most cases however, good enough is good enough.

Artificial Neural Network Tools

There are a large number of tools out there now that can be used to train networks from the high-level API cloud-based or control with python based scripts for image classifiers (such as Google Cloud AutoML[11], Microsoft Azure computer-vision ^[12] , and Amazon Rekognition ^[13] ) where you upload images and classify them into folders, then call a simple web API to find out what is in your picture. These tools typically use a Convolutional Neural Net (CNN) and various other magic sauces under the hood. These are relatively easy to use and little background knowledge in computer vision or CNNs are required to use them. However, these solutions do limit you to using that specific provider which incidentally are usually not compatible with portable / “on the edge” devices.

The next step in complexity and flexibility is to train your own ANN. The main tools for this are open-source, with the main tool being Google's developed Tensor Flow ^[14] , and Berkeley’s Caffe ^[15] . TensorFlow is probably the easiest tool to use with a Python interface and is relatively easy to get up and going, with a large number of tutorials on the web and a number of different use cases ^[16] .

To improve speed and decrease the memory requirements on embedded devices CNN models are usually converted to a TensorFlow Lite model. TensorFlow Lite typically quantizes the model - this means all the weights are converted from floating-point (32-bit) to 8-bit integer. This, in turn, means that only 8-bit calculations are required, which essentially enables the NN to operate on much smaller/embedded processors. Although quantization will reduce the accuracy of a model by roughly 3% (depends on each model) the benefits are considerable - specifically the huge memory storage savings (~4x) and significant energy savings ^[7][8] . Sometimes pruning can also be run, which removes weights that have little effect.

Artificial Neural Network Training

Training ANN can take an immense amount of cloud processing time and requires a huge number of labelled images and time to classify. Recently it has been discovered that a way to decrease the cloud computing time is to use an already trained CNN and re-train on a new set of specific images - which is called “transfer learning”.

There are two methods of transfer learning.

The first method involves fine-tuning the pre-trained model. This can be susceptible to over-fitting (for example, operating well for the training images but not new ones) and still takes time and a large number of additional training images.
The second method involves “feature-transfer”. The pre-trained CNN is exposed to the training set for the specialised task, and information is then extracted from the intermediate layers of the CNN, capturing low-to-high level image features (such as the number of legs on an insect). The feature information is then used to train a simpler machine learning system, such as a support vector machine (SVM), on the more specialised task. Feature transfer in combination with SVMs can work better than fine-tuning when the specialised task is different from the original task. It is computationally more efficient, works for smaller image sets, and SVMs are less susceptible to over-fitting when working with imbalanced data sets, that is, data sets where some categories are represented by very few examples.

Most of the pre-trained networks used for feature transfer come from the ImageNet challenges in which a number of competitions were run, in which large numbers of labelled images were made available for the best CNN to identify the images the most accurately. These are quite big networks. The winning networks have < 10% error which is on par with the human-level performance ^[19] . The most common of these networks are:

AlexNet
GoogleNet
VGGNet
ResNet

Embedded Processors

A look at the Edge:

‘The Edge’ is a term which basically means processing data on the embedded device, rather than in the cloud. This can have a number of advantages. For example, if you are running a cellular device with a camera on the device, sending every image back to the cloud server will use considerable data - not to mention the cost in cloud computing time. However, if you can run an ANN algorithm on the embedded device (a.k.a The ‘Edge’), then only the data you are interested in needs to be transmitted to the cloud server. The downside of this is that the embedded device needs to have enough computational power to run the ANN. This is especially important for video applications where it is important to have high frame rate identification of objects.

Can ANNs operate on the Edge?

The short answer is “ YES ”. Advancements in embedded processing power, low power processors and more efficient Neural Networks mean many ANNs can now be run on the Edge.

Of course, the longer answer is “ IT DEPENDS ”. There are a wide range of factors to consider - including restrictions on Power, Memory, Processing Power, Overall Price etc. All of these factors must be balanced - when considering whether or not to operate an ANN on the Edge.

For a more detailed and technical explanation of various development approaches available to the engineer - read on.

Embedded System on Chips (SoC’s)

There are more and more embedded System on Chips (SoC’s) emerging that are capable of ANN processing at the Edge ^[5] :

Raspberry Pi
Movidius NCS
NCS2
Nvidia Jetson
Coral USB Accelerator
(A plug-in USB peripheral that could be used in conjunction with (say) a Raspberry Pi)

Figure 3: Running neural networks on embedded systems. ^[22]

Benchmarking has shown that these ANN accelerators can get pretty close to a high-performance PC performance ^[5] . However, these embedded devices still use a lot of power and so are likely not suitable for battery power devices. They can also be relatively expensive, so will be only suitable for some embedded applications.

The Raspberry Pi 4 is probably the most affordable, most accessible way to get started with embedded machine learning right now. One can use it on its own with TensorFlow Lite for competitive performance, or with the Coral USB Accelerator from Google for ‘best in class’ performance. However, the idle current draw was between 410 and 600mA (~4 hours on a pair of AA batteries) with peak current 860 – 1430mA @ 5V (~3 hours on a pair of AA batteries).

For a commercial solution similar to Raspberry Pi, Nvidia's Jetson device could also be a good option. Google’s Coral SoC can perform at 4TOPS at 8-bit, at 2TOPS/Watt ^[17] , but this requires PCIe or USB 2.0 to interface to it. Therefore while low/lower power, it still requires a higher-end Linux type embedded device with large amounts of memory to interface to it.

Another approach for small ANN’s would be to run a low power processor such as a commonly available ARM Cortex-M4. These can allow for ~1uA sleep current (+10 years on AA batteries) and can have very efficient uA/MHz power usage. These usually are fairly limited in FLASH and RAM with the largest around 512kb of RAM and up to 1Mb of internal flash which limits the size of the ANN that can be run. However, the ANN can be stored in an external SPI flash. This approach may work for smaller ANN’s or if low operation speed is acceptable. The ARM Cortex-M CMISS-NN software framework can also provide up to a 5x boost compared to not using this method [11]. The basic approach to running with these MCUs is to trigger a wake-up and run the ANN when required. Another example of a Cortex M4 framework is ST Micro’s ‘STM32Cube.AI’ framework ^[18] . This helps to simplify the process to get your low power embedded device running a Neural Network. This supports a number of different processors in the STM32 lineup.

A low power approach is to use one of the new dedicated CNN (Convolutional Neural Net) SoC’s that are now coming out such as the Chinese chip Kendryte K210 [6]. This SoC is a “low end” MCU for ANN but packed an impressive punch of processing power. It quotes 1-Tera Operations per second (TOPS) with an 8-bit quantized network at around 0.3Watts. It can run a small CNN ~5-6MB at QVGA@60fps or VGA@30fps. Alternatively, if you have a large CNN you can store in external flash and run from there (though at a slower rate). It is a very affordable option compared to most SOCs, and much lower power. The processor has two 64-bit RISC-V CPUs each with an independent Floating Point Unit for additional processing. The maximum (quantized) network size is 5-6MB when using the on-board memory. Typical power consumption is < 1W and 200-300mA ^[9] .

One way to accelerate CNNs on the Edge is to use FPGA’s and ASICS [5] [13]. FPGA’s typically require the network to be quantized which reduced the accuracy by roughly 3%. Most research done with FPGA’s only uses small CNNs. FPGA implementations can give a 15% speedup compared to an MCU solution alone ^[7] . FPGAs can be complex beasts and power supply and power requirements are not insignificant.

There are a large number of modules out there for doing AI on the Edge with some found at GitHub repo collection of Edge AI devices ^[10] , or for vision-based solutions see edge-ai-vision.com ^[4] . More and more processors are including CNN processing accelerators.

Figure 4: List of Embedded Vision AI ICs. ^[4]

Conclusion

Operating Artificial Neural Networks on the Edge is now possible even with fairly low power MCUs. This unlocks exciting opportunities for AI based applications that were previously not possible with non-AI algorithms - such as image analysis and classification, etc.

However, there are still a wide range of factors and trade-offs to consider if developing an ‘AI on the Edge’ type product. Power consumption, processing power limitations, and overall cost are three important such factors.

Do you have a product idea that might benefit from Embedded Artificial Intelligence? Contact or call Beta Solutions to see how we can help you!

References:

https://en.wikipedia.org/wiki/History_of_artificial_neural_networks
Isaacson, Walter. The Innovators: How a Group of Hackers, Geniuses, and Geeks Created the Digital Revolution (p. 203). Simon & Schuster. Kindle Edition.
By Glosser.ca - Own work, Derivative of File: Artificial neural network.svg, CC BY-SA 3.0, https://commons.wikimedia.org/w/index.php?curid=24913461
https://www.edge-ai-vision.com/resources/industrymap/
https://www.hackster.io/news/the-big-benchmarking-roundup-a561fbfe8719
https://canaan.io/product/kendryteai
TinyCNN: A Tiny Modular CNN Accelerator for Embedded FPGA
William Dally. High-Performance Hardware for Machine Learning. Tutorial, NIPS, 2015
K210 Datasheet
https://github.com/crespum/edge-ai - Connect to preview
https://cloud.google.com/vision
https://azure.microsoft.com/en-us/services/cognitive-services/computer-vision/
https://aws.amazon.com/rekognition/?p=tile&blog-cards.sort-by=item.additionalFields.createdDate&blog-cards.sort-order=desc
https://www.tensorflow.org/
https://caffe.berkeleyvision.org/
https://www.tensorflow.org/learn
https://coral.ai/products/accelerator-module/
https://www.st.com/content/st_com/en/stm32-ann.html
CNN Architectures: LeNet, AlexNet, VGG, GoogLeNet, ResNet and more…
https://www.youtube.com/watch?v=K_BHmztRTpA
https://www.analyticsvidhya.com/blog/2020/02/cnn-vs-rnn-vs-mlp-analyzing-3-types-of-neural-networks-in-deep-learning/
https://www.quantmetry.com/neural-networks-embedded-systems/

5G for IOT?

Wed, 29 Jul 2020 22:13:44 GMT

Author: Jason Cleland, Senior Engineer and Project Manage r

5G Technology Today:

The 5G technology that has been launched in New Zealand already is a great first step. One of the immediate benefits of this technology (over existing 4G technology) is a lower latency (or ‘delay') connection. Less delay means websites will be snappier and video calls will feel more 'lifelike’. More importantly, low latency connections will unlock new disruptive technologies - such as real-time navigation of autonomous vehicles; ultra-realistic online VR gaming/simulations; and many more yet-to-be-developed ideas.

Initially, one of the biggest applications for 5G in New Zealand is 'fixed wireless' according to Vodafone's technology director, Tony Baird. "... the main investment case for 5G is fixed-wireless and getting ourselves ready to do a lot more wireless [internet] to the home and wireless to the business," - Tony Baird.

The 5G networks in New Zealand will first operate in the 3.5Ghz band. Spectrum within the 3.5GHz band is limited to existing rights holders and temporary licenses but the entire band will go up for auction in 2022.

5G can also operate in mm-wave bands (>10GHz). Mm-wave 5G can provide ultrafast short-range connections suitable for use in dense city areas. The government body who manages the spectrum, Radio Spectrum Management (RSM), say that mm-wave bands will play an important role in providing capacity for 5G networks. However, they are delaying the allocation of mm-wave bands until there is more certainty in which bands will become mainstream worldwide. It is likely New Zealand consumers won't see a commercial mm-wave 5G network for the medium term (3+ years).

For the first few years, 5G will likely just add speed and capacity to existing 4G networks, which perhaps isn’t tremendously exciting. Thankfully, 5G has been architected as an evolutionary technology and new, more exciting, features will be added as the technology matures. Read on …

Future 5G Technology for IoT:

IoT or Internet-of-Things devices generally have different needs to those of smartphones. 5G could supply them with massively fast speeds but this also requires significant battery depleting power. This is acceptable for smartphones as their batteries only really need to last for a day or two on a charge. But battery-powered IoT devices want to last many months or even several years! During this time they’re mostly asleep and only wake up occasionally to spit out a little bit of data.

At the present time, 5G can’t really support these kinds of low-power/long-life devices, but the future looks promising.

3GPP is the group who are defining the standards from which our NZ 5G networks are being built. ‘Release 15’ is the first and only 5G standard document which is ‘Frozen' and ready for deployment. Future 5G releases, 16 and 17, are available to view but might still be revised/tweaked. It’s these two documents which we can use to peer into the future of 5G. Refer to figure 1 below for the expected timeline of these standards.

Figure 1: 3GPP RAN Milestones ¹ .

The future 5G feature which excites us the most is the likely introduction of support for Massive Machine-Type Communications (mMTC) from low-power and low-complexity devices (I.E. Battery sensors and wearables) in Release 17. This feature will enable the following types of devices to be implemented:

Industrial wireless sensors with low latency (5-10ms) and medium data rate (<2 Mbps)
Medium to high data rate (2-25 Mbps) video transmission
High data-rate wearables (5-50 Mbps) with long battery life (1-2 weeks)

It’s important to note that the purpose of 5G mMTC support is not to replace the likes of NZ’s existing LTE-CatM1, LoRaWAN, and NB-IoT networks. These networks are meant for super-low-power devices where the latency and data-rate of their communications is not critical. 5G Release 17 (mMTC) is aimed to fill-the-gap for devices which require:

Lower device complexity and reduced energy consumption compared what can be provided by 5G Release’s 15/16; AND
Have higher requirements in terms of data rates and latency compared to what can be provided with LTE-CATM1, LoRaWAN, and NB-IoT protocols.

Conclusion

The first release of 5G is here today but it’s not terribly interesting from an IOT perspective. However, the future of 5G is something we are excited about.

Release 17 is not scheduled to be frozen until at least early 2022 so it will still be a while yet before it is implemented in New Zealand. Nevertheless, Beta Solutions will be prepared and waiting so that we might best support our client’s future innovations.

If you have an idea, or need advice regarding hardware or firmware design techniques for low power wireless devices or to discuss 5G - feel free to get in touch with the team at Beta Solutions!

References

Figure 1: 3GPP RAN milestones. Retrieved from https://www.3gpp.org/release-17
https://www.ericsson.com/en/reports-and-papers/ericsson-technology-review/articles/5g-nr-evolution
https://www.5gamericas.org/wp-content/uploads/2020/01/5G-Evolution-3GPP-R16-R17-FINAL.pdf
https://www.3gpp.org/specifications/67-releases
Banner image vectors are retrieved from / Vectors by Vecteezy
IoT image by jeferrb from Pixabay
Final image retrieved from https://pxhere.com/en/photo/1575607

A Closer Look at the MISRA C Firmware Coding Standard

Wed, 17 Jun 2020 22:13:43 GMT

Author: Jonathan Kapene, Firmware Engineer

B is for … Bug

Have you ever been working on your computer and suddenly your mouse stops moving? Maybe you’ve been on your phone and suddenly your app crashes for no apparent reason. I’m sure like most people you’ve had the experience of something not working only to find that turning it off and on again miraculously solves the issue. If this sounds familiar there’s a high chance you’ve encountered a “bug“, which is just another way to say “a defect in the firmware”.

As firmware contains varying degrees of complexity, writing completely bug-free code is usually not a straightforward exercise. In any case, these intangible bugs can be problematic.

Most firmware bugs probably cost you no more than a couple of minutes of time and a slight loss in productivity. However, bugs aren't just limited to your everyday PCs, TV remotes or smartphones. Complex Embedded Systems are becoming increasingly more and more prevalent in safety-critical industries such as Healthcare, Automotive, and Aerospace - and in these applications firmware bugs can result in injury or loss of lives. For example, the Patriots missile bug resulted in the terrible death of 28 people and injuring 96 (more on that later).

Needless to say firmware quality is important and there are many ways firmware engineers can mitigate bugs in the code they write. One way is to use test-driven development which I encourage you to read about in one of Beta Solutions' previous blog posts here. Another way, and the topic of this blog, is to follow a coding standard.

The particular coding standard which is the topic of this blog is called is MISRA C.

MISRA stands for Motor Industry Software Reliability Association which is an organisation founded by a number of vehicle manufacturers, including Ford and Jaguar. As the name suggests, the organisation was formed to help make the software (or firmware) developed in the motor industry more reliable. MISRA C is the coding standard MISRA came up with. Although this was intended to be used in the motor industry many other fields have adopted it because of just how reliable it made their firmware. Companies like Boeing, NVIDIA, and Apple all utilise MISRA C in some way. Continue reading to learn how MISRA was started, what MISRA C is, and why and where it is used.

MISRA Background

MISRA was founded off the back of a government-funded program in the UK when it became concerned with safety-critical electronics. MISRA was the working group responsible for the motor vehicle industry, and consisted of vehicle manufacturers, component suppliers and engineering consultancies. Their mission statement was:

“We provide world-leading, best practice guidelines for the safe and secure application of both embedded control systems and standalone software.“ - MISRA

This all started in the early 90s and by 1994 they would release “Development Guidelines for Vehicle Based Software“ - a document which outlined a set of guidelines for the creation of embedded firmware in the motor vehicle industry. Government funding stopped the following year but MISRA continues to operate on a self-funding basis with a number of organisations contributing. MISRA was, and still is, a collaboration of the following companies who are or have contributed to MISRA ^[1] :

Ford Motor Company
Jaguar
Lotus
Land-Rover
Rolls Royce and Associates

The original documents they published have since morphed into what's now known as MISRA C, which has three revisions. MISRA 1998, MISRA C 2004 and MISRA C 2012 - all released in their respective years.

MISRA representatives at the MISRA C 2004 release.

What is MISRA C?

In short, MISRA C is a set of firmware coding rules and guidelines which, if followed, are intended to help mitigate firmware bugs.

The C in MISRA C simply represents the programming language it's intended for ie: 'C'. The C language is generally great for embedded systems. It has low overhead, it's very efficient, and it's portable to other micro-controllers with little modifications. That said, there are some downsides to C.

For example, if you’ve had C programming experience you may know of the functions Malloc and Calloc, which are used for allocating memory on the heap (dynamic memory allocation). However, the use of such functions has the potential to lead to memory leaks which, in turn, can lead to undefined firmware behaviour (aka: "bugs").
MISRA C fixes this problem with directive 4.12. which simply states: “Dynamic memory allocation shall not be used“. This is a simple yet very effective rule which can dramatically improve code reliability.

Let's take a quick step back to define what a directive is and the different classifications/categories each guideline can have. Firstly a directive, in this context, is simply something that the code must adhere to if it is to be called MISRA C compliant. So directive 4.12 mentioned above states that for code to be MISRA C compliant it must not contain the use of any dynamic memory allocation. However, MISRA recognised that on the rare occasion it might be impractical to follow some guidelines. This is why each guideline is categorised into three different categories:

Mandatory: A guideline in the mandatory category must be followed for the code to be MISRA C compliant. Deviation from said guideline is not permitted.
Required: Code should conform to these guidelines. If it must deviate then formal documentation is required.
Advisory: Code should conform to these guidelines however formal documentation is not required if a deviation from this is required. MISRA state that these should not be ignored, but rather should be followed as is reasonably practical.

The example directive 4.12 has the category “required”. This means if a firmware engineer has carefully considered the use of dynamic memory allocation then he/she can still call the code compliant if it is well documented. Documentation is very important and serves two purposes:

It is very helpful for when other engineers are reviewing the code or performing code maintenance.
It makes the engineer fully consider the decision to deviate from MISRA C.

Why use MISRA C?

As stated above MISRA helps to eliminate firmware bugs and thus improves the reliability of the system.

Code reliability is especially important in systems that are considered ‘critical’ - whereby a failure of that system could lead to: (i) Injury/death of a person(s) (“Safety Critical”), (ii) Breach of sensitive information (“Security Critical”), (iii) Prevention of accomplishing an important Goal (“Mission Critical”).

Needless to say, in a critical system the code reliability should be the highest priority.

How dangerous can firmware bugs be? As an example, we refer back to the aforementioned Patriots missile bug. The Patriot air and missile defence system was used by the US and its allies during the gulf war. The Patriots job was to detect incoming Scud missiles from the enemy and send its own missile to intercept them mid-air before it landed on any allied targets. These were set up around the allies' bases and outposts. The system was designed as a mobile/portable system that would never have a long continuous runtime. (This happens to be quite relevant to the firmware bug.)

The patriot uses radar to scan for incoming threats from up to 80 km away, track that threat, and guide a missile to intercept it. Essentially, a firmware bug caused a miscalculation and the enemy missile was never intercepted. It hit its intended target, a U.S. Army barracks in Dhahran, Saudi Arabia consequently killing 28 people and injuring 96 others.

The miscalculation was caused by a rounding error. Namely, the clock would gradually get more and more out of time the longer the patriot had been running (This is known as clock drift). After about 8 hours running continuously, it would be ineffective. This particular patriot missle was operational for 100 hours when the incident occurred.

Please note: The point overall I am making here with this example is not to discredit the firmware engineers who wrote the patriot missile code, but simply to illustrate to the reader how important firmware quality can be and the consequences of unreliable firmware.

Incidentally, MISRA does address such clock drift issues by stating in directive 4.1 that “Run time errors shall be minimized”, which essentially forces the firmware designer to consider such effects during the code's development.

The Patriot missile sytem and Patriot surface-to-air missile system. © REUTERS/Osman Orsal

MISRA C Recommendations
If a program is to be considered to comply with MISRA C, it is shown using a compliance matrix consisting of all the guidelines and how each is checked. MISRA C recommends the use of a static analysis tool to verify most of these. A static analysis tool is simply a program that automatically checks the code for violations of rules. A firmware developer would typically use it right after they have written a block of code and certainly before the code is released out into the field.

There are many different static analysis tools for embedded systems that contain their own unique features but most support checking for MISRA C.

Where and When is MISRA C used?

Though it was intended to be used in the automotive industry MISRA C has been adopted by a wide array of other safety-critical industries including medical, aerospace, and military to name a few. This is a credit to how well MISRA C achieves its intention of reducing firmware bugs and improving the overall quality of firmware.

While the advantages of following MISRA-C are clear, perhaps the main disadvantage of MISRA C is the increased development cost - due to:

Using a restricted set of the C programming language, meaning that firmware engineers have less tools to write the code.; and
Documentation of every deviation.

Therefore prior to the firmware development, one must always consider the question: “What are consequences of a system/product failure due to a potential firmware bug”?

Perhaps one can live with those consequences - perhaps one can’t. The key thing is to be aware of the implications of firmware bugs, and to develop your firmware accordingly. Nevertheless, more often than not, in the long run, it is cheaper and more time effective, to invest in writing quality code from the outset - rather than patching bugs more months (or years) down the track.

For non-critical systems, one could consider using the “Barr Group C” standard which is compatible with MISRA but not as strict. Beta Solutions Ltd, will often use the Barr Group C” standard for non-safety critical projects.

References:

MISRA. Who we are. Retrieved from https://www.misra.org.uk/MISRAHome/Whoarewe/tabid/67/Default.aspx
Image of MISRA C Working Group, at the official launch of MISRA-C:2004. Retrieved from http://www.misra.or.kr/misra_c_team.html
Development Guidelines for Vehicle Based Software. Retrieved from https://ia800108.us.archive.org/26/items/misradevelopmentguidelines/misra_dev_guidelines.pdf
MISRA-C:2004 Guidelines for the use of the C language in critical systems. Retrieved from http://caxapa.ru/thumbs/468328/misra-c-2004.pdf
The Patriot missile system. Retrieved from https://www.ainonline.com/aviation-news/defense/2019-06-12/big-claims-big-cost-surface-air-missile-systems
Patriot surface-to-air missile system. © REUTERS/Osman Orsal. Retrieved from https://tass.com/world/1123009
Banner Image retrieved from https://www.pikrepo.com/fjfjq/black-porsche-911-on-road-during-night-time

Agritech Trends

Wed, 20 May 2020 22:13:43 GMT

Author: Brenda Wormgoor, Marketing and Operational Manager

According to a United Nations Report the world's population is expected to reach 9.7 billion people in 20501 - and to feed that population, crop production would need to double2 . With a growing global population and the effects of climate change, farmers are turning to technology for increased productivity, increased yield, improved harvest quality, and sustainability. Enter: Agriculture 4.0 - artificial intelligence, highly automated technology and agricultural robotics. With the goals of optimised processes and sustainability, digitisation is no longer a foreign concept for framers.

The New Zealand government has included Agritech (also known as AgTech) as a sector in its Industry Transformation Plan (ITP), and defines Agricultural Technology - Agritech, as follows:

The ‘agritech’ sector refers to manufacturing, biotech and digital-based technology companies that are creating product, service, IP and value chain solutions for the agriculture, horticulture, aquaculture, apiculture and fishing sectors, with the aim of improving yield, efficiency, profitability, sustainability, reliability, quality or adding any other kind of value. (Forestry is excluded because forestry and wood processing is the focus of another dedicated ITP)’. 3

Figure 1 - Agritech's Broad Applicability. ³

As you can see from the definition and Figure 1 above, Agritech spans over many industries. There is an increasing amount of technology (hardware and software) available to farmers to optimise output. Agritech hardware includes robotics, sensors, cameras, and drones whilst software enables the collection and interpretation of data in order to make real-time changes.

For the purpose of this blog, we are going to look at a few technology trends deemed to change farming in the future.

Agricultural Robotics

The Global Agriculture Robots Market is projected to reach USD 11.58 Billion by 2026 ⁴ , driven by labour shortages in developed countries, the global growing population, need for increased yield, and demand for better quality products. The fusion of rapid developing technology and the agriculture sector has resulted in innovative results with Agbots automating labour-intensive crop production processes and data collection for crop development and management. There are already many examples of highly automated robots with sensitive sensors and camera vision employed in harvesting, fruit picking, soil maintenance, etc. One such example is Root AI's Virgo tomato harvester that "see" ripe fruit, has a customised convolutional neural network to detect fruits of interest, move inside vines, and picks fruit without damage. ⁵

Figure 2 and 3: Root AI's Virgo Harvesting Robot.

Another example is from an innovative New Zealand company, Robotics Plus, who have launched two commercial innovations Āporo apple packer and an automatic log scaler. Below is another one of their projects, a robotic Kiwifruit Harvester:

Video: Robotic Kiwifruit Harvester - Robotics Plus. ⁶

Smart Sensors

With IoT-connected sensors placed near plants, farmers can capture and record environmental conditions, and then send the data back to the farm's data center, via a wireless connection, for analysis and action. In the quest for precision agriculture, a trend is for farmers to use more and more connected IoT sensors, like sensors on the ground to test soil moisture levels (to identify flooding, over-watering, or ground freezing) combined with IoT-enabled water and fertilizer delivery valves. Following are some examples of developments using smart sensors.

Solar powered with rechargeable batteries, ecoRobotix’s AVO autonomous weeding robot is using in scan-and-spray technology. Its multi-camera vision system detects weeds and sprays them once they pass under its spraying tool whilst LIDAR & ultra-sound sensors are used for obstacle and human detection. The fully autonomous navigation is based on vision and real-time kinematic (RTK) GPS positioning. Farmers can control and monitor the Avo robot through the ecoRobotix™ mobile app and define its missions via the desktop web app. ⁷

Figure 4. ecoRobotix’s AVO autonomous weeding robot. ⁷

Robotic scanning technology Smart-N, from Vantage NZ Ltd, is fertiliser application technology that use VIS/NIR (Visible/Near Infrared) sensors to identify urine patches on pastures. The system avoids wasting fertiliser over these patches and can also limit nitrification inhibitors to only these patches. By doing this the system both saves the farmer money and helps improve the environment.

This is not limited to crops, herds can also be monitored in real time. Using Internet-connected collars and tags farmers can have automated livestock tracking and use herd management software apps, like Breedmanager, that displays the herd based on breeding status.

Soil sampling, electromagnetic mapping and sensor technology assist farmers in determining the soil anatomy of different parts of a single field, and adjusting water or nutrients. SmartCore, from the start-up company Rogo Ag, is an autonomous robot that navigates fields and takes soil samples from specific locations. The machine is guided by obstacle-detection algorithms and GPS to take samples from the same spot each year, enabling farmers to track how their soil is evolving. Sampling and mapping tools collect soil data which farmers combine with a tractor’s, and even an aeroplane’s, internal software to determine where and how much water and fertiliser to apply to each patch of ground.

In the past, it has been a challenge to wirelessly connect each sensor to the internet - especially in very remote areas, where cellular coverage can be patchy. Advancements in Wireless mesh networking technology (eg: Bluetooth Mesh) provide a means for all sensors to “talk” to each other, as well as to the internet, and go a long way to solving issues with sub-standard wireless coverage. This mesh networking technology can also be integrated with (say) smart tractors and harvesters that can easily "carry" the network wherever they are working without requiring a pre-built networking infrastructure.

Semi/Autonomous Tractors and Vehicles

Driverless farm vehicles have been around for quite a while but needed a human in the cab in case of emergencies, such as when livestock wander into its path. Using GPS technology a farmer can programme the field’s coordinates into the tractor’s system and the machine can plough, or spread fertiliser autonomously. The obvious next stage of external sensors taking over the supervisory role is what many leading tractor manufacturers are developing (including John Deere, Case New Holland, AGCO, CLAAS, Same Deutz-Fahr and Kubota).

Figure 4. A robotic tractor cultivates a field alongside a tractor operated by a human, during a demonstration in Fukushima, Japan. ⁸

John Deere recently revealed their new concept tractor: GridCON - a fully electric, permanently cable-powered vehicle with full autonomous capabilities, developed in a consortium with B.A.U.M Consult GmbH and the TU Kaiserslautern (technical university). This driverless electric tractor has the capacity to connect with smart grid infrastructures so farmers who are already generating their own on-site renewable energy can plug directly into their own supply. ⁹ The GridCON is designed to remain plugged in to an external power source while in use. A drum attached to the tractor carries about 1000 m of cable (which can be adjusted based on field size if necessary) which is fed out and reeled in while guided by a robot arm to keep the operation friction free and at low load. An intelligent guidance system is also used to prevent the tractor from damaging the cable. A 100 kW electric motor powers the continuously variable transmission, and there is an additional outlet for other implements - powered by a 200 kW electric motor. The autonomous operating speed can reach 20km/h, and the tractor can also be guided manually using a remote control. ¹⁰

Figure 5. John Deere GridCON Electrical Tractor. ⁹

Small, autonomous self-propelled vehicles are seen to be the future of farming, however legislation (liability), cost and acceptance by farmers will determine the uptake rate.

Drones

With the drone, you can go from visual data to multispectral data (image data at specific frequencies), to thermal data, to hyperspectral data (from across the electromagnetic spectrum) all in one flight .
- Thomas Haun, Senior VP of Partnerships. PrecisionHawk (South Carolina-based commercial drone and data company).

Another trend is drones equipped with high-definition cameras, sensors (including Infrared (IR); Thermal; and NDVI (Normalized Difference Vegetation Index)), and image recognition capabilities that are used to monitor crops/pastures and increase efficiencies. They are also utilised for activities such as crop-spraying, real time management of livestock on distant paddocks, collecting crop and pasture quality information, whilst sending data via the cloud to a farmer’s device. In a recent study, drone operator PrecisionHawk found that farmers who used drone-based aerial intelligence instead of taking plot-based crop measurements by hand were able to collect data 2.5 times more efficiently and 25% more accurately, and the collection itself was more objective, repeatable, and standardised ⁸ .

Figure 6. Aerial Imagery and Seed Planting Drones ¹¹ .

Conclusion

What is clear from the above, is that one piece of technology isn’t the answer - it is about finding innovative ways of joining several technologies to find the best solution for increased productivity, increased yield, improved harvest quality, and sustainability. A farmer may use an autonomous tractor for precision farming, robots for harvesting and data gathering over smaller areas, whilst drones spray crops or gather data over larger areas. Cameras and various sensors are used in combination - fusing data sources via IoT devices that informs artificial intelligence and app solutions used to grow, monitor, and harvest crops at a much higher yield. We look forward to farms in the future being highly digital operations with sophisticated connected data collection technologies (sensors, drones, autonomous tractors, robotic filed scanners, etc.) that feed into integrated farm management platforms for real time decision making, precision agriculture, increased profits, and enhanced environmental sustainability.

IoT Fuel Monitoring Devices; Smart LED lighting arrays for horticulture; electric fence monitoring systems; guidance systems for vehicles; and health & safety monitoring devices are just some of the Agritech products we, at Beta Solutions, have developed for our clients. And in a recent exciting development, three of our team members won the AgTech Hackathon 2020 with their Bugkilla entry. Combining technologies from pest control, artificial intelligence and IoT, the innovative device is designed to lure, identify, and kill the brown marmorated stink bug that is a huge threat to New Zealand's kiwifruit industry. You can read more about Bugkilla in this Farmers Weekly article and we look forward to sharing more on the development of this exciting product in the future.

Do you have an idea or problem that needs an AgriTech solution? Contact us , or better yet, call us today and come and discuss your ideas with us over a coffee.

References:

United Nations. June, 2019. Retrieved from https://www.un.org/development/desa/en/news/population/world-population-prospects-2019.htm l
The Future of Food. National Geographic Magazine. Retrieved from https://www.nationalgeographic.com/foodfeatures/feeding-9-billion/
Growing innovative industries in New Zealand. Agritech in New Zealand Industry Transformation Plan. February, 2020. Retrieved from https://www.mbie.govt.nz/dmsdocument/5878-growing-innovative-industries-in-new-zealand-agritech-in-new-zealand-towards-an-industry-transformation-plan
Agriculture Robots Market Size And Forecast. Retrieved from https://www.verifiedmarketresearch.com/product/global-agriculture-robots-market-size-and-forecast-to-2025/
Root AI. Retrieved from https://root-ai.com
Video. Robotics Plus Kiwifruit Harvester. Retrieved from http://youtube.com/watch?v=b4L-oMd0yVk
ecoRobotix’s AVO autonomous weeding robot. Retrieved from https://www.ecorobotix.com/en/avo-autonomous-robot-weeder/
Technologizing Agriculture. February, 2019. Retrieved from https://cacm.acm.org/magazines/2019/2/234343-technologizing-agriculture/fulltext?mobile=false
John Deere GridCON Electrical Tractor. Retrieved from https://www.electrive.com/2018/12/12/video-john-deere-premiers-electric-tractor-in-action/
John Deere GridCON Electrical Tractor. Retrieved from https://www.ivtinternational.com/news/agriculture/john-deere-develops-fully-electric-autonomous-tractor.html
Aerial Imagery and Seed Planting Drones. Retrieved from https://builtin.com/robotics/farming-agricultural-robots
5 ways technology will change farming in 2019 and beyond. January, 2019. Retrieved from https://www.walleniuswilhelmsen.com/insights/5-ways-technology-will-change-farming-in-2019-and-beyond
Five technologies changing agriculture. October, 2017. Retrieved from https://idealog.co.nz/tech/2016/10/five-technologies-changing-agriculture
AgTech Trends in 2019: Synthetic Biology, Precision Agriculture, and Millennial Farmers. December, 2018. Retrieved from https://research.g2.com/insights/2019-trends/2019-agriculture-agtech-trends
Banner Image by Photo by Kaboompics .com from Pexels
Banner graphics by CleanPNG

Changing How We Work Post COVID19

Thu, 30 Apr 2020 20:17:41 GMT

Author: Brenda Wormgoor, Marketing and Operational Manager

I am happy to report that Beta Solutions has managed the transition to remote working relatively easily. As we had already adopted integrated, cloud-based software solutions and cloud-based communication and data storage solutions we were able to successfully continue working on our clients’ projects remotely with minimal disruption.

I cannot seem to stop wondering - how would we have coped with COVID19 lockdown without the technology available today? Yes, under lockdown a large amount of economic sectors just had to shut down, and physically we are cut off from our friends and family. But we could connect remotely with many sectors able to continue operations and people connecting socially via a plethora of platforms. Is this just a forced once off phenomena or has the way we work and socialise changed forever?

The pandemic forced people across all sectors of society into the extensive adoption of technologies and new ways of working that, due to various social and cultural barriers, have previously only existed in isolated pockets.

According to Ismail Amla, chief growth officer at Capita Consulting:

“This is just the start of a journey. Historically the technology has been ahead of the cultural and societal acceptance of what a new general operating model might look like. Coronavirus has been a ‘jolt to the system’ that will force us to embrace technology and ways of working – from video calls to virtual reality – that we were previously reluctant to use, because we had a choice not to.”1

With hardly any warning, companies have been forced to embrace remote working and digital tools to adhere to government guidelines. In just a few weeks, businesses have shifted to something experts thought would take years.

Remote Working

Remote working has proven to be beneficial to both employers and employees.
Reduced travel time to and from work means employees have more time to spend with their families, attend important events, be more rested and less stressed. This can result in improved productivity and inventiveness.
Concerns over productivity have been a historical barrier to let employees work remotely. If the lockdown proves that workers can in fact do their jobs just as efficiently from home, companies in certain sectors may embrace remote working and adopt entirely new organisational structures.
With remote working and flexible work schedules, employers can reduce their costs by reducing their real estate footprint - only needing to supply hot desks, meeting rooms, and labs.

Technology

As people were confined to their homes, technology became the glue that held both their work and social lives together.
Lockdown has accelerated the pace at which companies embrace digital collaboration tools like Zoom, Slack, Asana, Skype for Business, Google Meet, etc. Augmented and virtual reality tools have proven to be useful in replicating the more personal connection you get from face-to-face meetings - as evidenced in reports of “virtual tea-breaks” where employees congregate on messaging services at the same time to chat.
In the future, technology will bring teams together, both online and offline, as the nature of work continues to shift and be disrupted.

Culture

Another key barrier to remote working has been trust - companies relying on “presenteeism” as a crucial way of managing staff. Remote working has forced employers to treat their employees like adults – and for that to work in the long term companies will need to engage their staff in a renewed sense of collective corporate purpose.
Businesses that showed the ability to nurture and strengthen their company’s culture during this trying time will come out of the crisis more resilient and better equipped to face the challenges and disruptions that lie ahead.

Learnability and Inventiveness

As companies and employees are facing a new world of working, the ability to adapt, innovate, and learn new skills or tools faster than the competition will become the benchmark for success. The crisis forced companies to reinvent work flows, make technology fit for purpose, and perhaps highlighted redundancy. Companies have been forced to change their products/services, distribution channels, etc. - as evidenced in the sudden emergence of gyms offering online workouts, virtual training, and starting online shops selling performance tracking apps, equipment and apparel. The financial outfall will also force companies to reinvent their organisational structures, looking for synergies, efficiency, and outsourcing opportunities.

The question is not whether the working landscape has changed, but by how much and will it have a lasting effect?

As mentioned above, Beta Solutions has managed the transition to remote working relatively easily and by taking office equipment home before the lock down (indeed many of our engineers already had their own electronic tools at home for their own personal projects) we have even been able to do a reasonable amount of physical prototyping. The delays we have experienced are almost exclusively related to shipping delays when procuring project materials etc.

Now that we have moved to Level 3 in New Zealand, staff will continue to work from home unless they have to use the lab at the office for building prototypes, assembly, and small production runs. All staff have been briefed on the COVID19 heath guides and we have prepared the office for safe hygiene practices.

Our wide-ranging experience and knowledge of leading technologies enable us to design electronic products for a wide range of industries. With the collective knowledge and experience in our team, we can meet almost any electronic design challenge. Below are some of our company stats:

So, if we can help you with your project or if you are looking for an outsourced electronic design partner, call us today for your 1-hour free virtual consultation.

References:

"How COVID-19 is Changing How We Work" from https://www.viewpoint.com/blog/how-coronavirus-covid-19-is-changing-how-we-work-part- 1
"How The COVID-19 Pandemic Is Accelerating Change" from https://www.experfy.com/blog/future-of-work-how-the-covid-19-pandemic-is-accelerating-change
Banner image graphics from Technology Vectors by Vecteez y
Image of Video Conferencing by Tumisu from Pixabay https://pixabay.com/photos/video-conference-business-meeting-4221401/
Image of nurse holding the globe by Anna Shvets from Pexels https://www.pexels.com/photo/man-people-woman-hand-4167539 /
Image of New Zealand Government COVID19 Communication from Whangarei Distric Council http://www.wdc.govt.nz/Pages/Default.aspx

Bluetooth 5.1 - A game changer for local positioning?

Wed, 08 Apr 2020 20:33:38 GMT

Author: Siddharth Venkatesh, Intern

Bluetooth 5.1 is the latest iteration of the wireless technology, and introduces the ability to detect the Angle Of Arrival of the wireless signal. Such disruptive technology introduces a vast range of new opportunities - such as going a long way to solve the present issues with Local Positioning Systems.

Bluetooth Technology - A Quick Recap

Bluetooth is no doubt a word we have all heard. Since the advent of version 1.0 in 1999, the technology has become synonymous with short range wireless communication found in consumer electronics today. Indeed, whether it is to transfer files between devices or listen to our favorite songs using wireless headphones - Bluetooth is often the technology of choice.

Bluetooth operates by using (short range) UHF radio waves at a frequency band of 2.4GHz to 2.48GHz. Incidentally, this is the same frequency band as another popular wireless communication protocol - Wi-Fi. However, these are two very different protocols and intended for very different applications. Bluetooth technology devices transmit a much weaker signal than Wi-Fi and therefore are intended for short range applications. The upside of this is significantly reduced power consumption, and therefore longer battery life.

Bluetooth has evolved markedly over the years.

Versions 1-3 (now known as “Bluetooth Classic”) was a point to point (1:1) topology - enabling various devices to pair together and exchange data.
Version 4 (known as “Low Energy”), introduced in 2010 - is a point to many (1:M) topology. That is, Beacons can “broadcast” their data to many Receiver Nodes. Additionally, the technology was designed to be ultra low power - enabling low power wireless sensing.
Version 5 (“Mesh”) was first introduced in 2016, and is a many to many (M:M) topology. Any nodes can talk to any other nodes, enabling large device networks. You can read more about this in our Blog A Closer Look at the New Bluetooth Mesh from 2017.
Version 5.1, as we will see, is a further advancement on the BT Mesh protocol - and introduces a very exciting new feature.

Bluetooth 5.1

Bluetooth Special Interest Group (SIG) announced in January 2019 that 5.1 would include new “direction finding” technology. Undoubtedly, being able to determine the position of an object will unlock a wide range of commercial opportunities.

While Global Navigation Satellite Systems (GNSS) are relatively good at resolving positions, they will only work if the GNSS receiver can “see” 3 or more satellites - making them unsuitable for indoors.

Current Indoor Positioning Systems use the Received Signal Strength Indication (RSSI) technology to triangulate the position of other Bluetooth devices. This often requires multiple receivers (more the better) to “resolve” the transmitter’s location - and the end results are still often not highly accurate.

This new direction-finding technology, however, calculates the angle of transmission of the Bluetooth device to accurately pinpoint its location. This technology is three times more accurate than RSSI technology ¹ .

There are two methods for direction finding, the Angle of Arrival (AoA) and Angle of Departure (AoD) methods - which are briefly described as follows:

Angle of Arrival (AoA):

The Angle of Arrival method consists of an antenna array on the receiver end. At least two antennas are required - but more antennas will enable more direction precision.
The transmitter emits what is called a “Constant Tone Extension” packet, which is received by the receiver's antennas and ultimately sampled and digisited both In-phase and Quadrature components (IQ) - and by the Bluetooth receiver’s circuitry.
Essentially, as each of the antennas will receive the packet at a different phase, this phase delay can be used to determine the AoA of the transmission.

Figure 1. Angle of Arrival (AoA) ¹

Angle of Departure (AoD)

The Angle of Departure method also consists of an antenna array, but on the transmission end.
Each antenna in the array transmits a Constant Tone Extension (CTE) packet which again is IQ sampled in the receiver by a single antenna.
Based on the samples, the angle of transmission can be calculated.

Figure 2. Angle of Departure (AoD) ¹

It is important to note that 5.1 technology only provides inherent object direction information and not object position - as distance information is not included. However, as per existing methods, Distance (and therefore position) can be approximately deduced by measuring the received signal strength.

Bluetooth 5.1 Applications

While most people are familiar with the consumer sector of the Bluetooth market (e.g.: audio, wearables, entertainment, personal handhelds, etc.), it is increasingly being used in industry and other IOT applications. Some possible applications are listed below:

Smart Buildings
There is an increasing demand for smart buildings in today's economy. Smart buildings encompass airports, offices, hospitals, museums etc. Bluetooth 5.1 applications such as wayfinding and asset management in buildings help improve the occupant experience and operational efficiency of the building.

Figure 3. Wayfinding in Buildings

Figure 4. Asset Management in Offices

Wayfinding in Buildings:
Devices such as phones, tablets and PC’s can be used to connect to bluetooth nodes placed inside buildings to accurately provide pathfinding functionality. These nodes can be strategically placed to provide the optimum route to your destination.
Asset Management in Offices:
Large offices and hospitals are equipped with many assets. Accurately tracking the location of these assets is presently difficult and can lead to delays and inefficiencies. With embedded bluetooth tags in all assets, the location of each one can be monitored continuously, improving asset management.

Smart Homes
With the advent of home automation, smart homes have become widely popular within the market. Home automation encopasses a range of products such as portable speakers, smart locks, smart switches etc. These products aim to provide security, dependability and reduced power consumption to homes.

Personal Property Tags
The introduction of Bluetooth 5.1 has opened up another avenue in home automation. The new direction finding functionality allows for tracking personal items within the household. With bluetooth enabled nodes placed in each room, there will be no more frantic searches for keys in the morning before work! Favoured personal items can be embedded with a bluetooth 5.1 tag which will allow for continuous tracking of those items.

Figure 5. Personal Property Tags

Smart Industry
With the industrial IOT revolution, bluetooth has helped manufacturers achieve an increase in productivity. Smart Industry has helped reduce operational costs and increase efficiency2.

Asset Tracking
With thousands of assets moving in and out of warehouses, a reliable method of tracking these assets is required. This is where Bluetooth 5.1 comes in. With its accurate direction finding capability, manufacturers will be able to track the location of these assets in the warehouse at any time. Moreover, since the tags embedded in these assets are bluetooth enabled, they will be able to store and communicate details about the asset themselves.

Figure 6. Asset Tracing in Warehouses

Conclusion

The introduction of Bluetooth 5.1’s new direction finding capabilities is certainly changing the game in terms of indoor direction finding - with it now being possible to resolve an object's position to a much greater resolution than traditional RSSI triangulation methods.

Beta Solutions has experience with developing BT 5.1 technology and if you have an idea of how you could use 5.1, or need advice regarding hardware or firmware design techniques for Bluetooth Networks - feel free to get in touch with the team at Beta Solutions!

References:

Bluetooth Direction Finding: Angle of Arrival (AoA) and Angle of Departure (AoD). (2020, January 24). Retrieved from https://www.silabs.com/products/wireless/learning-center/bluetooth/bluetooth-direction-finding
2020 Bluetooth Market Update. Retrieved from https://www.bluetooth.com/bluetooth-resources/2020-bmu/
Images retrieved from https://www.pexels.com

COVID-19 Update: We Are Still Open

Thu, 26 Mar 2020 02:26:52 GMT

Author: Terry Southern, CEO

Beta Solutions COVID-19 Trading Update.

To our valued Clients,

Have you tried turning 2020 off and then on again? It is certainly a surreal situation and our thoughts go out to everyone who is affected by COVID-19.

I just wanted to let you know that we’re still open for business.

As per the NZ Government's directive - from Tuesday 24 March our office will be closed for at least 4 weeks, and our team will be working remotely.

Existing projects:
We have a work-from-home plan in place with good systems and procedures that will allow us to continue working on many projects.
You can contact our staff during office hours via:
Office phone (+64 (0) 62802830) - please be patient as your call will be diverted to a staff member remotely.
Mobile - You may contact your project manager on their mobile phone, which you will find in their email signature.
Email:
All staff are contactable via their work emails.
Alternatively email: contact@betasolutions.co.nz
Most of our current projects booked into our schedule are still on target, although some projects may be impacted.
In any case, your Project Manager will be in contact with you.
If, on your side, your project or potential project is likely to be held up (for whatever reason), it would be appreciated if you could let us know as soon as possible, so we can plan accordingly.
New projects:
We look forward to hearing from you regarding any new projects. Please contact me on the contact details supplied below.
We are able to offer video and telephone meetings.

If you have any questions, please feel free to contact me.

Let's work together to lessen the impact of COVID-19.

Take care.

Kind Regards,

Low Power Firmware Design for Wireless Sensor Networks

Fri, 29 Nov 2019 02:24:02 GMT

Author: Nathan Geddes, Electronics Engineer

Introduction

Wireless Sensor Networks (WSNs) are a rapidly growing technology which enable us to better understand and manage the environment we live in. Their applications are vast and they can bring benefits to a wide range of industries - including Agriculture (e.g.: crop monitoring sensors), Manufacturing (e.g.: inventory management sensors) and Healthcare (e.g.: monitoring and diagnostics sensors) ⁽¹⁾ .

One of the "holy grail" aspects of wireless sensor design is to be able to achieve very low power consumption - enabling the devices to operate for long periods of time (many years) on a single small battery. Many design techniques can be used to minimize this power consumption spanning across both hardware and firmware.

This blog specifically focus on a few key firmware design techniques to help maximize that all-important battery life-time.

Quick Recap on WSN Topology

Wireless sensor networks generally consist of up to 3 types of nodes - defined as follows:

Gateway node - the connection for your network to the internet.
End (sensor) nodes - gathering data, packaging into a protocol that the network can understand and transmitting the data across the network.
Router nodes - nodes placed between sensors and gateway to forward data toward the gateway (NB: These can also be sensor nodes).

Figure 1: Wireless sensor networks ⁽⁵⁾ .

Wireless sensor networks consist of several operational "layers of communication" - as indicated in the diagram below. Firmware techniques generally apply to the link layer and higher layers.

Figure 2: Wireless sensor network layers of communication.

Low Power Firmware Design Techniques for WSNs

#1. Sleep and Wake-up

Generally, keeping a wireless transceiver on all of the time is very expensive - in terms of power consumption. Therefore, it makes sense to turn the transceiver off whenever it is not being used. 'Sleep and wake-up' is the general term for such a technique - where the transceiver 'sleeps' when wireless communication is not being performed then 'wakes up' to transmit or receive data.

This is elaborated upon in more detail as follows:

1.a) Synchronous Wake-up (a.k.a Scheduled Rendezvous Technique) ^{(2) (3)}

A simple and often effective approach is to duty cycle this sleep and wake-up time for all nodes and use a scheduled 'rendezvous' - e.g. turn the transceiver on for two minutes - on the hour - every hour.

The challenges and drawbacks of implementing such an approach are:

Choosing an appropriate duty cycle period - as one has to balance power consumption with the need to transmit timely data. Therefore, this approach may be unsuitable for 'real-time' data networks.
Wasting energy due to 'redundant' wake-ups, when no communication is actually required.
Any nodes participating in communication must take into account 'clock drift', so they can ensure they wake-up at the same time - especially if there is a short time window to complete communication.
Preventing (or dealing with) possible packet collisions between nodes - since they share the same wake-up period.
Waking up all nodes is wasteful as only a few may be required to forward data.
Once also needs to consider that there is also an energy overhead in sending such time synchronisation data packets.

Figure 3: Synchronous wake-up.

Suitable Application Type:

Regular data trend logging / monitoring where real-time delivery is not required. E.g. ambient temperature.

1.b) Asynchronous Wake-up ^{(2) (3)}

An alternative to waking up the transceivers in a synchronous manner is to ... (no surprises here)... wake them up asynchronously. That is, nodes wake-up at independent times for data transmission. However, to ensure the sender's messages is received/relayed, the senders "wake-up period" must still overlap with a neighbouring receiver's wake-up period.

Pros: (Compared with Scheduled Rendezvous)

Wake-up intervals can be shortened for nodes with no data to send as they only need to detect traffic to be forwarded.
Strict time synchronization is not required between nodes.

Cons: (Compared with Scheduled Rendezvous)

Not suited to broadcast traffic.

Figure 4: Discovery of an asynchronous sender through periodic listening. The sender transmits a single long discovery message. ⁽³⁾

Application Type:

Regular monitoring where nodes transmit data at different intervals e.g. Health monitoring - where different sensor types require different sampling rates. (e.g.: Frequent Heart rate sensing Vs Infrequent Sleep sensing)
(This can improve on specific scenarios over Scheduled Rendezvous where nodes must wake-up for a longer window even if they have no data to send.

1.c) On-Demand Wake-up ^{(2) (3)}

A possible alternative to Synchronous and/or Asynchronous wake-up is to simply enable the transceiver nodes when communication is needed. (i.e.: If the node has sensor data to send or if it needs to relay data).

Pros: (Compared with Synchronous and/or Asynchronous wake-up)

Data can be sent across the network with minimal latency.

Cons: (Compared with Synchronous and/or Asynchronous wake-up)

A receiving radio is needed to listen for wake-up requests from transmitting node. This receiving radio understandably will draw it's own electrical power - which will need to be carefully considered. (i.e.: Will the energy consumption of this receiving radio out weigh the energy saved from only waking up the transmitting nodes when required?).
Additional hardware costs for additional receiver radio.

Figure 5: On-Demand wake-up.

Application Type:

Event based monitoring - when real-time delivery is essential, e.g. Surveillance - movement sensors, Forest-fire detection sensors.

#2: Transmission Power Control

If implemented ineffectively, a wireless transceiver network can consume a significant amount of electrical energy. Energy is consumed at the highest rate when transmitting data and so to that end, careful consideration needs to be given as to how much 'juice' to pump into the transmitter. This is a main factor which determines how far the wireless signal will travel while being recoverable at the receiver.

Transmitting at a higher power than necessary is wasteful as the transmitting node will deplete its battery much faster - especially if it sends data frequently. In network topologies where there are many nodes in range of each other, this is also wasteful for any receiving nodes which will 'overhear' messages sent at high power when they are not the intended recipient.

Transmission Power Control techniques tend to use the RSSI (received signal strength indicator) of a receiving node to set the transmission power of the transmitting node or even adapt it over time - since weather or other interference can affect the link quality over time. Often designers here want to find a balance between reliable data communication and power efficiency.

Transmission Power Control can be used to implement another technique called topology control. This defines the path the data will take from a transmitting node to the network gateway by adjusting the transmission power of nodes. The purpose of this technique is to favour nodes with a higher battery level remaining to perform relaying of data thereby fairly distributing the communications workload.

NB: The WSN lifetime is often only as good as the lifetime of the first node to fail!

Figure 6: Using transmission power control to implement topology control - balancing energy consumption between nodes.

Application Type:

Transmission power control in general is applicable and essential to all WSN deployments.
Toplogy control is more applicable to large distributed mesh networks e.g. agriculture, warehouse inventory tracking.

#3. Data Reduction

As mentioned above, one of the key factors in reducing power consumption in WSNs, is to only have the transmitter switched 'on' for as short a period as possible. Therefore, it follows that it would be an objective to transmit data as efficiently as possible, so as to prevent unnecessary transmitter 'on time'. There are several techniques to accomplish this - some of which are listed below:

3.a) Data Aggregation

Does every reading actually need to be transmitted or can several be collated and sent? Aggregating data greatly reduces overhead associated with communication. Therefore, for instance, it will be more efficient to send 24 readings once per day, as opposed to sending one reading 24 times per day.

NB: Data aggregation is not suitable for when real time data is required. However, it will be possible to create 'events' that, when triggered, could transmit all stored data. (e.g.: a tracking device might send all it's 'bread crumb' data if it detects a significant impact - indicating a crash event.

3.b) Compression/Coding

Data compression is used everywhere in the modern digital age - to save storage, processing requirements or reducing internet data usage. Audio, video, images and text files all make use of some compression method. E.g. Spotify uses audio compression so you don't use up your mobile data too quickly.

'Raw' data, say from a microphone, camera, or other sensor often contains much more information than we really need. In the world of digital electronics, information is stored in 'bits' and therefore more information = more bits. In turn more bits = more energy required to transmit those bits.

Compression can be 'lossy' or 'lossless'. As those names suggest, 'lossy' compression loses information in the process of compression. The receiver of such data can not recover the original uncompressed data but instead an approximation of it. This approximation is often 'good enough' as we either don't notice the loss of information or we simply don't need it. This approach tends to allow data to be compressed further than 'lossless' approaches - in which the original data can be exactly recovered.

Figure 8: Compression that removes blank values (white) and encodes repetitive values. ⁽⁴⁾

Quantization

An extremely simple form of lossy compression applicable to wireless sensors is quantization - which basically reduces the number of bits used to represent a unit of data. For example, let's assume we want to measure and transmit a temperature reading:

Scenario 1: 16 bit resolution.
If the full scale temperature range of a temp sensor was 0-99°C, then 16 bits of resolution (65536 in decimal) would mean we have a theoretical temperature resolution of 0.0015° degrees. This is likely far in excess of what is required - and likely below the noise floor of the electronic circuitry anyway.

Scenario 2: 8 bit resolution.
Alternately, if that same temperature sensor was quantized into 8 bits of resolution (255 in decimal) - this would result in a theoretical temperature resolution of 0.39° degrees. This (8 bit resolution) temperature will likely be all we need - and therefore we will reduced the energy required to transmit such data (compared to 16 bits).

Other Techniques

Many other more complex approaches take advantage of the fact that most data in the real world is not entirely random, but rather follow some sort of pattern - e.g. temperature rises at day, decreases as night. In such cases, the data can be cleverly represented in another form rather than individual data points. For exaple Fourier transform can represent signals (e.g. audio and images) as frequency-based information rather than time-based, then some quantization can be applied to the very-high frequencies which humans don't notice.

Additionally, in large data sets, there are often measurements that have a higher probability of occurring than others. Therefore, it is prudent to use fewer data bits to represent higher probability measurements and more bits to represent less likely measurements (e.g. Huffman coding) to reduce the total number of bits to represent a set of data.

Application Type:

Quantization can be applied in many applications as it is computationally inexpensive to run on small wireless sensors.
Other techniques are more applicable to high data rate sensors - where the energy savings can outweigh the computational energy requirements. E.g. audio, electrocardiogram monitor.

3.c) Adaptive Sampling

There are often scenarios when frequent data transmission is not required most of the time - but only in certain conditions. This is where adaptive sampling comes in highly useful, as it is a less computationally expensive way to reduce data transmission.

Application Type:

Where the application may not require regular sampling of data - e.g. a river monitoring device might only transmit data a few times per day under 'low flow' conditions - but this might increase to several times per hour in 'flooding' conditions.

#4. Routing Protocol / Clustering

Clustering is a useful technique when a WSN consists of a large number of nodes. In such scenarios, data from a transmitting node can make several shorter 'hops' through intermediate nodes toward the gateway, so that transmission power can be reduced. As mentioned already, the lifespan of a WSN is often only as good as the first node to fail - so clustering approaches aim to distribute the routing tasks evenly amongst nodes to maximize this overall lifetime.

As the name suggests, nodes form 'clusters' - in which there is a designated 'cluster-head'. The cluster-head is used for forwarding data from its cluster nodes to the gateway. Ultimately, the cluster-head will consume the greatest energy over time due to forwarding data form several other sensor nodes - so the cluster-head is dynamically changed to a new node over time (one with a high amount of remaining energy).

Figure 8: Routing Protocol - Clustering. ⁽⁶⁾

Application type:

Large distributed mesh networks e.g. agriculture, warehouse inventory tracking.

Other Firmware Considerations

The WSN firmware techniques listed above are by no means an exhaustive list of all possible ways to write power efficient firmware. Designers will also need to consider many other aspects associated with their specific design - which is beyond the scope of this Blog.

For example, designers should consider:

How to sleep their specific micro-controller - to reduce significant power.
How to use slower clock rates when possible to reduce power.
How to power down other hardware peripherals when not required.

Conclusion

This blog has taken a high level look at four firmware design techniques for Wireless Sensor Networks. As can be seen, there is certainly no 'one size fits all' for reducing energy consumption, and the approach taken will vary depending on the specific application.

If you have an idea, or need advice regarding hardware or firmware design techniques for Wireless Sensor Networks - feel free to get in touch with the team at Beta Solutions!

References:

https://radiocrafts.com/applications/wireless-sensor-networks/
https://hal.archives-ouvertes.fr/hal-01009386/document
http://cnd.iit.cnr.it/andrea/docs/aohoc09.pdf
Faust, David. 2013. "Shrink Your Storage Needs with c‑treeACE Data Compression." Faircom Corporation, September 30. Updated 2015-11-17. https://www.faircom.com/insights/shrink-your-storage-needs-with-c%E2%80%91treeace-data-compression
Adapted from https://www.elprocus.com/wp-content/uploads/2014/03/28.jpg
Adapted from https://media.springernature.com/lw785/springer-static/image/chp%3A10.1007%2F978-981-10-7323-6_22/MediaObjects/455586_1_En_22_Fig1_HTML.gif
Banner background image retrieved from: https://www.pexels.com/photo/brown-and-green-mountain-view-photo-842711/

Precision GPS for NZ... on it's way

Tue, 30 Jul 2019 03:08:44 GMT

Author: Terry Southern, CEO

Background

Back in September 2017, we wrote a Blog on New Zealand Trialling Centimetre Level GNSS . We are excited that the New Zealand government have now announced that they have budgeted nearly $2m to support the development of a regional satellite-based augmentation system (SBAS) to significantly improve GPS accuracy. This blog takes a quick recap of the SBAS technology and why it will be so valuable for New Zealand.

What is GNSS again?

For those who want a more in-depth read (on the history and technical nature) of the Global Navigation Satellite Systems check out our Blog on New Zealand Trialling Centimetre Level GNSS .

In any case, a very brief recap is provided below:

The first Global Navigation Satellite System (hereafter "GNSS") was called the Global Positioning System ("GPS") network - which was first launched by the USA in the late 1970s.

Figure 1: U.S. Global Positioning System (GPS) ⁽¹⁾

Since then modern civilisation has unquestionably developed an ever increasing dependence on global navigation systems.
Along with GPS, there now exist other GNSS providers - including the (European based) Galileo network, the (Russian based) Glonass network and more recently the (Chinese based) BeiDou network.
Common to all GNSS networks are:
A network of satellites (orbiting approximately 20,000kms above earth) with extremely precise atomic clocks.
Ground based GNSS receivers.
Provided that a single GNSS receiver is able "see" 4 or more satellites, it can calculate its position on the planet.

Figure 2: GNSS receiver with GNSS satellites transmitting their information needed for receiver to calculate its position. ⁽²⁾

However, accuracy (for a single GNSS receiver configuration) is typically only ±10 meters.
Accuracy errors are due to: Satellite clock errors, Orbital errors, Atmospheric disturbance errors, GNSS receiver electrical noise, and multipath propagation.
While an error of ±10 meters is perfectly acceptable in some situations (eg: To "Find My Friends"), it is totally inadequate for other applications (eg: Autonomous vehicles).

Figure 3: Contributing factors to GNSS positional error. ⁽²⁾

In the 1990s, a technology called Differential GNSS (DGNSS) was implemented to improve the accuracy of GPS to around 1-2 meters. More recent technology (such as RTK) improves the accuracy to 1-2 centimetres.
DGNSS typically works by having a reference station in a known fixed location, and some form of radio link to a "rover". Since the reference station has a known location, any difference in the calculated position and its actual position can be considered as an error. This error or difference is then sent out to the rover via the radio link, allowing for the error to be subtracted at the rover.

Figure 4: Diagram showing key features of DGNSS.

Traditionally, the major downside of classic DGNSS is (high) setup cost and personal responsibility. That is, each operator wishing to implement a DGNSS solution has to pay thousands of dollars in setup costs, and is entirely responsible for the setup and maintenance of the network. This naturally inhibits many users from access to such technology.

So what is SBAS technology?

Much like DGNSS, the Satellite Based Augmented System (SBAS) also makes use of reference stations to provide correction data to the rover. However, rather than using a single reference station, SBAS makes use of several reference stations spread over the entire operational area. By using several reference stations, a map of the atmospheric effects is generated and this allows the system to cover large areas. Along with the atmospheric effects, the reference stations also allow for correction data of satellite orbital error and satellite clock errors to be accounted for. The correction information is then beamed to a geostationary satellite, which in turn broadcasts the correction data back to earth, allowing the rover to make a local correction. Since this correction data is sent over the existing GPS L1 channel, this can be received by existing GNSS receivers without the need for any additional antennas or hardware. This means that the existing single receiver GNSS devices could make use of this correction data with just a firmware update. Systems similar to SBAS have been available in the USA (Wide Area Augmentation System WAAS) and Europe for a number of years and allows for a positional accuracy of 1 - 2 m.

Figure 5: Diagram of SBAS core features. ⁽³⁾

Why doesn't NZ and Australia have SBAS already?

Like all infrastructure, it takes plenty of time and money to roll out nationwide.

Time:

As this technology will ultimately be used for Saftey Critical applications (eg: Aviation), it is understandable that it needs to be thoroughly trailed and tested first.

In 2016 Land Information New Zealand (LINZ) announced that it was collaborating with Geoscience Australia (now called "Positioning Australia") in a two year trial of SBAS. As part of the SBAS test-bed trial, three technologies were assessed:

Single frequency service SBAS - which is the equivalent of the current systems used in the US and Europe. Should enable accuracies of ±1m.
Dual frequency/Multiple Constellation SBAS - known as "next generation SBAS" and will use the recent development of a civil frequency, known as L5 for GPS and E5a for Galileo. Should enable accuracies of sub ±1m.
Precise Point Positioning (PPP) - which is a method that provides highly accurately position solutions with accuracy better than ±10cm. (Refer to our previous GNSS blog regarding the pros/cons of PPP).

Figure 6 – Present GPS channels. ⁽²⁾ These trials (between January 2017 and January 2019) are thought to have been highly successful.
Read more at Australian Flying .

Money:

Indeed, implementing this technology will not be a low cost exercise. Land Information New Zealand (LINZ) has set aside $NZ1.992 million in their 2019 Budget while Positioning Australia received $A160.9 million in their 2018-19 Federal Budget. Both will co-operate in delivering a regional satellite-based augmentation system for both countries. The funding will be used by LINZ and Positioning Australia to jointly develop specifications and to undertake initial procurement processes. Once a preferred provider for delivering SBAS has been identified, approval for further funding will be sought to implement a regional SBAS.

“ GPS usually provides positioning information accurate to about 5-10 metres. This new system will improve the accuracy to less than a metre, and in some devices to 10 centimetres. ”

“ This data is fundamental to a range of applications and businesses worldwide. It increases our productivity, secures our safety and propels innovation. ”
Matt Amos, LINZ National Geodesist. ⁽⁴⁾

What are the benefits of SBAS?

In short, SBAS will make positioning data, like GPS, even more accurate than it is now.

Examples of SBAS applications include:

Precision navigation for: manned and unmanned aircraft, driver-less cars and other autonomous vehicles.
Eg: SBAS could provide accurate vertical guidance for landing procedures for rescue helicopters, meaning they can reach patients in difficult terrain more quickly and in more challenging weather conditions such as low cloud.
Virtual fencing where livestock in the future may wear GPS-enabled collars to stop them going where they are not supposed to – replacing some physical fences, keeping livestock out of waterways, and making grazing more efficient.
Workplace health and safety, particularly in the forestry sector, where more accurate GPS can alert workers using equipment to other people in the area.

When will SBAS be available?

As mentioned above, it does take time to roll out such safety critical technology. Therefore New Zealanders (and Australians) will need to "hurry up and wait" before getting their hands on this new tech.

Investigations and procurement will be carried out this year and next. Tenders will apparently be out in 2021 and hence the technology is only slated to be operational in 2023. Read more here .

Conclusions:

At Beta Solutions we will continue to follow SBAS roll out closely as we are excited about the possibilities this new technology will unlock in the product design solutions we can offer our clients.

If you have a problem which could possibly be solved through the use of accurate Global Positioning, feel free to contact us to talk to someone in our knowledgeable team.

References:

Figure 1. Retrieved from https://www.google.com/search?q=precision+gps&hl=en&tbm=isch&source=lnt&tbs=sur:fmc&sa=X&ved=0ahUKEwjAr4uu1MzjAhUKfysKHdFrDzAQpwUIIw&biw=1731&bih=836&dpr=1.1#imgrc=mY4y84fY1W_4yM:
Figure 2 and 3. Retrieved from Jeffrey, C., (NovAtel, 2010). “An Introduction to GNSS.”
Figure 5. Retrieved from CRCSI, (July, 2017). “Technical Specifications Document for Satellite-Based Augmentation System (SBAS) Testbed.”
Matt Amos, LINZ National Geodesist. Retrieved from https://www.linz.govt.nz/data/geodetic-services/satellite-based-augmentation-syste m
Images retrieved from:
Banner: https://www.nbcnews.com/technology/x-47b-navy-drone-take-first-stab-unmanned-carrier-landing-6C10591335
Rescue helicopter: https://coastguard.dodlive.mil/2014/02/from-air-station-to-ice-station/
Virtual fence: https://www.sott.net/article/134719-General-Law-for-Cows-Clever-collar-keeps-cows-in-a-virtual-paddock /
GSP Satellite: https://commons.wikimedia.org/wiki/File:GPS_Satellite_NASA_art-iif.jpg

Key Takeaways from Techweek19 - Innovation and Industry 4.0. What to do when there is no off-the-shelf solution

Fri, 21 Jun 2019 02:19:38 GMT

Author: Brenda Wormgoor, Marketing and Operational Manager

Introduction:

With over 500 events, in more than 30 towns and cities around New Zealand, Techweek is a National Tech Innovation Week aimed at building enthusiasm around new technologies being developed in New Zealand.

This year Beta Solutions decided that it was about time that we sign up and represent Tech in our region. We wanted to host an event that would give local entrepreneurs and staff from innovative companies some industry insights and knowledge that they can use in their workplaces and careers. There is a lot of buzz around Industry 4.0 in the media and we hear clients asking - what is it, why must we know about it, what should we do about it, what does research and development mean for my/our company, where do we start? Thus, we decided to host an event that would address these concerns and questions in the form of a Business Breakfast: Innovation and Industry 4.0 - What to do when there is no off-the-shelf solution .

We were delighted to host over 50 attendees at the event in our office. It was a pleasure to share a continental breakfast with everyone as a small thanks for all their contributions to technology and/or innovation in our region. It was also very encouraging to see that woman made up 20% of the attendees.

There has been such positive feedback on the presentations and client case studies that we decided to share some of the key takeaways in this Blog.

"Firstly congratulations to you and the team at Beta Solutions for a fantastic breakfast session yesterday. By far the best 'tech' event I have attended in Palmerston North! For me the critical factor was that the event was Industry led. I am looking forward to going through your presentation again once they are on your website."

We asked the attendees what their key takeouts were in a post-event feedback questionnaire, and following is what they shared (grouped in the topics addressed).

Industry 4.0 - An overview and Industry 4.0 Readiness:

The avalanche of disruptive technologies that is fast changing the world as we know it, is blending the physical, digital and biological worlds. During the presentations, Aaron Fulton (Beta Solutions' Systems Developer), focused on:

End-to-end digitization and data integration of the value chain.
Cyber-physical Systems
Internet of Things (IoT)
Cloud Computing
Cognitive Computing
We also looked at what it would take for companies to be ready to capitalise on opportunities created by this 4th Industrial revolution.
Our challenge as businesses/entrepreneurs today includes:
Responding to new opportunities and challenges.
Learning how to adapt and innovate quickly.
Expectations of our customers are ever increasing, and so our ‘baseline’ of service has to as well.
To compete successfully, there are two areas of focus for businesses - Digital Maturity and Organisational Mastery Areas.
This Slide is a good guide to work through to determine your company's readiness digitally and across the four organisational mastery areas ⁽¹⁾ :

Key takeouts from attendees:

"We need to adapt to keep up with the competition"
"Industry 4.0 isn't a term I'd heard before, and Beta Solutions' R&D process is inspiring."
"Need more industry sector leadership."
"There is a vibrant tech culture in Palmerston North."
"Interesting ideas coming to market locally. Beta Solutions is working on some exciting projects."

Case Studies:

We were very excited to share two of our Clients' case studies, which made the content more real. A huge thank you to our guest speakers:

Mark Glenny - Innovation Research Manager at Resene Paints Ltd.
Mark shared the Resene Conductive Coatings R&D journey (from market research, ideation, development to commercialisation) of the newly launched Resene SmartTouch ^(R) product.
This a great example of Industry 4.0 as it incorporated: a cyber-physical system; user-centric function and performance; system control; mechanical and electrical engineering; micro-controller; and connectivity.
You can read more about the project here .

Kevin Halsall - Founder, Director & Designer at Omeo Technology
Kevin shared the Omeo (personal mobility device) product development and commercialisation journey.
During his presentation he highlighted some of the research and development process steps they had to overcome, such as prototyping and user testing.
Be sure to watch this video of how Omeo is transforming people's lives in their video: Bush adventure .

R&D Process and Leverage Partnerships:

60% of the respondents in our survey of the event rated the R&D Process the most valuable session of the event.

During these sections, Terry Southern (Beta Solutions' CEO), discussed:

Research and Development Process:
First we looked at what is Research and Development.
Terry explained that R&D comprises of two distinct parts - "Research" and "Development" - and each needs to be understood for its role and attributes.
An insightful discussion followed on how using a decision making framework can help with understating the complexities of R&D projects.
Not all ideas, systems nor R&D projects are created equal!
This session created a lot of interest as Terry shared how Beta Solutions has adopted the Cynefin Framework ⁽²⁾ (by Dan Snowden) to help understand and manage our clients' R&D projects.

Over the last 10 years Beta Solutions has developed and is constantly refining its R&D Process , which involves:
Project Management
Designing
Prototyping
Validation and Testing
Regulatory considerations
Manufacturing
Due to the response to this session, we will look to discuss this in a future Blog.
Leverage Partnerships
Partnerships can be the difference between product success … or not, so Terry shared that companies and entrepreneurs should bear in mind the following when considering innovation partnerships:
Partnerships that provide Value
Evidence of shared common values
Understanding the optimum structure for your (company's) R&D - in-sourcing, outsourcing or perhaps a hybrid approach.

Key takeouts from attendees:

"The realization of the complexity involved in getting an idea to market."
"Best explanation of the R&D process I have heard."
"Learning about Beta's approach to R&D."
"R&D Process and Project Management information was valuable."
"We're investigating R&D Processes in our Tech consulting business. This session was very timely and helped me understand methodologies and what to look into."
"Doing it alone is the exception. Find the right partner."
"Networking and finding potential partners to work with."

Conclusion:

After such positive responses to the event, we are looking forward to hosting the next event during Techweek 2020.

As for Industry 4.0 and Innovation - the future is upon us, so... What to do when there is no off the shelf solution? Put your best foot forward … and innovate.

Remember: Ideas do not equal Innovation.
R&D is difficult but not all R&D is the same:
It is better to understand the nature of your R&D through a decision making framework, such as Cynefin.
Up-skill on the R&D process in order to have realistic expectations.
Leverage partnerships that add value and get your idea to market faster.
You can read our Blog on "Why Outsource Your Electronic Design? "

Or better yet, contact us and come and discuss your takeouts or ideas with us over a coffee.

References:

PWC Industry 4.0 Self Assessment - https://i40-self-assessment.pwc.de/i40/landing/ and https://www.strategyand.pwc.com/industry4-0
Cynefin Framework Introduction - http://cognitive-edge.com/videos/cynefin-framework-introduction/

Why Outsource Your Electronic Product Design?

Thu, 28 Mar 2019 00:58:54 GMT

Author: Brenda Wormgoor, Marketing and Operational Manager

Introduction

Outsourcing involves the contracting out of a business process to another party. To outsource is primarily a business decision for companies because it can reduce their overhead to produce a product, thereby increasing their profit.

According to Brandon Gaille ^[1] , the top reasons why businesses outsource are to:

Reduce or control costs
Gain access to resources unavailable internally
Free up internal resources
Improve business or customer focus
Accelerate company reorganisation/transformation
Accelerate project
Gain access to management expertise unavailable internally
Reduce time to market

It is expedient and advantageous to leverage expertise not available in-house to take ideas from concept to production. In this blog we will take a closer look at four key benefits to outsourcing electronic product development.

Benefit 1: Expertise

Some companies may do some of the work in-house, but with each unique project comes the requirement of expert knowledge and experience depending on the technology involved in the project. Accessing the skills and knowledge of experienced electronic design engineers, whose bread and butter is to be at the forefront of new technology, can give your product the edge to succeed.

The right electronic development partner should add value to your company in the following areas:

Working exclusively in electronic product design and development, they acquire knowledge about various markets, industries, and technologies that can make your product the best it can be.
Multidisciplinary perspectives (hardware, embedded software, mechanical integration) that leads to innovative ideas, offering the right technology and design solutions to your project's problem.
Knowledge of the process, stages and requirements of electronic product development.
Understanding and communicating the risks associated with particular decisions, and recommending ways to mitigate such risks.
Knowledge and advice of the regulatory compliance required for your electronic product.

Benefit 2: Speed to Market

Product life cycles for electronics are significantly shorter today, and experienced electronic product developers with the correct in-house resources can deliver new products with speed. This is due to:

Inherent knowledge that cuts down learning and exploration time. Your project benefit from their years of experience and established industry relationships from technology to production suppliers.
In-house resources to deliver rapid prototyping so that it can be tested with end users and refined without delay.
Design For Manufacturing (DFM) techniques that ensures the intended product design can seamlessly enter production - without the need for costly and time consuming modifications.

Benefit 3: Cost

Opportunity cost

By boosting your resources with outsourced design partners, you can avoid lost sales due to not having a product to market on time.
Accessing world class capabilities with outsourced electronic product design engineers you can prevent not having the right product for consumers' needs resulting in lost sales.
Hiring an outsourced team ensures focused attention and project management with greater efficiency and time management to get a stronger product to the market faster.

Learning cost

Years of experience and breadth of knowledge from design companies results in a reduced learning-though-failure costs for their clients.
Electronic product design companies are passionate about new technologies and are constantly investing in updating their technical tools and equipment and upskilling their engineers. To attain the same level of knowledge and resources for every new product and its unique technology requirements, can become a costly endeavour for product companies and can lead to wasted resources.
The profit advantage of hiring an outsourced electronic product design team hence can be even greater for complex projects. These projects might include undertaking significant amounts of research and/or developing complex electronic hardware, firmware (embedded software) or electro-mechanical systems.

Operational and labour cost

Hiring an electronic design company, like Beta Solutions, give you access to a whole team with multidisciplinary design experts which is more economical than hiring them individually.
The infrastructure costs (design, diagnostics, and test equipment, etc.) to establish an effective electronics design lab can be considerable. In-house teams may have the know-how but not the right tools and resources for the design and development of your product. This may include complex software programmes, testing and validation equipment, prototyping tools and production equipment, etc. You benefit from the electronic design company's investment in time and money to procure and learn to operate new tools and equipment.

Benefit 4: Focus and Amplified Capacity

Outsourcing the parts of product development that is not your expertise - like the electronic workings of the product - expands your product design capacity.

If your company is facing a resource crunch, an outsourced development team can expand the workforce.
By hiring an electronic product development team (and their immediate knowledge and skills) and thereby delegating non-core practices, your internal team can focus on other tasks that will ensure the commercial success of your product.
Whilst your outsourced design partner focus on research and development, testing and experimenting to get a viable product to your specification, your team can focus on the management of the product and the marketing required for launch.

Conclusion

As discussed, there are many advantages to outsourcing in the realm of technology and electronic product design. Ultimately, combining skilled outsourced resourced with your in-house efforts will lead to superior products, quicker turnaround times and speed to market. There is never a wrong time to outsource electronic product development, which could be (for example), right at the product's inception stage or perhaps somewhat closer to the product's release date. The correct electronic product design company can add value at every step of the design process, from innovative idea generation and cost reduction analysis to critical design reviews and specialist product testing.

Over the past ten years we at Beta Solutions have gathered a considerable amount of expertise in designing electronic products for many industries. We have experience with a wide range of technologies, obtaining regulatory compliance / certification, designing for manufacturing - as well as, undertaking fast (in-house) low volume production or outsourced high volume production. We have successfully partnered with both clients who required fully outsourced R&D as well as clients who outsourced only a part of their electronic product design (where a specialised service was required to complement their in-house team's work). Our purpose is to apply all our knowledge, skills and passion to get the product you need to market at the opportune time.

It always starts with a conversation ... so feel free to contact us for a free 1-hour consultation.

References

Brandon Gaille. (May 2017). 27 US Outsourcing Statistics and Trends. Retrieved from https://brandongaille.com/26-us-outsourcing-statistics-and-trends /
Banner modified image retrieved from https://pixabay.com/photos/web-network-technology-developer-3963945/
Conclusion modified image retrieved from https://www.pexels.com/photo/about-us-achievement-agreement-business-533405/

Artificial Intelligence - A Closer Look

Thu, 28 Feb 2019 00:50:46 GMT

Author: Morten Kirs

Introduction

Believe it or not, the concept of artificial intelligence (AI) has been around for quite a while. The first digital computer was developed in the 1940s, and with medicine at the time having discovered that the brain is effectively an electrical network of neurons, it wasn't long until research into developing an electronic brain was started. Perhaps due to not appreciating the full complexity of the problem, the original researchers sparked a bit too much hype and tended to over-promise on deliverables. This led to periods of funding cuts and lack of continued interest, known as "AI winters", followed by periods of resurgence whenever new developments arose. Recent developments in "deep" neural networks and developments in training through "big data" and machine learning techniques have sparked a new hype cycle. This time it appears that AI could be here to stay, as it seems to have proved that it can solve problems that humans have yet to solve.

What is AI and Where is it Used?

There is still a big difference between the application specific AI that is in the limelight today, and general AI, or the electronic brain, that has been theorised since the late 1950s. More than 60 years on we're still not expecting robots to have a hostile takeover of the world anytime soon. The application form of AI on the other hand is now everywhere. Whenever you post a picture on Facebook, complex image recognition AI will automatically find and tag you and your friends. Whenever you ask Siri or Alexa a question, speech recognition AI will translate your voice into appropriate commands. And eventually, when you jump into a car, AI will be able to take you to your destination without you ever touching the wheel.

Despite its apparent complexity, at the heart of modern AI lies a more simple optimisation problem of curve fitting. If you've taken high school math or statistics, you likely would have encountered a similar problem on a 2D graph. Given a bunch of points (x,y), what is the line of best fit that describes the data set? Once you've found the line of best fit, you can then predict the output y, given any x value.

Figure 1. Simple line fit to known sample data. At the heart of AI lies this same data-fitting principle.

AI solves a problem much like this, except with a fair bit more calculation than you would have used in high school. AI is tuned with potentially millions of data points, where each point could have many inputs and many outputs. For example, an input may be an image with each pixel effectively being an input. An output from this could be a simple yes/no decision for a question such as "Is there a person in this image?", or it could be multiple outputs for coordinates for each person found in the image. AI finds a solution model that best fits the original data set, with the model potentially using thousands of coefficients. AI can then use this model to predict the result of any future inputs. The output of a model is usually a probability rather than a direct yes or no answer.

Modern advances into "deep learning" have shifted complex problems into neural networks. Each node has its own function and solution, and the output of one node is then fed into further nodes. Splitting the problem into nodes allows it to be optimised in parallel using modern GPUs or cloud based server farms. There are currently very many architectures available for deep learning, some of which are pictured below.

Figure 2. Example neural network architecture. In between the inputs and outputs lie multiple hidden layers.

Training AI and Machine Learning

Since the early days of AI, board games have been a simple but effective way of determining AI "intelligence". One of the first AIs in the early 60s was, in fact, a checkers program that managed to beat experienced amateurs. Today AI is pretty much unbeatable in games like this and is able to teach itself games at astonishing rates. An example of this is Google's AlphaZero chess engine, which trained itself to play chess in under 4 hours to a superhuman level by simply playing itself over and over ^[1] . Given more time, a 9-hour trained engine was able to defeat the most notable engine of the time, Stockfish 8.

Training and machine learning is at the heart of AI. AI needs a lot of previously categorised data to learn from. For board games this is relatively simple, as the engine can effectively play itself and learn along the way. There is no consequence to losing a board game. Clearly this is not the correct approach for all scenarios. In a fully autonomous vehicle mistakes are critical, and needless to say a trial and error approach to learning driving could be deadly. An autonomous vehicle AI actually learns from a human driver as an example. For an AI to be effective at driving it requires thousands of hours of observed inputs.

Extensive training is the reason why AI might not be applicable to all problems. Facebook and Google have access to millions of images, data points, etc. to train their AI algorithms. Obtaining thousands of data points and categorising these for your needs is very difficult and time consuming for a smaller company that's just starting out. However, as long as the function is non-critical, an AI system may be able to train itself over time, given a pre-built model of similar data.

Figure 3. A realistic picture on what AI is all about. A large part of AI "coding" lies in getting and categorising sufficient training data.

The Magic of AI

There are by now several examples of AI being used in medicine to predict patient current and future ailments. A prominent example of this is Deep Patient, an AI which has learned to predict conditions like liver cancer and is also strangely good at diagnosing complex disorders like schizophrenia. It seems to do this far better than most doctors can. How does it do this though? As of today, the answer is, no one knows. Deep Patient, much like most AIs, is a complex network of neural pathways that have been tuned through the use of a huge data set. The output is currently a result of optimised math that makes no sense to someone looking at the final system and is a prime example of one of the issues with AI. Although an AI has found a solution to the problem, we currently have no concrete way of verifying the end result's reliability, nor of any prejudices in the system due to biased data.

Google's Ali Rahimi, an award winner at the 2017 Conference on Neural Information Processing (NIPS) ^[4] , likened the current state of AI to the medieval magic of alchemy. Some things just seem to work. In medieval times these examples were metallurgy and glass-making. The processes around these examples worked like a charm when using the scientific knowledge of alchemy at the time, whereas other things like turning common metals into gold simply didn't. AI is much the same at the moment, where some architectures, processes and optimisations seem to work for specific problems, whereas other attempts at solving problems though AI do not. There is currently no full understanding of the processes that underlie machine learning. For alchemy, we now understand that magic is not the right explanation, but we still need to fully figure out the mysteries of AI.

An interesting attempt at understanding the inner workings of AI has been made by Google in its Deep Dream project ^[2] , which I'd encourage the reader to check out. It is a tool for finding things like faces in images, but can also be run in reverse to see what an optimised and biased AI might be looking for. The result is strange images that are now classified as a new art form.

Figure 4. Example image from Google's Deep Dream project. An example of what AI may see when it tries to look for patterns.

Problems with AI

Deep learning and AI does have its downsides, and there are a few key reasons for this, which have been outlined by Gary Marcus, CEO of Geometric Intelligence (an AI company now acquired by Uber) ^[3] :

AI is greedy
It requires huge sets of training data.
AI is brittle
If the AI is transferred and confronted with scenarios which differ from training data, the AI is very likely to break.
AI is opaque
AI is effectively a "black box" whose outputs cannot be explained. Parameters of neural networks exist in mathematical geography and are not debuggable like traditional programs.
AI is shallow
AI does not have any innate knowledge and holds no common sense checks about the world.

There is current research into "Artificial General Intelligence" which might make AI less shallow and less brittle by providing a basic model of the world, but such a feat seems incredibly complex. Research is continuously also trying to fully understand how to best optimise and architect AI, which is likely to make it less opaque and less greedy, if we knew exactly what data needs to fed to AI. However, for the moment it seems we are a little way off from any of these goals. We could experience another AI winter before these problems are solved.

There are also a few social arguments against AI. An extreme example of this is an AI that was used in prison systems ^[5] . The goal of the AI was to try and predict which prisoners were likely to re-offend once they were released. As people of certain races are over-represented in prison populations already, it appeared that the AI ended up having a heavy racial prejudice itself. It seems as though prejudices held by humans have also been learned by this AI. This general human bias likely extends to all forms of AI, such as that which determines your search results and the the advertisements you see online. As the AI has been trained by a multitude of previous users, there is likely to be an inherent bias in this data to a general worldview. The full extent of the bias, as argued above, is completely unknown as the AI behaves as a black box. There is caution in the community that AI might be breeding intolerance, may be influencing elections, and could be damaging individual and original thought in general.

Embedded AI

AI has ingrained itself in the embedded world. Currently the common AI approach has been to send data from a simple IoT device to the cloud, where the more complex and involved AI and machine learning algorithms run. This is because AI algorithms often require a lot of computational power. Once trained on more powerful computers, the AI can be transferred onto more local specialised hardware. IC developers are currently working on integrating AI accelerators into the silicon of microcontrollers, which should eventually allow for faster processing and better allow these algorithms to run and train locally.

Conclusions

AI is now everywhere, and it has proved itself to be a powerful tool that has helped us solve or given insight to a variety of problems. The black box nature of AI and our lack of thorough understanding of it means that critics think of it as unreliable and dangerous (Elon Musk has stated it's more dangerous than nukes ^[6] ). Nevertheless, AI seems to have found its niche in the market for non-critical applications where it is able to find patterns and make predictions far better than most humans.

If you have a problem which needs pattern recognition then it may be worth considering incorporating AI into your next project. Contact us to talk to someone in our knowledgeable team.

References

AlphaZero Engine https://en.wikipedia.org/wiki/AlphaZero
Google's Deep Dream Generator: https://deepdreamgenerator.com/
Greedy, Brittle, Opaque, and Shallow: The Downsides to Deep Learning - Jason Pontin:
https://www.wired.com/story/greedy-brittle-opaque-and-shallow-the-downsides-to-deep-learning/
Ali Rahimi's talk at NIPS (NIPS 2017 Test-of-time award presentation) https://www.youtube.com/watch?v=Qi1Yry33TQ E
MIT Technology review: AI is sending people to jail—and getting it wrong - Karen Hao:
https://www.technologyreview.com/s/612775/algorithms-criminal-justice-ai/
CNBC: Elon Musk: ‘Mark my words — A.I. is far more dangerous than nukes’ - Catherine Clifford: https://www.cnbc.com/2018/03/13/elon-musk-at-sxsw-a-i-is-more-dangerous-than-nuclear-weapons.html
Figure 3 retrieved from Common mistake: AI is all about building neural nets, https://hackernoon.com/%EF%B8%8F-big-challenge-in-deep-learning-training-data-31a88b97b282
Machine Learning & Artificial Intelligence image via www.vpnsrus.com
Banner image retreived from https://pixabay.com/photos/artificial-intelligence-robot-ai-ki-2167835/

20 Things to Consider When Planning an IoT Solution - Part 2

Thu, 17 Jan 2019 00:35:46 GMT

Author: Aaron Fulton, Systems Developer

Part 2

This post is Part 2 of a 2-part series on what to consider when planning an Internet of Things (IoT) solution.

As we discovered in Part 1, IoT is a complex subject which crosses many disciplines ranging from embedded electronics through to communications, cloud infrastructure, and software design. Architecting an IoT solution is a multifaceted business, there are many things to consider. For most organisations, working with companies who have IoT experience is critical to architecting and implementing a solution which works on all levels. IoT crosses 4 key domains: Devices, Communications, Cloud/Server infrastructure, and Applications.

In Part 1 we looked at some higher-level aspects to consider. In this article, we'll take a deeper look at some of the more technical hardware-related aspects.

11. Power & Battery Life

Power is a major consideration for IoT devices. Electronics need power to operate and this must come from somewhere. There are really only three mainstream options: mains power, solar power, and battery power. Where the IoT solution is deployed will dictate what type of power is available, and it is from this that a lot of other decisions are based. Sensors require power, computation requires power, and communications require power. Typically IoT devices are battery powered and need to last a really long time (many years). This means that the battery has to be really good and the power consumption needs to be really low. This presents an engineering challenge. Batteries not designed for this purpose can suffer from self-discharge and this can dramatically affect the life expectancy of the solution. There is a huge range of batteries on the market now and choosing the right one is very important.

12. Communications

The communications platform is a key decision. When it comes to communications there is a triad of inter-related factors. Range, Bandwidth, and Power consumption. Each communications technology has different range characteristics. Some technologies can jump directly onto the internet while others need some kind of hub or gateway. It is important to choose the communications protocol that is fit for purpose.

13. Transmission Rate & Quality of Service

Another important aspect of communications is the transmission rate and quality of service. Protocols that operate on free radio spectrum have "good behaviour" restrictions placed on them. These restrictions are like "fair use policies" and limit how often devices are allowed to transmit. As a consequence, the amount of data that can be sent from protocols such as LoRa or Sigfox is limited to at most a handful of small bursts of data every few minutes. For sensors that just need to send in periodic readings such as measuring the office temperature, this is perfectly adequate but for more critical applications (such as monitoring the temperature in a cool store) this is not good enough.

How important is it for messages to get through? This is where quality of service (QoS) comes in. Due to the "good behaviour" restrictions, it is not feasible to send acknowledge messages back to the devices after every transmission. This means that often data needs to be sent with the hope (not guaranteed) that it will be received. For applications that do not require every single piece of data to make it through the journey that is fine, but for more critical applications the quality of service can be a real issue. Not all protocols suffer from the quality of service problems, so careful selection is needed.

14. Two-way Communications

Do you need to talk to the device? If so, how real-time does that need to be? The main way that devices are made low power is by sleeping. Much like with humans, talking to devices in their sleep is not a very effective way of communication - you need to make sure they are awake first. Low power devices wake themselves up and do what they have to do (usually taking a reading from a sensor and transmitting it). There is then usually a small window after the transmission when low power devices "listen". It is a really difficult engineering problem to have low power devices for applications that need to control stuff in a timely manner.

Figure 1: Communications Networks Comparison spider graph as of July 2018 (Graph made by BSL)

* Commercial networks are not yet fully deployed therefore coverage for NB-IoT and LoRa WAN is marked as low. For the same reason, some other evaluations are based on information at hand rather than measured performance.

15. Network Selection & Availability

Communications networks are not ubiquitously available. Cellular providers use terms like "Our mobile network covers over 98.5% of the population" which is code for, we've got the main centres covered really well and rural coverage is spotty. It is likely that LoRa WAN, NB-IoT and SigFox will be similar in their approach. It only makes economic sense to deploy to places where there are enough people to make the network financially viable. Some networks like Wi-Fi and LoRa can be rolled yourself but with that comes the burden of maintenance.

16. Remote Updates

How good is the firmware on the devices? Will it ever need to be updated? Updates might be necessary to fix bugs, apply security patches, and add new features. Due to the limited downstream capabilities (server to device) of low power networks, remote firmware updates at this point in time are not possible. This means that the firmware on the devices needs to be rock solid at the time of installation. If remote update capability is needed, then a higher power communications protocol is required

17. Data Storage Policy & Quantity

IoT devices can create a lot of data. Consider a device that takes one single reading every 10 minutes; that is 4,300 readings every month. By the time data is stored in a database this would be around 2.5 MB of data growth every year. Now multiply that by a deployment of 1,000 sensors and you have 2.5 GB of database growth every year.

Now, if instead of one reading every 10 minutes, the application had 10 sensors. That is now 25 GB of database growth every year. While this in-and-of-itself is not a problem, but it must be planned for. Databases can become a challenge when they get this large. Cloud infrastructure is really good at dealing with scale and should be the default go-to for any large IoT application.

Data storage policies are also important. How long should the data be kept for? Does all the raw data really need to be kept for 10 years? Or can some of it be deleted sooner?

18. Data Quality, Accuracy, Integrity

Data quality must also be considered. Is near enough good enough or is a high degree of accuracy needed. For applications measuring consumables like water, power or gas, the data must be good enough to bill from. Sometimes sensors or devices can have a bad day and send in poor quality data and this must be considered, for example, something as simple a spider crawling into a height-measurement sensor could cause a blip in readings. Accuracy is also a consideration. Do readings need to be good to 10 cm or 1 mm? Data quality, like all quality, is something that does not happen by chance. It must be engineered in.

19. Edge vs Cloud Compute

There is a trade-off between putting compute on the edge (where the devices are) or in the cloud. On one hand, it is really convenient to do all the data processing in the cloud, the software can be updated easily, and all the data is kept centrally. However, there is no off-line capability when there is invariably a network outage. Doing all the processing in the cloud also means more data to transfer (which comes at a cost) and more cloud compute costs. On the other hand, doing as much processing on the device as possible means that there is a lot more resiliency to internet outages and the devices are more self-managing. Doing computing on the devices is also free as the processor is probably not fully utilized. However, edge computing does not work if one of the constraints is to conserving as much resource (battery) as possible to prolong life.

In practice, a compromise between the two needs to be reached. Usually, there is some data processing capability on the devices, but anything complex gets outsourced to a server in a far-off data-centre.

20. Sensors

Finally sensors. What would IoT be without sensors? Sensors are why the Internet of Things exists in the first place. Without a way of capturing interesting data the rest of the system is redundant. Good quality, robust and resilient sensing is one of the hardest parts of the whole IoT thing. While many things are interesting to measure, they are not always easy to measure. Sensors have to work in the harsh environment of the real world and need to be engineered appropriately. There is usually a few options for sensor types which could be used to measure something, each with their own set of pros and cons. Talking to sensing experts like Beta Solutions is a really important step in putting together a successful IoT project and getting the right data.

References:

Photographic images retrieved from https://www.pexels.com

20 Things to Consider When Planning an IoT Solution - Part 1

Thu, 15 Nov 2018 00:27:59 GMT

Author: Aaron Fulton, Systems Developer

Part 1:

You have heard of IoT and can see the benefit of sensors and data collection, but don't know where to start... You are not alone, many entrepreneurs feel this way, so this post is part 1 of a 2 part Blog series on what to consider when planning an Internet of Things (IoT) solution.

IoT is a complex subject which crosses many disciplines ranging from embedded electronics through to communications, cloud infrastructure, and software design. Architecting an IoT solution is a multifaceted business, there are many things to consider. For most organisations, working with companies who have IoT experience is critical to architecting and implementing a solution which works on all levels. IoT crosses 4 key domains: Devices, Communications, Cloud/Server infrastructure, and Applications.

In this Blog (Part 1) we will look at higher-level points to consider.

1. What are the Benefits

It may seem like an obvious question to ask, but the most important thing about planning an IoT project is to clearly understand the benefits and how they can impact/profit your business.Benefits can take many forms:

A common benefit could be making better decisions through the use of data . Sometimes important decisions have historically been made based on intuition, gut feel, or forecast from historical trends. With the inability to collect real-time data, this was the best that could be achieved. With IoT technologies this changes as real-time data becomes available. Decisions can be made based on measurements rather than hunches and as a result will be higher quality. More data enables better long-term planning as well.
Quality of service is another common benefit of IoT, particularly in service industries. Consider a commercial distribution system which needs to proactively keep their customers from running out of products. This is especially difficult if the customer consumption rate is variable. An IoT solution can be used to monitor each customer's inventory in real time, allowing for the supplier to service the customer much more effectively.
Efficiency is a third major benefit. Getting real-time information means that systems can be optimised, which leads to increased efficiency.
Sometimes the data can be on-sold to partners. From a financial perspective, this can certainly be a benefit!

The internet is also full of IoT solutions to problems that don't exist. Sometimes technology can look enticing, but without clearly understanding the benefit the technology brings, the solution could be a waste of money. For instance, do you really need an internet-enabled toaster?

2. Data Ownership & Sovereignty

Data is a huge asset in today's digital economy. Who owns IoT data? This might sound like a straightforward question if you are gathering data for yourself about yourself, but there are many scenarios where this gets more complicated. Consider a company which supplies grain to a farm. The supplier installs a monitoring device on a grain silo to measure the volume and schedule deliveries. The data set is as much about the farm as it is about grain logistics. Does the data "belong" to the grain supplier or the farmer? Or is it shared?

It is important to work out the policies around who owns data, what is shared, and what each party is allowed to do with the data.

The other dimension to data sovereignty is by which laws it is governed. Data kept off-shore is subject to the laws of the country in which it is kept. For example, there are penalties in some countries for significant data breaches, or the data might be subject to being accessible to local law enforcement if it is deemed appropriate (watch out for the NSA). Services which European citizens access need to comply with the EU General Data Protection Regulation (GDPR) regulations.

Once sorted out, data ownership & sovereignty can be spelt out in terms and conditions documents or other agreements, but it deserves some serious consideration.

3. How Will the Data be Analysed?

OK, so you've got some IoT data streaming in. Now what? Data is really powerful if used correctly. It is really easy to gather data but is a much harder proposition to extract out the maximal amount of information from it. Some data is designed to be used for real-time monitoring while other data is put into a data lake for later analysis. There are many many ways of slicing and dicing data. Consulting a data science expert who understands the range of tools and techniques is often a useful exercise.

There is often great insight to be had from cross-correlating different data sets and using the information for predictive and forecasting purposes. There is a famous example of how the Target stores very successfully correlated purchases across 25 products to estimate a pregnant woman's due date ⁽¹⁾ .

The use of neural networks (AI) or machine learning techniques has become much more popular in recent years, as the field has matured and the tools much more powerful and models much better.

4. Visualisation of the Data

Being able to visualise the data is another really important aspect of an Internet of Things solution. Row upon row of raw data might be interesting to a computer, but is not the optimal way to present data to a human. Summarising and displaying data in some kind of graphical form is much more useful.

Visualisation is important for:

Finding trends in data.
Finding anomalies in data (or points of interest).
Seeing the data in context.

5. Criticality

Most systems are non-critical. But some are. Critical systems require special attention when it comes to design. With non-critical systems, it doesn't matter too much if there is a failure or fault. A failure is more of an annoyance rather than anything catastrophic. However, with critical systems, it is a different story.

There are four types of criticality:

Safety Critical - Where failure or malfunction may result in death or serious injury to people. Examples include fuses, fire alarms, emergency systems.
Mission Critical - Where failure or malfunction will cause cascading problems across the entire business or affect many businesses or organisations. Examples include aeroplane navigational systems, banking systems, and the power grid.
Security Critical - Where failure or malfunction will mean private, personal or classified information is breached. Examples include systems that provide encryption and transmit personal or classified information.
Business Critical - Where failure or malfunction will cause some parts of the business to cease functioning but will not cause a cascading effect. Examples include key bits of software used by specific departments.

If the IoT solution is in any way critical, special care must be taken in the design. Risk planning (through the likes of a Failure Mode and Effects Analysis (FMEA)) and a lot more engineering will be required to make sure that the solution is sufficiently robust. Calling on experts with experience in designing critical systems and conducting Risk Assessments and Failure Mode and Effects Analysis is highly recommended.

6. Security

Security is always a significant consideration in this day and age. With IoT solutions there are two key types of security to consider: Network security and Physical security.

Any decent enterprise IoT solution will have security baked into the solution, this is not the case for every provider though. Sometimes security gets forgotten. There have been examples of internet-connected security cameras being compromised and being conscripted into a bot-net attack, or cars being controlled remotely.

Luckily many IoT solutions are inherently much more secure in terms of their transmission, with all data being force-encrypted, also much of the communications happening on private networks. Wi-fi and internet-connected devices are most at risk of security vulnerabilities. Low-power IoT devices have very limited capabilities and spend most of their time asleep which greatly decreases the "attack surface" for anything malicious.

Physical security also needs to be considered. IoT devices and sensors sometimes need to be in locations which are easy to access. This opens up the possibility of the sensors being tampered with, especially those in public places.

7. Total Cost of Ownership

Three of the defining characteristics of the Internet of Things are the sheer number of devices, the length of time they are deployed for, the low total cost of ownership. The whole life cycle of the product must be considered however: procurement, installation, maintenance, and removal. If thousands are being deployed this cost can mount up to a significant amount of money.

When working out the total cost of ownership of IoT devices it is important to take into account:

What it will cost to install the devices
How often the devices will require maintenance (changing out batteries, fixing faults, performing upgrades)
How much replacement batteries cost
The cost of transmitting the data
The cost of storing & analysing the data
Any costs associated with decommissioning devices

8. User Experience

Why collect data in the first place? Well, it is usually for some human to use to make decisions. It may not be surprising then that good user experience design plays a big role in the success of an IoT project. It is important to consider who the data is for (stakeholders), and how it will be used. A poor user experience will mean that people won't engage with the data and use it to its full potential, thus wasting a lot of opportunity IoT brings.

User experience is most important when selling the data as a service. The interface is the part that customers will be interacting with on a day-to-day basis and is how they will judge the solution. For machine-to-machine communication and integrations, the "user" experience takes the form of a well-designed API.

In 2014 Google acquired the Nest for $3.2 billion. Nest was one of the first internet-enabled thermostat controls, what set them apart however and made the company so valuable was their exceptional user experience design.

9. Density

IoT devices need some way of getting back to the internet. Protocols such as Bluetooth Low Energy, Zigbee, and LoRa need to go through some form of base-station or hub which bridges them to the internet. Often it is only cost-effective to deploy these base stations if there is a reasonable density of devices to connect to them. Some protocols such as LoRa WAN or Sigfox require a network provider to deploy the network, and for these operators, the success of their business is all about a high density of devices.

On the other hand, if the device can hook into a local Wi-Fi connection or be plugged into a physical network, density is not so much of an issue.

10. Maturity of Technology

IoT technology, as a field, is still maturing. A lot of IoT sensing applications are technically very difficult problems to solve and there is still lots being learned on all fronts. IoT technology is certainly on a pathway to maturity but is not there yet. It is an exciting time to get involved in IoT and be an early adopter. There are a lot of benefits to be had from IoT but it is good to consider slow adoption to learn what works well for your needs before going large.

Conclusion

There is a lot to consider when it comes to deploying an IoT solution. Talking to experts such as Beta Solutions is a good idea to get you on the right track.

Subscribe or look out for Part 2 that will complete the series of 20 Things to Consider When Planning an IoT Solution.

References:

Hill, K. How Target Figured Out A Teen Girl Was Pregnant Before Her Father Did. Retrieved from https://www.forbes.com/sites/kashmirhill/2012/02/16/how-target-figured-out-a-teen-girl-was-pregnant-before-her-father-did/#2aeed4ef6668
Vector - In Love with Technology: Vectors by a href="https://vecteezy.com">vecteezy.com
Vector - In Love with Technology: 2 Free Vectors via vecteezy.com
Vector - Technologia Free Vectors via Vecteezy
Photographic images retrieved from https://www.pexels.com

Formula E - Racing into the future

Thu, 18 Oct 2018 00:09:43 GMT

Author: Phillip Abplanalp, Hardware Engineer

Introduction

Motor racing with the absence of screaming engines and grid girls would have been a recipe for catastrophe only 20 years ago. But here we are, 2018, and this is exactly what Formula E (FE) offers and it seems be working, really well in fact – although it seems the removal of the grid girl has resulted in quite a saga! For those that may have missed the memo, Formula E is a single seater, fully electric racing series, competing at a range of street circuits around the globe. Started in 2014, it’s seen a rapid growth in both viewership and automotive manufacturer interest. Approaching its 5th season, FE boasts a staggering number of automotive manufacturer entries including: BWM, Audi, Nissan (Renault group), DS Automobiles (Citroen-Peugeot group), Mahindra (large Indian car company) and Jaguar. Two further mega manufacturers are set to join at the end of 2019 - Porsche and Mercedes. For this reason we at Beta Solutions thought it was high time that we take a deeper look at what all the commotion is about!

What is a Formula E car?

Looking at the exterior of a FE car, a lay person would be forgiven for confusing it with almost any other single seater formula car. With the typical single seater concoction of front and rear wings, open cockpit, central roll hop and radiators, these are all features synonymous with those of say a Formula One car. However, even here the under lying philosophies of the car's aerodynamics are actually significantly different. Whereas the aerodynamic surfaces on Formula One cars are primarily designed to generating copious amounts of downforce (at the expense of drag), FE is far more focused on reducing drag (at the expense of less downforce). These differences in aerodynamic philosophies stem from the very core of the FE series - its energy storage medium and the need to preserve energy.

Figure 1: 2018 Ferrari Formula 1 Car

Figure 2: 2018-2019 Formula E Car

But as the old adage goes, it’s what’s inside that counts, right? So what's under the skin of a Formula E car? Read on ...

Essentially a FE car is a battery powered electric vehicle which is made up of four critical electrical components, these are set out below:

Electric Motor/Generator

This component is responsible for both propelling the car under acceleration and recovering energy under braking. Typically the motor/generator is some configuration of a Permanent Magnet Synchronous Motors (PMSM) - which is in essence a permanent magnet 3-phase AC (Alternating Current) motor – similar also to Brushless DC (Direct Current) motors. This section of the regulations are open to all teams to develop and try find the most optimal solution.

Inverter

This component takes in the DC current supplied from the battery and turns it into a 3-phase AC used to drive the motors, as well as controlling the motors speed based on the input from the driver. The Inverter, like the electric motor/generator, is open to development.

DC to DC Converter (power supply)

The DC to DC power supply is responsible for converting the 500 VDC (Gen 1 battery voltage) to a more usable voltage range such as 12-16 VDC (McLaren MCU-500 ¹ ) for all the auxiliary components like the numerous sensors on the car or steering wheel computer. Some inverters (such as McLaren MCU-500) will include both the DC to DC power supply & inverter all as one single component.

¹ McLaren MCU – 500 is an example of an Inverter similar to those used on an FE car. The exact components are not publicly available.

Battery

The battery is the component which stores electrical energy which the car uses throughout the race to both propel it and also power the numerous on board sensors. This component is a “standardised” part that is provided by an FIA (Fédération Internationale de l'Automobile) selected manufacturer. For the first four seasons, the battery pack was supplied by Williams Advanced Engineering and from season five it will be supplied by McLaren Applied Technology.

The diagrams below show a high level representation of the interconnection of the components discussed above.

Figure 3: High level system diagram – Gen 2 Formula E car.

Figure 4: Gen 1 car with showing the general layout packaging.

Formula E Technical Evolution

As part of the 5th season of FE, the series is rolling out its second generation (Gen 2) car. Along with some exterior cosmetic changes to the body work, the cars have undergone a significant overhaul of some of the key technical areas. The most significant of these changes are outlined in the table below as follows: Note: At the time of writing, the full FIA Technical regulations for 2018-2019 had not be made public.

Figure 5: Gen 1.

Figure 6: Gen 2.

Perhaps the most significant change between Gen 1 and Gen 2 cars is the removal of the in race ‘car change’. Under the Gen 1 regulation (first four seasons) it was mandatory for all drivers to make a ‘car change’ pit stop which required drivers to change from a FE car with a depleted battery to a new fully charged FE car. This was essentially due to the Gen 1 battery pack (at 28 kWh) not being large enough to last an entire race (~50 mins) and the swapping of batteries during the race were deemed to be too dangerous and complex. To allow the Gen 2 FE cars to last an entire race on a single charge the battery capacity has been raised from 28 kWh to 54 kWh. While it is great to have had such a large energy capacity increase, the knock-on effect has been an increase in the battery pack weight from ~200 kg (Gen 1) to 385 kg (Gen 2). This unfortunately means the overall energy density of the battery packs have only slightly increased over the intervening 5 years ² !

² Gen 1 = ~ 140 Wh/kg to Gen 2 = 140.2 Wh/kg

Possibly the second largest change being implemented for the Gen 2 regulations is the increased power allocation for regenerative braking. Regenerative braking power limit essentially determines the amount of energy that can be recovered during a braking phase, the higher the amount set by the regulation means less wasted energy and therefore an increased overall system efficiency. The Gen 2 regulations will allow an increased regeneration power of more than 60% over the previous Gen 1 cars, going from 150 kW to 250 kW. Interestingly enough under racing conditions this will mean that the Gen 2 cars will for the first time have the ability to recover/regenerate energy at a higher rate than what is allocated for acceleration (250 kW (regenerate) vs 200 kW (accelerate)). Perpetual motion, right? Not quite, but it’s likely to allow for energy recover in the range of ~1 kWh (~3.6 Mj) per lap, potentially meaning that as much as 50% of the used energy is being recovered over the lap ³ .

³ Calcs are based on a 100 s lap, 15 s braking duration per lap at the full 250 kW recovery during the entire braking phase, 50 min race length and 54 kWh battery.

Technology Transfer – From the Race Track to the Road

Over the past decade the development of road relevant technology on the race track has become somewhat of the holy grail for motor sport regulators. This has become increasingly important as society demands motor sport to be not just entertaining but also environmentally sustainable! Sustainability in the context of motor sport has to be automotive manufacturers using it as a tool for developing the road cars of tomorrow.

So, if the vision is to assist with the development of future electric vehicles, how is FE doing? Well in my opinion, both good and bad (or should that be positive and negative!). Perhaps on the bad side is the closed nature of the existing FE regulations for arguably the most important component to the success of electric vehicles. This of course is the battery, and more importantly their present lack of energy density. For a number of reasons including to limit excessive spending, stopping runaway scenarios (where one team has gained a significant advantage) and to encourage development in other areas the FIA (FE regulators) have opted for “standardised” batteries for FE. This means an identical battery is supplied to all teams, not allowing individual teams to try develop their own optimised solution. However without a high level of competition in this area, the development of the batteries in FE is likely to show very little progress and this is already visible with essentially zero gains to energy density having been achieved over the 5 seasons of FE! If you are thinking energy density isn’t that big a deal, consider for comparison, the energy density of these batteries compared with petroleum powered combustion engines. Namely, the battery packs on these Gen 2 FE cars will have energy densities of only ~1% of that of petroleum ⁴ ! This is a gigantic difference and even when considering the potential for energy recovery and the significant efficiency differences between a combustion engines and electrical equivalent (20-35% vs 80-90% respectively) the total energy storage for an electrical system is still likely to be ~30x heavier to do the same work. What’s more disappointing is that the regulations are unlikely to allow any battery development by teams until at least 2025. So, if you are after the state-of-the-art in battery technology it is unlikely that this will be found in FE before 2025!

⁴ petroleum=~12.8 kWh/kg and Gen 2 Battery = 0.14 kWh/kg

On the good side, is that the present regulations are very open on the drive train and power electronics of the car with both the electric motor and the inverters being free to develop. With teams partnering with the likes of semiconductor giant Rohm, significant gains have be made on both the physical packaging size, weight of the inverters and a reduction in energy losses. These gains can be seen in the image below:

Figure 7: Design evolution of inverters.

These improvements have in large come from the developments Rohms has made in the area of silicon carbide (SiC) MOSFETs used for the switching of the inverter. These developments have been implemented in a two stage implementation, starting in season 3 with a hybrid solution of standard silicon IGBTs (Isolated Gate Bipolar Transistor) but SiC based Schottky Barrier Diodes (SCB) rather than the conventional pure silicon based diodes. SiC Schottky Barrier Diodes offer reduced reverse leakage (at high switching frequencies) through faster reverse recovering times and also reduced package size, resulting in the inverter being having lower switching losses and reduced size and weight. Season four saw the implementation of full SiC modules with both SiC MOSFET and Schottky Barrier Diodes, bringing further reduction in high speed switching losses, resulting in increased efficiency and smaller packaging. See overall switching gains in Figure 8 and a Half Bridge SiC Module in Figure 9 below:

Figure 8: SiC MOSFET Reduction in switching losses vs IGBTs.

Figure 9: Half bridge using SiC MOSFETs and Schottky Barrier Diodes.

Conclusion

While there are improvements to be made to the series such as opening battery development, it is a racing series with a positive future and one that is rapidly gaining traction, both with fans and manufacturers! So be sure to join the electrification movement and tune into the first race of season 5 on the 16th of December (2018).

Feel free to pop in for a coffee and join our lively discussions about the technology and the driving strategies as the seasons progress.

References:

Figure 1, 2018 Ferrari Formula 1 Car Retrieved from https://formula1.ferrari.com/en/2018-f1-chinese-gp-free-practice-1/
Figure 2 and Figure 6, 2018-2019 Formula E Car. Retrieved from http://formula-e.wikia.com/wiki/Spark_Gen_2
Figure 4, Blueprints Powered by Mahindra. Retrieved from https://current-e.com/blueprints/powertrain-explained-m3electro/
Figure 5, Gen 1 - Felix Rosenqvist at the 2017 Berlin ePrix. Retrieved from https://en.wikipedia.org/wiki/Formula_E
Figure 7, Design evolution of inverters. Retrieved from http://micro.rohm.com/en/formulae/sic.html
Figure 8, SiC MOSFET Reduction in switching losses vs IGBTs. Retrieved from https://www.rohm.com/sic/full-sic-power-modules
Figure 9, Half bridge using SiC MOSFETs and Schottky Barrier Diodes. Retrieved from https://www.rohm.com/sic/full-sic-power-modules
Vector Racing Flags Illustrations by https://www.vecteezy.com/
Vector of Racing Car Layout https://www.vecteezy.com/
Vector of E Graphics by: https://www.vecteezy.com/
Vector F1 Car Vectors https://www.vecteezy.com/
https://www.google.co.nz/imgres?imgurl=https%3A%2F%2Fcurrent-e.com%2Fwp-content%2Fuploads%2F2016%2F12%2FBlueprints-Mahindra-Racing-December-2016-M3Electro-powertrain-cheat-sheet-Current-E.jpg&imgrefurl=https%3A%2F%2Fcurrent-e.com%2Fblueprints%2F&docid=UK-4Ws54J20A4M&tbnid=DtLlvlkJoQESZM%3A&vet=10ahUKEwi0kciGudzdAhXGa7wKHcpcCh0QMwg_KAEwAQ..i&w=2667&h=1667&bih=938&biw=1920&q=mahindra%20racing%20blueprints&ved=0ahUKEwi0kciGudzdAhXGa7wKHcpcCh0QMwg_KAEwAQ&iact=mrc&uact=8#h=1667&imgdii=DtLlvlkJoQESZM:&vet=10ahUKEwi0kciGudzdAhXGa7wKHcpcCh0QMwg_KAEwAQ..i&w=2667
http://micro.rohm.com/en/techweb/knowledge/sic/s-sic/03-s-sic/4360
https://www.fujielectric.com/products/semiconductor/model/sic/box/pdf/MTET0-3160-EN.pdf
https://www.rohm.com/electronics-basics/sic/what-are-sic-power-modules
https://www.motorsport.com/formula-e/news/fe-battery-supplier-competition-987025/1382880/
https://www.infineon.com/dgdl/IISB_SiC_Studie_Part2_v2.pdf?fileId=5546d461580172fe0158249537a00222
https://www.williamsf1.com/advanced-engineering/what-we-do/case-studies/formula-e
https://innovation-destination.com/2018/02/21/faster-longer-range-e-racer-debuts/
http://www.fiaformulae.com/en/championship/regulations/
https://www.fia.com/news/fia-formula-e-gen2-unveiled-geneva
https://www.scribd.com/document/362674734/Model-3-epa#download&from_embed

The 4th Industrial Revolution

Thu, 06 Sep 2018 00:47:30 GMT

Author: Brenda Wormgoor, Marketing and Operational Manager

“The possibilities of billions of people connected by mobile devices, with unprecedented processing power, storage capacity, and access to knowledge, are unlimited. And these possibilities will be multiplied by emerging technology breakthroughs in fields such as artificial intelligence, robotics, the Internet of Things, autonomous vehicles, 3-D printing, nanotechnology, biotechnology, materials science, energy storage, and quantum computing.”(1)

- Klaus Schwab, Founder and Executive Chairman, World Economic Forum

The avalanche of disruptive technologies that is fast changing the world as we know it, is blending the physical, digital and biological worlds. This phenomenon is called the Fourth Industrial Revolution.

So, how did this evolve?

Figure 1: Timeline depicting the development of the industrial revolutions.

What does this mean?

Society and economic systems are changing as the fourth revolution is evolving at an exponential rate. As advances in technology cause rapid change and transformation in numerous industries, the blurring of technology into every part of our lives and business practices is becoming the norm.

There are many documented potential threats, such as:

Businesses failing to respond, adapt and innovate in time.
Governments failing to regulate the technologies appropriately.
Increased security and privacy risks.
Increase in inequality as automation and robotics threatens low-skill jobs whilst artificial intelligence (AI) and technology drives demand for high-skill jobs.

However, just as society and businesses had to adapt to the consequences and opportunities created by the previous three revolutions, we today need to choose how we are managing the challenges and embracing the opportunities of Industry 4.0. By intelligently employing and harnessing the benefits and potential of smart technologies and data we can drive positive change.

It is critical for New Zealand businesses to respond to this rapidly changing environment and ceaselessly innovate:

Finding new ways to serve existing customer needs.
Using the global digital platforms to speed up research and development, facilitating speed to market, agility and competitiveness (price and quality).
Anticipating and understanding changing consumer needs as a result of smartphone enabled platforms that are causing consumption and behaviour changes.
Using data to better design products whilst improving marketing and finding/creating new distribution channels.
Adapting management and business strategies to be agile and responsive to change, with new leadership styles and entrepreneurial intelligence.
Questioning and redefining strategies and business models whilst investing in continuous internal innovation and skills training.

In the informative video The Fourth Industrial Revolution ⁽²⁾ by World Economic Forum, William McDonough (Stanford University) says, “The goal is really about a diverse, safe, healthy and just world with clean air, clean water, clean soil, clean energy.”

One example of a New Zealand organisation that is pioneering the way in blending the physical, digital and biological worlds for the greater good is The Cacophony Project .

The Cacophony Project is built with a wide variety of technologies including embedded systems, web applications, backend services and machine learning (artificial intelligence) in order to solve the problem of invasive predators in New Zealand ⁽³⁾ , by:

Luring invasive predators with sound and light,
Observing predators using a thermal camera,
Identifying predators automatically using machine learning algorithms,
Eliminating positively identified predators,
Monitoring the bird song over time to measure the impact.

Figure 2: The Cacophony Project’s System ⁽⁴⁾

We, at Beta Solutions, are very excited to have recently partnered with the Cacophony Project to produce their embedded platform, the Cacophonator, and are looking forward to further future involvement.

The engineers at Beta Solutions are experienced and well versed in many current and emerging technologies. We are constantly researching, discussing, employing and testing new technologies and are passionate about finding the right technology for our clients’ new electronic products.

Conclusion:

Be part of the revolution! Feel free to contact us, call us or pop in to discuss how technology can improve and future proof your current product, or enable your new product idea to have the competitive edge.

I leave you with this quote from the The Fourth Industrial Revolution video mentioned above:

“ We need to take responsibility at every level of society... to adapt to these technological challenges and changes which are redefining what it means to be human, what it means to work, what it means to be completely embedded in this world. ” ⁽¹⁾

References:

World Economic Forum. "The Fourth Industrial Revolution: what it means and how to respond".
World Economic Forum (Apr 13, 2016). The Fourth Industrial Revolution | Full Version. Retrieved from https://youtu.be/khjY5LWF3tg
The Cacophony Project https://cacophony.org.nz/
The Cacophony Project’s System retrieved from: https://cacophony.org.nz/technology
Vector of Steam Engine sourced from: Vector Graphics by www.vecteezy.com
Banner photograph by Tan Danh from Pexels https://www.pexels.com/photo/selective-focus-photography-of-right-human-hand-1358534/
Images from https://www.pexels.com/

Firmware Design Patterns in Embedded Systems

Wed, 15 Aug 2018 00:32:19 GMT

Author: Jonathan Kapene, Firmware Engineer

Introduction

This article discusses firmware development in embedded systems, specifically common design patterns implemented today. The article defines what a design pattern is, why it should be used in firmware development, and outlines one of the more popular patterns used in embedded systems today. However, this will not go in depth into any one specific pattern and rather touches on some popular ones. To be able to do this we have to first define what the terms "firmware" and "embedded systems" mean.

An embedded system is a "computerised system dedicated to performing a specific set of real world functions, rather than to providing a generalised computing environment" ^[1] . It consists of a combination of (electronics related) hardware and software (known as firmware). In fact, most of the products we at Beta Solutions design are considered embedded systems. These systems might be a standalone device or form part of a larger system. Embedded systems can be found in a broad range of places:

At home, embedded systems can be found in toasters, kettles, washing machines, dishwashers, food processors, televisions and more.
In the telecommunications field, examples of embedded systems include network routers and modems, low power wide area network enabled sensors, cell towers and of course the mobile phone.
The medical field has x-ray generators, ECG monitors, MRI scanners, even the hospital beds have embedded systems.
Embedded systems can be found in the science technology field with electronic microscopes, oscilloscopes.
Your car is also made up of dozens of embedded systems, ranging from electronic speedometer, ABS, power steering, car alarm, air conditioning and just the general heads up display.

Figure 1: Examples of Embedded Systems

So you can see embedded systems make up a large part of everyday life. It is likely that the average person interacts with over 30 embedded systems every day.

Embedded Firmware

Simply put, firmware is a type of software specifically written for embedded devices.

The vast majority of electronic devices today include some form of programmable componentry - usually in the form of a micro-controller. The firmware is the brains of system - telling the hardware what to do and when to do it. eg: "Read" the value from that sensor at exactly 08:15:00 and "if" the value of the sensor is 0, "then" turn on the error light. It is the firmware that communicates directly with the physical world via the electronics hardware.

When it comes to designing an embedded system there are inevitably a considerable number of design considerations and constraints. For example, hardware considerations include: size, computing power, thermal distribution and EMC - to name just a few. What about considerations for firmware?

Well, unlike computer/smart phone software, which are regularly popping up with “software update” messages, it is much more difficult (or impossible) to update the firmware in embedded systems. Ultimately this means that the developer has to get the firmware code ‘right’ before the device enters production. In short - design considerations for firmware are immensely important.

Performance constraints, for instance, are heavily considered and different systems will have different performance requirements. For example, the ABS braking system in your car - for obvious safety reasons - is critically required to be a highly responsive system. Other systems may value accuracy over responsiveness. Accuracy is an example of a quality of service constraint. Other examples of quality of service properties include:

Throughput
Reliability
Cost
Availability
Maintainability
Flexibility
Responsiveness

Design Patterns

A design pattern is a term used in the software development community that means "a generalised solution to a commonly occurring problem". A design pattern's purpose is to be a template for a firmware programmer when developing a system. Any one design pattern can be used in any programming language and is not restricted by the field in which the system is being designed for. (Note: If the pattern was specific to a certain field this would be considered an analysis pattern. An analysis pattern is problem specific, where as a design pattern is driven by the quality in which the system performs its requirements.) One system may have multiple design patterns implemented and often do.

A developer’s goal is to use a design pattern to optimise one, or a few, of the important systems quality of service properties at the expense of others. A good developer will pick a set of design patterns that optimises the important properties while keeping the properties of others to an acceptable level.

Popular design patterns used in embedded systems are listed below:

Observer pattern: Also known as the publish-subscribe method. It is a method which allows data to be shared to multiple elements and makes it easy to add more elements to share the data. Thus the system becomes more flexible.
Hardware proxy pattern: Elements specifically responsible for accessing certain hardware. These make it easy to update firmware when changes in the hardware design are made. Again, this increases the flexibility of the design.
Interrupt pattern: Used to pause what its currently processing and handle events as soon as they happen. This in turn increases the Responsiveness.
Queuing pattern: Uses queued messages to share information among multiple processes.
Rendezvous pattern: Synchronises 2 or more separate processes.
State pattern: Implements a state machine inside firmware.
CRC pattern: Verifies data which makes the system more reliable.

Observer Pattern

The observer pattern is a popular pattern in event handling systems. That is a system which monitors multiple inputs and may display these for the user in some way and/or respond by initiating some actuators. It's called the observer pattern because a subject (data object from input sensors) has a number of dependents which observe the subject for changes and react accordingly. In reality the subject will notify the observers when it has changed.

The observer pattern optimises the execution efficiency and flexibility of a system. It is flexible as it makes it easy to add more subjects and more observers to those subjects. The pattern is execution efficient, as all relevant observers are only notified when needed so not to waste processing time.

A nice example which demonstrates an observe pattern is in a car alarm. Car alarms can consist of a range of sensors, but let's look at a typical example. A car alarm that has three sensors to detect thieves, a control panel so the owner can arm and disarm the alarm, and three actuators that act as alerts. The actuators are observers to the sensor data. Using an observer pattern here makes it easy to add more observers without having to change the subject’s code.

Figure 2: Possible firmware architecture of a car alarm.

Summary

In summary, a design pattern is used by a software developer as a template to build part of an overall system. Most embedded systems will use more than one of these design patterns in practice and these should be chosen to fit the quality of service requirements of the overall system.

Figure 3: Beta Solutions Engineer writing signal processing firmware.

At Beta Solutions, our experienced embedded electronics engineers have the skill and proven track record to use the best suited firmware design pattern for our clients' products.

You can get in touch with us to discuss any idea you have in mind via our contact page or give us a call .

References:

Retrieved from https://pdfs.semanticscholar.org/a312/430dc5a54a1cae78e19f06c3ffa8509015a7.pdf
Photographic images retrieved from https://www.pexels.com
Photo by Mikes Photos from Pexels - https://www.pexels.com/photo/audi-vehicle-interior-1009871/
Photo by Torsten Dettlaff from Pexels - https://www.pexels.com/photo/space-gray-iphone-6-193004/

T+10 years and counting

Tue, 03 Jul 2018 23:31:59 GMT

Author: Terry Southern, CEO

It doesn't seem so long ago that Beta Solutions was founded, and now as I write this, the company has just had it's ten year birthday. That's ten whole laps around the Sun - something certainly worthy of celebration and reflection.

Excitement ... Doubt ... Satisfaction ... Intrigue... Passion ... Stress ... Enjoyment. These are all emotions that founders typically experience at some point as their "baby" transitions from a Startup to an Established business. Without a doubt I certainly remember times where I have experienced each and every one of these at some point in the past.

In this blog I reflect on these emotions and share a few musings about the journey to date.

Key Milestones:

2008
Business founded by Terry Southern and Matthew van der Werff.
1st Premise: Self-contained unit with adjoining garage.
First clients.
2009
2nd Premise: Larger office, within a larger Tech Hub at Palmerston North's Bio Commerce Centre (Now called "The BCC").
First private sector clients.
2010
New website launched.
Building our client base.
2011
First full time employee.
Optimising business structure (Terry CEO; Matthew CTO).
2012
Expanded services into Electro-Mechancial and PCB Production.
Number of clients reaches 25.
2013
Number of employees reaches 5.
2014
Clients based in all the main centres - Auckland, Wellington, Christchurch and beyond.
Industries worked now include: Transport; Industrial; Music; Aviation; Consumer; Scientific; AgriTech; Assistive Devices; IoT; Research and more!
2015
Number of employees reaches 8.
2016
3rd Premise! New custom-fit out 300sqm office in Palmerston North CBD.
2017
New website launched.
2018
Number of projects completed exceeds 100.

Left: Premise #1: Self-contained unit with adjoining garage.
Middle: Premise #2: Larger office, within a larger Tech Hub at Palmerston North's Bio Commerce Centre.
Right: Premise #3: Custom-fit out office in the CBD.

Inception

In 2008, Obama was elected as the first African-American US president. Uber and Spotify didn't exist and neither did the iPad or Instagram. Nor did Airbnb, Kickstarter, Google Chrome or LPWAN. In 2008, the world experienced the worst Global Financial Crises since the 1930s.

So it was in this year, when Dr Matthew van der Werff and I, first mooted the idea of establishing an Electronics R&D consultancy business. We had both been employed in the electronics industry prior to this and the opportunity of starting our own business intrigued us immensely. Due diligence was undertaken; the business case assessed; and the initial business plan devised.

Of course, there is always an element of risk involved in starting a business and no amount of planning will completely eliminate all risk. It is therefore natural for budding entrepreneurs to have some doubts about making the leap from (perhaps stable) employment into running one's own business. In fact, statistically, we know that many businesses don't even reach the 10 year mark - with one study putting these odds at only (approximately) 1 out of 3 surviving the decade ^[1] . Nevertheless, for us it was a case of "nothing ventured - nothing gained" and the opportunity was too good to pass up. Our quiet confidence and passion for electronic product development won out over any doubts we had and Beta Solutions was officially incorporated on June 24th 2008.

Left: The Beta Solutions team in 2008.
Right:The Beta Solutions team in 2018.

Humble Beginnings

Yes, it was humble beginnings. Our first office was somewhat stereotypical of a Kiwi Tech Startup - situated in a self-contained unit with adjoining garage serving as our workshop. At this time, our equipment consisted of the bare essentials - Power Supply; Oscilloscope, Signal Generator; Multi-meter, Soldering iron, PCB CAD software etc. Yet, despite some space and equipment restrictions, we worked hard, attracted clients (both long and short term) and delivered projects.

Through these early years, we also strategically re-invested significantly back into the business, enabling us to fund further marketing and the acquisition of additional equipment.

Left: In 2008, our technical gear consisted of the bare essentials.
Right: Over time we have expanded our arsenal to an array of quality equipment.

Onward and Upward

Of course, at Beta Solutions " We Design and Build Electronic Products ". More specifically, we are a consultancy business specialising in hi-tech electronics and electo-mechanical research and development. As the number of these R&D consultancy projects increased, during the period of 2010- 2012, it became clear that further professional resource was required. We hired our first employees in the areas of administration and engineering.

It was also during these years that we optimised our business structure, with myself leading the Business wing of the business (as CEO) and Matthew leading the Technical wing (as CTO). This allowed us to work with more efficiency and effectiveness and undoubtedly unlocked the potential for further growth within businesses.

As time went on, we expanded the services to included not only Electronics Hardware and Firmware Design, but also Printed Circuit Board Assembly (PCBA) Production. Our client base and employees both also increased steadily as did our awareness within the market.

A Day in the Life of Beta Solutions

I would be remiss if I didn't comment about what we have spent most of our time actually doing over the past decade.

Working in the business

Of course as per any business there is always daily "paper work" to attend to but typically most of our working day is engaged in undertaking actual project R&D work. In no particular order, tasks might include:

Meeting with our clients;
Writing product specifications;
Devising product architectures;
Researching for solutions;
Designing circuit boards (PCBs);
Writing software/firmware;
Modelling mechanical components;
Considering Regulatory Compliance;
Conducting simulations;
Building prototypes;
Project Managing;
Facilitating Production;
... to name just a few things ...

As we both design and build products , it would be fair to say our work consists of both highly academic aspects, as well as hands-on practical engineering. Our team consists of some very smart professionals, and it's an honour to see them really enjoy rising to the challenge of their work. (Personally, I also appreciate the opportunity to learn new things from them daily).

We love working alongside our clients and over the years we have been privileged to partner with many entrepreneurs and innovative businesses to help bring their product ideas to life. Our clients have been local, national and international - across a wide range of industries. We have enjoyed working with both Public and Private businesses; Startups as well as Established businesses. Perhaps the biggest satisfaction of all is when, after all the effort that goes into a project, we finally get to deliver a market winning product. To see our clients fully commercialise the product (in large numbers) to countries all over the world, is really quite a rewarding feeling. Their success is our success.

Working on the business:

Yet, another significant part of our day involves something else - not working "in the business", but rather working "on the business". We have always invested significant resources into strategically developing the business - as innovation is at the heart of our business and who we are as people.

Put another way, we are not content with merely accepting the status quo. It is in our DNA to continually search for new and better ways of doing things. We will never stop learning, we will never stop getting better.

“ If you are not moving forward, you are moving backward. ”
- Mikhail Gorbachev

Our Passion

Growing up, I didn't know exactly what I wanted to do as a job - but I did spend "a lot" of time tinkering with electronic and mechanical components/systems. As often as I could I would pull apart old electronic widgets - radios, TVs, robots, appliances, toys etc. I learned to solder at a young age and loved the feeling of being able to fix things. (OK, I may have also broken more than a few things along the way too - sorry Mum & Dad!). I built countless mini machines out of Meccano and Lego Technics, all the while learning about motors, gears and other mechanisms.

I guess then looking back, it wasn't a surprise that I chose to study a Bachelor of Engineering - majoring in Mechatronics at University.

My story is not unique and our staff (including Matthew) can share very similar experiences. Technology (or more specifically, Electronics and Software and Mechanical engineering) is our genuine passion. We don't have to force ourselves to like what we do - it comes naturally.

In fact, the reason we get out of bed and go to work is because " We enjoy the challenge of designing electronic products that will empower our clients and make a positive difference in the world we live in ."

Business is easier when you enjoy what you are doing.

Left: 2008 - The Beta Solutions team doing what they love.
Middle & Right: 2018 - Our (somewhat larger) team still doing what they love.

Our Culture

Every business has their own unique culture. This culture may have developed intentionally or unintentionally. Perhaps it developed for no known reason, or perhaps it's obvious that the culture in the organisation exists because one or several people in the organisation define it by their own character traits. Regardless, it is certain - just like death and taxes - that a business will have a culture.

From 2014 onwards, we chose to be more deliberate about establishing the culture we desire. A culture that is ingrained into the very fabric of the company and independent of particular staff - management included. From an early time in our business we decided to help pro-actively shape our culture using the following guiding values:

Integrity - Say, do & think what is right. Take responsibility for your actions.
Excellence - Strive for and set excellent standards in everything you do.
Enjoyment - Create and maintain an enjoyable work place environment.
Innovation - Continually search for new & better ways. Don't be content with the status quo.
Relationships - People are people. All people matter.

Values x Behaviour = Culture . Our values and culture are inextricably linked, and everything we do as a business strives to reflect these.

He aha te mea nui o te ao. He tangata, he tangata, he tangata

What is the most important thing in the world? It is the people, it is the people, it is the people .
- Maori proverb

There is one value for me that stands out, which is aptly illustrated by an experience I had around 2015.

Our team had recently been working on a complex project for a client. The deadlines were crushingly and uncompromisingly tight. We are pleased to say the project was completed on time, but truth be told, there were periods of palpable stress (for both parties) as the deadline neared. It was what happened next that really stuck with me.

One day soon after project completion, our office doors surprisingly burst open and we welcomed in a fully catered morning tea for our entire staff - courtesy of that client. I promptly expressed my thanks - to which our client replied: " Life and business are about people, and no matter what happens I try to remind myself of that ".

I couldn't agree more, and this is why we honour the value of "Relationship".

Business/Family Startups

On a personal note, the last 10 years has consisted of enormous personal change and growth for myself. For my wife and I, 2008 was also "life before kids" - which does feel like a very long time ago indeed. Becoming a Dad has been a most awesome journey in itself and now it's impossible to imagine life without kids. Matthew, would echo the same story.

Interestingly, our experience is that starting up a business has many similarities to raising a child. At the start of their life both babies and startups are naturally highly dependent on their parents and co-founders respectively. As time goes on children grow and learn - becoming increasingly more mature and more self-sufficient. The same can be said for business startups.

Both require a lot of time and attention. Both can be tremendously challenging and exhausting. Similarly both can cause stress & anxiety. Yet despite all of that, both can elicit remarkable feelings of excitement, reward, purpose and deep satisfaction. It's those latter emotions that make the roller coaster ride all the worthwhile.

We value our relationship with our families. It's especially rewarding to have " kids at work ", where we can share our passion for electronics with the next generation.

Thank you!

I would like to finish this blog by taking the time to acknowledge and thank all of our Stakeholders - past and present. Business co-founder Matthew; our staff, our suppliers & collaborators and of course our clients. Without you all we wouldn't exist.

If you haven't already (or even if you have before!) we sincerely welcome you to come and meet us and see our (kind-of) new office. Perhaps you'd like to find out more about what we do, and why we get out of bed in the morning and go to work! Or perhaps you'd love to discuss a new Product idea over a coffee? In any case, don't hesitate to get in touch .

Cheers to past 10 years and also to the future.

Terry

References

https://www.cbinsights.com/research/startup-failure-post-mortem/

Electromagnetic Compatibility (EMC) Compliance - Answers to Frequently Asked Questions

Wed, 30 May 2018 23:14:22 GMT

Author: Jason Cleland, Electro-mechanical Engineer

So you've got an idea for a brilliant new product that you want to develop and then put to market? And you've heard about this thing called 'EMC' but you don't know exactly what it is or why it should matter to you? Or how long will it take? Or how much will it cost? The purpose of this blog is to answer some of these questions and explain what 'EMC' is and why you really do need to know about it.

If you haven't already, please read our previous blog on compliance, Electronic Product Compliance - Answers to FAQs , which answers common questions our customers have had about complying with the regulations which govern electronic products here in NZ and abroad.

What is 'EMC'?

EMC or Electromagnetic Compatibility is a term which refers to how well electronic devices operate within the same vicinity without interfering with each other. A device which has poor Electromagnetic Compatibility could either cause problems with nearby devices or those same devices could cause it to have problems. You may have experienced this yourself if you ever used an older wireless home phone at the same time that a microwave oven is operating. The electromagnetic energy being emitted from a 'leaky' microwave oven can be detrimental to the audio quality on those old handsets. Wifi performance can also be slowed down or interrupted by 'leaky' microwave ovens or other wireless products like baby monitors or security cameras.

Almost all electronics emit at least some electromagnetic energy during their operation. This is just a consequence of using electrical energy. Given a high enough magnitude, almost all electronics can be susceptible to the energy being radiated from the devices around them. Due to this inherent relationship between devices, it is imperative that designers consider EMC during the development process else risk poor compatibility with the environments that their product operates in.

The result of unintended interactions between devices can range from imperceptible to disastrous. An interrupted phone discussion is a minor annoyance but a passenger aircraft navigation system is critical and the potential result of poor EMC design could be injury or death. In today's modern world there are electronics everywhere which all emit invisible electromagnetic energy. It's only through good design and testing that this has been achieved with few unintended interactions.

Radio Spectrum Management (RSM) is the governing body which regulates EMC products in New Zealand.

Why should you concern yourself with EMC?

In short, because it is a legal requirement ^[1] .

All markets have laws and regulations which govern how electronic products interact with the environments around them. There are often tests which are mandated by these regulations which have to be done before the device can be sold in that market. There are 'Standards' documents which define all the tests a device must pass in order to prove compliance with the regulations ^[2] . Devices which fail any of these tests would then be disallowed from being sold in that market until it is modified or fixed so that it passes the tests.

You could have both an amazing product and customers lined up but if you can't prove it's compliant with the EMC regulations of that market, you won't be able to sell a single unit ^[3] . Concerning yourself with EMC during the development process reduces risk and the liability of a product with poor EMC.

What is EMC testing?

To prove that your device is compliant you must have documentation which shows that the device has passed all the tests specified for that product category ^[4] . This documentation usually comes in the form of test reports produced by an independent test laboratory. These test laboratories are accredited by a governing body so that the test reports they produce can be recognised as proof of compliance.

There are two main categories of testing which can be done to measure the electromagnetic compatibility of a device. The most commonly required category is 'Emissions' testing which measures how much energy the device radiates or conducts out and compares this to set limits. The other main category of EMC testing is 'Immunity' testing where the device is put in prescribed electromagnetic environmental conditions to test it's susceptibility to electromagnetic phenomena it may experience in the real world.

Emissions Testing

Emissions testing typically measures how much energy the device emits with the following modes:

Radiated Radio-frequency (RF) Fields - Radiated RF fields are electromagnetic disturbances caused by the electrical activity in a device, and radiated out through the air.
Conducted RF Noise - Conducted RF emissions are also electromagnetic disturbances caused by the electrical activity, but are conducted out along its cables.
Magnetic or Electric Fields - Electrical activity in a device can produce magnetic or electric fields which can have near-field effects. Fluctuating voltages or currents in one device can induce or couple unwanted voltages or currents in other nearby devices.

The image below ^[5] shows various RF antennas used to measure the radiated RF fields of an electronic device:

Immunity Testing

These tests cover a range of situations where the device is put under 'worst-case scenario' conditions to see if its performance is affected. Testing your device for susceptibility can go a long way to ensuring that it will operate as designed in the environmental conditions it is likely to experience. Immunity testing typically seeks to measure the susceptibility of a device under the following conditions:

Radiated RF Fields - Same as described in emissions testing but now the device is made to operate under a barrage of RF noise from an external source. The conductors in that device can act like an ‘accidental receiving antenna’ and absorb part of that RF energy.
Conducted RF Noise - Same as described in emissions testing but now the noise is in injected into the devices cabling.
Transients - Surges, lightning strikes, and switching noise from electric motors can induce high voltage transients onto the cables which supply your device. These transient pulses could cause permanent damage to your device if not designed for properly. These transients are artificially induced into the cabling to test its effect.
Electrostatic Discharges (ESD) - Lightning strikes or even humans can cause electrostatic discharges to harm your device. Surfaces which can be touched by a user are subjected to prescribed ESD events (read 'sparks') and then tested to ensure no loss of functionality can occured.

The image below shows ESD validation testing we did at Beta Solutions, prior to sending a device to an accredited test laboratory for final testing:

What can be done to ensure that my product passes EMC testing?

Failing EMC testing can be very expensive. Not only because the cost of redoing testing, but there is the additional R&D cost associated with the redesign time and the opportunity cost of a delayed product launch. So naturally investing time in designing for EMC early on in the product development process can pay dividends if your device then passes EMC testing first time.

Understanding and predicting how electronic components interact so that they can be designed to pass EMC testing is a tremendously complex task. As such, it is impossible to ensure with 100% certainty that a product design will not fail EMC testing each time it is sent to the lab. Here are some principles which we apply to our projects to reduce the probability of EMC test failures:

Use good EMC practices - Our engineers have years of experience and training which enables issues to be remedied before they become problems. We achieve this by:
Having a sound understanding of EMC principles
Over-design (safety factors)
Using known 'good' designs
Applying industry accepted 'rules of thumb'
Test early, test often - There are more solutions avaliable if issues are found earlier in the development process. Hence the need to test early and often so that these issues can be found and then remedied when the cost implications of design changes are the least. At Beta Solutions we have specialised tools in our lab which allow us to do limited pre-compliance testing so we can have more confidence that a device is compliant before it is sent away for testing.
Use Computer Aided Design (CAD) Tools - Using computer tools to model electronic circuits for running simulations can be powerful for validating designs before they've even been prototyped. Changing a computer model of a circuit takes significantly less time than changing a prototype.

The image below shows a Beta Solutions Engineer scanning a device for sources of RF emissions with our spectrum analyzer and a near field probe:

Conclusions

The legal requirement for EMC compliance is here to stay. Every electronic product sold in New Zealand must comply. Embrace it now to minimise liability and ensure that your new product is a success.

Designing for EMC can be a tremendously complex area, often requiring one to be familiar with hundreds (or thousands) of pages of EMC standards. Therefore, it is often prudent to use the services of a professional electronics design consulting firm to take the burden off you.

To reduce the probability of EMC non-compliance, we at Beta Solutions, consider the implications for EMC compliance with every decision. We aim to design our clients' electronic products so that they pass the EMC testing process first time. Beta Solutions has years of experience with complying with standards which enables us to create products which do not cause interference and are immune to their environments.

If you would like to know more about EMC compliance in electronics product development, feel free to get in touch via our contact page or give us a call .

References:

Radiocommunications (EMC Standards) Notice 2015, Part 3 clause A, Retrieved from: https://gazette.govt.nz/notice/id/2015-go4671
Ministry of Business, Innovation and Emplyment (NZ), 2017. Compliance standards for EMC & Radio. Retrieved from https://www.rsm.govt.nz/compliance/supplier-requirements/compliance-standards-for-emc-radio
Radiocommunications Regulations 2001, Clause 37.1.a.ii, Retrieved from: http://www.legislation.govt.nz/regulation/public/2001/0240/latest/DLM71584.html
Ministry of Business, Innovation and Emplyment (NZ), 2014. Product documentation. Retrieved from https://www.rsm.govt.nz/compliance/supplier-requirements/product-documentation
Schneider Electric. Retrieved from https://blog.schneider-electric.com/telecommunications/2014/01/14/emc-standards-review-requirements/

3D Printing - The Move to Mass Production?

Thu, 03 May 2018 23:06:44 GMT

Author: Phillip Abplanalp, Hardware Engineer

Introduction:

The core technology of “3D Printing” or “Additive Manufacturing” has been around for almost four decades (the early 1980s), but it wasn’t until the advent of low cost Fused Deposition Modelling (FDM) in the late 2000s which really brought about the democratization of 3D Printing. This increased accessibility of 3D Printing generated a huge amount of public and journalistic interest which resulted in 3D Printing becoming one of the most hyped emerging technologies. This brought with it wildly optimistic forecasts of overnight wholesale changes to the manufacturing sector. Now, five years on from the peak 3D Printing hype, it is clear that 3D Printing has fallen short of the rosier predictions.

This blog will focus on how 3D Printing may now finally be shifting from the prototyping/ development space to becoming a viable manufacturing method at scale.

Before we investigate this, for those that are not familiar with 3D Printing, here are a few of the core printing technologies commonly used today.

Fused Deposition Modelling (FDM):

In essence, Fuse Deposition Modelling is a hot glue gun (known as the extruder) mounted to a 3-axis platform. The extruder moves around the build plate (base of the printer) ejecting molten plastic wherever material is required, this step is repeated layer by layer – see image below.

Common FDM materials:

Polylactic Acid (PLA)
Acrylonitrile Butadiene Styrene (ABS)
High-Impact Polystyrene (HIPS)

Figure 1 – Image showing the layer by layer application of plastics [1]

Stereolithography (SLA):

Stereolithography, simply put, makes use of a liquid material (resin) which is selectively hardened by a movable laser beam. The liquid resin becomes solidified at the points where light from the laser beam shines on the surface, this laser will generally be in the Ultraviolet end of the light spectrum. As with FDM (or virtually any 3D Printing method), this is done with a layer by layer approach.

Common SLA Materials:

Materials for SLA machines will usually be proprietary to the printer manufacturer due to the dependence on the wavelength of the UV laser used for curing.

Figure 2 - Image shows the workings of SLA printing. UV laser solidifying layer by layer [2]

Selective Laser Sintering (SLS):

Selective Laser Sintering, as the name suggests, uses lasers to selectively fuse (sinter) material together. Unlike SLA, the raw material is a fine powder and the fusing is achieved by thermally bonding powders together using a high-power laser. For the layer by layer process, a roller or slider applies a thin layer of powder after each z-axis increment. SLS is one of the most commonly used methods for engineering applications due to both the range of materials and its mechanical properties.

Common SLS Materials:

Steel
Titanium
Nickel Alloy
Stainless
Aluminium
Inconel
Nylons

Figure 3: SLS printing system with laser-based sintering [3]

Low Volume Production

3D Printing as a tool for assisting with prototyping, conceptual evaluation, and integration testing has been commonplace in several large industries such as airspace and automotive for at least the last 20 years. More recently though, high value and low volume industries have been shifting towards making direct use of 3D printed parts in their final product. This is particularly noticeable in industries such as the Commercial Space arena and Formula One. Companies such as Space X and the New Zealand based Rocket Lab are believed to make extensive use of 3D printed components, this even extends to components which require a high degree of mechanical integrity.

An example of this is the Rutherford engine which Rocket Lab claims to use 3D printed parts for “all primary components” [4].

Figure 4: Rocket Labs’ 3D printed Rutherford engine [4]

There are numerous advantages to using 3D Printing as a final production part, these are outlined below:

Greater design freedom, allowing for extremely intricate parts to be designed e.g. internal cavities and support structures which would be impossible when using conventional manufacturing methods.
Extremely fast design iterations.
Low tooling costs.
Reduced amounts of waste material (additive vs subtractive manufacturing).

Like the Commercial Space industry, Motorsport (in particular Formula One) has moved from not only printing components for their internal wind tunnel models and prototypes but rather printing components get used directly on the race track. It was estimated in 2015 that Formula One teams were using between 40 to 60 3D printed components per race car [5]. This number of 3D printed components used on Formula One cars is likely to have risen since then, with rumours circulating that Ferrari may, in fact, be using 3D printed engine pistons.

Figure 5: Formula One exhaust with 3D printed components [5]

Scaling to Medium Volume Production

As previously mentioned, 3D Printing is a fantastic technology for low volume and high-value production. However, when scaling to higher volumes and smaller profit margin products, 3D Printing as a production solution has in the past been too expensive. The main driver of high production cost is related to the printing speeds [10], where a component that may take (say) a few seconds to make in an injection mould, could take anywhere from tens of minutes to several hours using conventional SLS, SLA or FDM methods. Hence medium to high (1,000 - 100,000) volume 3D Printing production in most cases is economically unviable-even when taking into consideration the significant overhead cost of the tooling needed for injection moulding.

The conventional break-even point for injection moulding (vs 3D Printing) can be anywhere between 50-500 units (Depending on many factors, including component size, shape, and complexity). The graph below shows an example of a bracket printed in nylon using conventional SLS vs Injection moulding in polyamide nylon [6]:

Figure 5: Injection Moulding vs 3D Printing Breakeven [6]

The cost comparison given in Figure 5 details the large overhead cost (~$800) inherent in the injection mould tooling (shown in blue) with an exponential price drop as the tool cost become less significant. In this particular case, the break-even point is ~130 parts at a price of $33 per unit. This would make the case that conventional SLS 3D printers are not sufficiently fast/cost effective as a medium-to-high volume production tool.

Fortunately, technology is forever evolving and 3D Printing (still being in its infancy) is showing plenty of development potential. Siemens in 2014 predicted that 3D Printing machines would become 400% faster in the coming five years [10]. However, according to claims made by several large 3D Printer companies, developing new 3D Printing methods such as Carbon 3D, Desktop Metal, and HP the advances may, in fact, have been significantly higher. The claimed speed improvements made by these companies vary between 10 to 100 times (1000-10,000%)* faster than conventional methods and since printing speed correlates to the final component cost, this could bring significant cost savings.

*Note: These are marketing claims and presently there is a lack of data to give a good indication on its validity.

The following section will take a closer look at two of these 3D Printer Companies.

HP Jet Fusion:

HP in recent years has taken its printing business to a new dimension with the addition of a Z-axis! HP made a move into the 3D Printing space around one year ago with the launch of its Jet Fusion printing system. The new Jet Fusion system is an interesting new plastic 3D Printing method which even includes multi-colour printing. Its approach is a blend of 3D SLS printing and conventional 2D inkjet printing. The plastic powder is evenly distributed over the build platform (as with SLS) and is then followed by an ink jet like nozzle, distributing fusing and detailing agent which ensure only the desired sections become “fused” together. Finally, the powder and fusing agent is thermally fused together - see Figure 6 below:

With this new approach, HP has made significant improvements in both the print speed and component cost. In fact, HP is claiming its print speeds are around 10 times faster and are about half the cost of conventional FDM and SLS printers. Below is a graph (from HP) broadly showing how their printing system shifts the break-even point (to conventional manufacturing i.e. injection moulding) to ever higher part quantities.

Desktop Metal:

Desktop Metal’s new metal 3D Printing production system (multi-staged approach forming and sintering) - due to be released in 2019 - makes some extremely impressive claims. Printing speeds are claimed to be “up to 100 times faster” than that of conventional laser-based systems (SLS). Additionally, part cost being reduced by around 20 x compared to today’s metal 3D Printing systems [9]. According to an example given by Desktop Metal, their production system could bring the break-even point (when compared with metal casting methods) of up to 100,000 units - see the graph below:

Figure 8 – Comparison of Conventional SLS (grey), Casting (white) and Desktop Metals Production system (red) [9]

This particular comparison was based on a water pump impeller which has some fairly intricate details (see image below) and therefore is likely to be more favourable to 3D Printing than casting. Component complexities aside, the component price for this water pump impeller is stated to be ~$4/unit which makes this an extremely promising system [9]. The method in which Desktop Metal achieve this is similar to HP’s approach and I would recommend taking a look!

Figure 9: Desktop Metals comparison impeller [9]

Conclusion

With the advancements made by the likes of HP, Carbon 3D, and Desktop Metal, the 3D Printing industry is now finally approaching a point where 3D Printing may, in fact, become a genuine economically viable alternative to conventional manufacturing methods. This certainly looks to be the case at medium-volume production and even high-volume production in some cases, such as the “water pump impeller” example given by Desktop Metal. However, high-volume production will likely remain limited to intricate/complex components where conventional production methods have astronomical high tooling costs. The move to use 3D Printing as a production tool is likely to bring significant benefits to the manufacturing sector, including:

Reduced design cycle time
Reduced manufacturing tooling cost
Reduced manufacturing setup
Potential for personalized customization – with virtually zero cost
Simplified production automation

Finally, the proliferation of 3D Printing for manufacturing is good news for small-medium sized businesses with ambitious plans but the limited financial capacity to invest in expensive production tooling.

Here at Beta Solutions, we continue to use 3D printing as an effective method for rapid prototyping and are excited about the developments being made in this space. We continue to follow it closely and are always looking for ways in which the latest technological developments can be best applied to our clients' needs.

You can get in touch with us to discuss any idea you have in mind via our contact page or give us a call .

References

[1] Retrieved from - https://en.wikipedia.org/wiki/Fused_filament_fabrication
[2] Retrieved from - https://en.wikipedia.org/wiki/Stereolithography
[3] Retrieved from - https://en.wikipedia.org/wiki/Selective_laser_sintering
[4] Retrieved from - https://www.rocketlabusa.com/news/updates/rocket-lab-reveals-first-battery-powered-rocket-for-commercial-launches-to-space/
[5] Williams, N. (2018) Sauber Motorsport AG and Additive Industries: Formula 1® engineering meets metal AM. (Pg. 91). Retrieved from http://issuu.com/inovar-communications/docs/magazine_metal_am_spring_2018_pdf_s?e=32443561/59440823
[6] Retrieved from - https://www.xometry.com/blog/3d-printing-vs-injection-molding-breakeven/
[7] Retrieved from - http://www8.hp.com/us/en/printers/3d-printers.html - Technical White Paper
[8] Retrieved from - https://www.google.co.nz/imgres?imgurl=http%3A%2F%2Fwww.www8-hp.com%2Fus%2Fen%2Fimages%2Fvideo-poster-reinvent-3_tcm_245_2243684.jpg&imgrefurl=http%3A%2F%2Fwww8.hp.com%2Fus%2Fen%2Fprinters%2F3d-printers.html&docid=0o1lRPaz3GKcGM&tbnid=8rlDguz8v-XjIM%3A&vet=10ahUKEwjB24ycuOnaAhWCNpQKHcS1B_oQMwihAShTMFM..i&w=900&h=493&bih=949&biw=1920&q=hp%203d%20printing&ved=0ahUKEwjB24ycuOnaAhWCNpQKHcS1B_oQMwihAShTMFM&iact=mrc&uact=8
[9] Retrieved from - https://www.desktopmetal.com/products/production/
[10] Retrieved from - https://www.siemens.com/innovation/en/home/pictures-of-the-future/industry-and-automation/Additive-manufacturing-facts-and-forecasts.html
[11] Retrieved from - http://www.metal-am.com/wp-content/uploads/sites/4/2018/03/MAGAZINE-Metal-AM-Spring-2018-PDF-dp-1.pdf

Sigfox Technology Review

Wed, 28 Mar 2018 21:52:06 GMT

Author: Jonathan Kapene, Firmware Engineer

In this blog we take a closer look at the Sigfox technology that was first introduced in our December Blog: "LPWAN Benefits for IoT Connectivity". To recap, LPWAN - which stands for Low Power Wide Area Network - is a communication protocol which is designed to transmit small amounts of data over large distances.

Compared with traditional cellular-network connected devices, LPWAN devices consume significantly less electrical power, enabling these devices to last 5-10 years on a single battery. Large Telcos (such as Spark and Vodafone) are now recognising the low-power benefits of this technology and are in the process of trialing certain LPWAN infrastructure. It is expected that Telco-supported LPWAN technologies such as LoRa and NB-IoT will be officially rolled out in New Zealand sometime between 2018-2019.

Sigfox's network however achieved 50% coverage of New Zealand in Aug 2016! The rest of this article will detail what Sigfox has to offer to the LPWAN market and will provide a brief technical overview of how the underlying technology works.

General Overview

Founded in 2009, Sigfox is a French company that has designed and developed a Low Power Wide Area Network (LPWAN) technology by the same name. So to be clear, we have Sigfox "the company" and Sigfox "the technology".

Sigfox has been a pioneer in developing a network specifically designed for Internet-of-Things (IOT) applications. They foresaw the ever-growing demand for internet connected devices and no network suitable to do it - so they built their own end-to-end proprietary solution. Their vision was and is to "connect the unconnected" and they continue to work towards their goal of connecting billions of devices through their network.

Sigfox technology is suitable for applications where devices:

Need to be low cost
Send small amounts of data infrequently
Use one-way communication (device to base station direction only)
Transmits non-critical information only.

A good example would be a remote temperature sensor. Such a device could be manufactured in large quantities at low cost. Further, it would only ever transmit a small amount of data and infrequent measurements (say hourly) would be appropriate. Also, one-way communication would be suitable as data would only ever be sent from the temperature devices to the receiving base station.

On the other hand, if your application requires high amounts of data* or two-way communication** or message sent to the device - Sigfox is likely not the LPWAN technology for your device.

* Sigfox only permits 12 byte payloads which is enough for some applications but far too little for others.
** Technically, Sigfox does have the capabilities for two-way communication however it is very limited.

Technical Overview

The following provides more technical information about the Sigfox technology in the form of some "Frequently Asked Questions".

#1 Broadly speaking, what does the Sigfox Network Architecture look like?

Sigfox describe their network as having a horizontal and thin architecture comprised of two main layers as follows:

1. Equipment layer

You can think of the equipment layer as the hardware used to create the network. The majority of this layer consists of the base stations - which are responsible for listening and receiving messages from devices and forwarding them to the 2nd layer, support systems. The Sigfox base stations in New Zealand are built by Kordia and operated by Thinxtra (who is also the network operators in Hong Kong and Australia).

Figure 1: Sigfox Development Kit from Thixtra

2. Support system layer

The support system layer is the back-end software side of the network - where the devices' data gets processed. The device's data may then be displayed through a web page that can be accessed anywhere with internet. More technically speaking, you have the choice to setup what is called a "callback" which will forward your devices' messages to a place of your choosing. Once your data has been forwarded to the correct place on the internet (e.g.: a supported cloud platform), you can customise the way your data is displayed and who has access to it.

Figure 2: Sigfox Network Architecture

Security FAQs:

#2 How secure is my data?

The architecture of the network and the nature in which the devices communicates is inherently more secure than a typical cellular network. A Sigfox device is actually never "connected" to the network. It only ever sends and receives data when it was programmed to. The device will never be able to send data to other entities through the internet because it is never connected to it. Sigfox call this their "built-in firewall"

#3 What about someone intercepting the message intending for a base station

By default the message is encrypted but the payload information can be seen by the network provider. If you want to keep your data private, you can add your own encryption at the device end and then decrypt the information once it has reached its destination.

#4 What if someone imitates my device and sends incorrect/malicious data

During manufacturing, devices are given an authentication key. Any message from a device will have a cryptographic token in the header information which is generated from its authentication key. This token is checked in the support system layer to validate the device's identity.

Network coverage FAQs:

#5 What network coverage can I expect?

As of February 2018, Thinxtra have achieved Sigfox coverage of 85% of New Zealand's population (mostly city centers). This is one of Sigfox's biggest advantages as Spark and Vodafone don't have an offcial LPWAN network yet.

#6 What if my devices don't have general network coverage?

Unlike LoRa you can't just purchase your own gateway and set it up as a base station. Fortunately however, Thinxtra provide an option for users to rent a "mini base station" that essentially extends the network coverage to include your devices. It sounds a lot like having your own gateway but Thinxtra will install and maintain it for you. If you have lots of devices they will consider installing a full base station to include your devices free of charge.

Figure 3: Sigfox Coverage of New Zealand

#7 How does Sigfox achieve long range and low power?

Short answer: "It's technical". If you are interested ... read on ...

Longer answer:

They use a carrier frequency of <1GHz - which helps communications to penitrate further through the atmosphere. i.e increasing the range.
They use Ultra-narrow band modulation (UNB).

Even longer answer:

Sigfox transmits from device to base station using ultra-narrow band (UNB) modulation across a 192KHz band. The base station listens on this 192KHz wide band for any UNB messages. These UNB messages are 100Hz wide which helps with the overall capacity of the network as well as makes it easier for the base station to distinguish between the message and the noise floor. UNB is known for being spectrum efficient which makes sense in this case because each message only uses a tiny bit of the 192KHz bandwidth.

To improve the chances of the base station receiving the messages, Sigfox employs random access. Traditional cellular networks rely on the device being connected to one tower continuously and when the device move out of range it will connect to the next tower. With Sigfox devices this is not possible as the power requirements are too high. Instead the device and base station communicate unsynchronized. Random access helps to overcome the unreliability that comes with communicating unsynchronized. Every time a device sends a message it will actually transmit three copies of that message on differing frequency inside the 192KHz bandwidth.

Unlike traditional cellular, which communicates in licensed spectrum, Sigfox operates at 928MHz which is unlicensed in New Zealand. LoRaWAN technology uses 928MHz for their downlink messages which may cause interference when the density of devices increases. The downside of using the unlicensed spectrum is of course any one can transmit on it, so technical rules are put in place to mitigate possibilities of interference. These rules limit the "on-air" time to 1%. Although these rules are different depending on the country, Sigfox has decided to have their rules uniform no matter the country.

Rules for Sigfox:

A maximum of 140 uplink messages per day
12 byte payload per downlink message
A maximum of 4 downlink messages per day
8 byte payloads per downlink message
Transmits at 600bps 24dBm (in NZ)

An "uplink message" is a message sent by a device to a base station which is later displayed in the cloud back-end. The majority of messages will be uplink messages, carrying the usefull data your device is collecting. The payload is a modest 12 bytes which take six seconds to send. So you can see the effective data rate of payload infomation is 2 bytes per second. Compare that to 4G cellular which has a data rate in the range of 5Mbps and which is over 2 million times faster. But of course sending small amounts of data has the upside of only using a little bit of energy, allowing devices to last up to and past 10 years, where as my phone is lucky to make it through the day.

"Downlink messages" are configured in the back-end and sent from the base station to the device. A maximum of four downlink messages per day gives us a hint that Sigfox was not designed for downlink communication. It's simpliy there for configuration purposes but never the less it is still there. So if you have a deivce that needs two way communication Sigfox is probably not the technology for you. Downlink messages are requested by the device, so you can't just send one anytime you want either, the device will have to be programmed to request a downlink message at certain times. The range at which these downlink messages can be recieved will suffer from the low grade recievers on the device. unlike the cellular grade recievers at the base stations.

#8 What are the key differences between Sigfox and other LPWAN technologies?

The two main competitors to Sigfox are NB-IoT and LoRa. As previously mentioned, Vodafone and Spark intend to release their networks sometime during 2018. Vodafone plan to release just the NB-IoT network whilst Spark have proposed to release both.

LoRa offers a lot more network freedom to the user:
With LoRa anyone can set up their own gateway and operate your network free of charge. (Alternatively, you could use the likes of paid networks like Spark or KotahiNet).
With Sigfox there is strictly only one network available - provided through Thinxtra. (Even when you rent out a mini base station you have no control over it). This highlights the potential risk that if Thinxtra was to go under (however likely or unlikely) Sigfox devices in the Asia Pacific could become obsolete.
NB-IoT offers more data:
NB-IoT can acheive transmit speeds of around 20-60 kbps. This allows for "over the air" (OTA) device firmware upgrades.
Sigfox data speeds 0.6 kbps are just too slow for OTA updating.
Sigfox offers roaming:
Sigfox (and LoRa) both have the ability to roam between base stations.
NB-IoT devices struggle with roaming, and are better suited paried to a single base station.
Sigfox has the lowest cost devices - a significant benefit especially if users are considering large scale deployment.

Sigfox may be closed on the network side, however it is very open on the device side. The protocol stack is the firmware in modules that generates the radio frames in the correct way to be able to communicate to Sigfox's base stations. To get the cost of devices down, Sigfox give away their protocol stack royalty free to modem manufactures. This is why Sigfox have the most affordable modules out there at a quarter of the price of other technologies' modules.

See the table below for a more detailed comparison Overview of Low Power LoRa Communication...

Conclusion

There is no one size fits all LPWAN technology, each one has their pros and cons and Sigfox is no exception. Sigfox out shines the other technologies with its royalty free protocol stack allowing for super inexpensive devices. It has already covered >85% of the New Zealand population. The down sides of Sigfox are its very small payload sizes, limited messages per day and its two-way communication limitations.

Beta Solutions has experience in developing products containing various communications technologies, including: LoRaWAN, Sigfox, Cellular, Bluetooth, Wifi, and Satellite communication.

If you are after more information, subscribe to our Blog for updates as we will continue to look at emerging IoT connectivity technologies. Alternatively you can get in touch via our contact page or give us a call .

Refrences:

Figure 2: Sigfox Network Architecture. Retrieved from https://www.disk91.com/wp-content/uploads/2017/05/4967675830228422064.pdf
Figure 3: Sigfox Coverage of New Zealand. Retrieved from https://www.sigfox.com/en/coverage
Photographs from Pexels https://www.pexels.com

Budgeting for Innovation

Tue, 27 Feb 2018 21:33:50 GMT

Author: Brenda Wormgoor, Marketing and Operational Manager

" I believe in innovation and that the way you get innovation is you fund research and you learn the basic facts. " - Bill Gates

Can you afford NOT to invest in innovation?

Being new product developers in the high-tech electronics field, we have a front row view into how disruptive, emerging technologies open up new possibilities and change industries. Seemingly overnight, established industries and the companies in them become obsolete or left behind when they do not invest in innovation.

Perhaps this is why Apple Inc. continues to increase their spend on R&D (spending over US $10B in R&D for 2016) as they look for other products and industries to broaden their product portfolio and lessen their dependence on the iPhone for growth.

Figure 1: Apple Inc. Research and Development Expenses, by Quarter - Globally, in Billions.

Innovation is not always about spending money on the next big idea. Investment in new technologies can improve productivity and profitability by:

Improving efficiencies, saving time on labour or production,
Improving product costs,
Updating existing products to close the competitive gap or, even better, gaining the leading edge in your market.

Plan for Innovation Investment with an Innovation Budget

Here are some helpful tips we came across that could help your business plan for innovation:

Establish a Strategic Innovation Roadmap

As companies start to get to grips with the way rapid innovation is shifting the landscape of markets, more and more are moving towards an agile approach in various departments, from strategic goal setting, marketing methodologies, to even financial planning.

Corporate strategies can thus no longer be strictly linear, as it has to allow for agile possibilities that come with new technologies such as Augmented Reality, Artificial Intelligence, Internet of Things, Precision Farming, to name a few. These technologies can at any time change the customer requirements of current products or alter the scope of existing new product development projects.

A strategic innovation roadmap is thus a guide that aligns innovation efforts to a company’s long-term goals whilst allowing the flexibility to react to and budget for uncertainties and risks as they arise. This ensures the required nimbleness to capitalise on new opportunities.
Change your Budgeting Approach

One thing is clear - traditional budgeting processes are rigid in nature and can not be used to plan for innovation investments that require agility and speed in responding to new developments in technology.

The flexibility of a responsive budget process is required in order for a company to allow innovation to succeed. Such allowance can be in the form of building in contingencies for either under or overestimation in the uncertain realm of emerging technologies where there is no historical knowledge to draw from.

How you measure innovation investment is also important because it is an experimental approach to learning, charting new territory with unpredictable outcomes. Transformational innovation mostly follows a J-curve pattern and is not comparable to traditional linear growth patterns that investments are normally measured against.

Figure 2: J-Curve Growth Chart Associated with Transformational Innovation.

3. Set Aside an Innovation Budget

An innovation budget is crucial in giving companies the freedom to experiment with and benefit from new technologies with the speed and agility required to succeed ahead of the competition.

However, even more important, it facilitates a mindset of innovation and willingness to change that creates a corporate culture of forward thinking that can set you apart from your competition.

Be sure to know what the budgeted split of funds (and resources) are between operational expenses, continuous improvement activities, and disruptive innovation. Challenge this split of investment in its ability to ensure future value creation. Also ensure that innovation budgets are protected from being pilfered by revenue-producing parts of the business.

Conclusion:

“ Innovation has nothing to do with how many R&D dollars you have. When Apple came up with the Mac, IBM was spending at least 100 times more on R&D. It’s not about money. It’s about the people you have, how you’re led, and how much you get it. ” - Steve Jobs

As Steve Jobs said, innovation is not just about the money spent, it is about a broader investment in:

Adaptive management practices
like responsive budgeting processes that enable companies to capitalise on opportunities that come with emerging technologies;
Individuals
as their knowledge and skills need to keep up with new technologies;
A business culture
that enables company wide idea generation, tolerance for risk, and learning from failures.

All in all, “ transform or perish ” is a reality and it is clear that investing in innovation is no longer a nice to have, but a survival necessity.

We are happy to discuss our flexible and proven innovation process and project estimation techniques (designed to help minimise risk for our clients) with you in a free 1-hour consultation.

References:

Figure 1: Leswing, K. (2017). Apple Inc. Research and Development Expenses, by Quarter - Globally, in Billions. Retrieved from https://www.businessinsider.com.au/apple-rd-spend-charts-2017-2?r=US&IR=T
Graphic vectors retrieved from Vecteezy https://www.vecteezy.com/
Photographs from Pexels https://www.pexels.com

Five SpaceX Insights for Startups

Tue, 30 Jan 2018 21:21:47 GMT

Author: Terry Southern, CEO

Perhaps you identify with being an entrepreneur? Maybe you already own or manage a startup business, or have an interest in them? This blog takes a high-level look at five key traits which have enabled SpaceX to “concur the Startup”.

Space Exploration Technologies Corp, trading as the more well-known and hip-sounding name “ SpaceX ”, is a privately owned American aerospace company. Not one to miss an opportunity to grab a headline and wow the media (much in the same way as Apple Inc does), there is a high probability you have already heard of SpaceX - along with its billionaire founder and CEO Mr Elon Musk. In fact, the story of Musk and SpaceX are so inextricably linked that this blog sometimes uses the two names interchangeably.

Founded in 2002, SpaceX is a relatively new player to the space transport business, and yet despite what is a highly entrenched industry, SpaceX has recently achieved remarkable success. In the process, they have fended off some traditional aerospace juggernauts such as Lockheed Martin and Boeing (collectively known as the United Launch Alliance “ULA”), making their achievements all the more impressive.

SpaceX’s achievements to date include:

The first privately funded company to:
Place a liquid-propellant rocket into orbit (Falcon 1 in 2008);
Successfully launch, orbit, and recover a spacecraft (Dragon in 2010); and
Send a spacecraft to the International Space Station (ISS) (Dragon in 2012).
Additionally, they were also the first company outright to achieve a propulsive landing for an orbital rocket (Falcon 9 in 2015);
The first to reuse an orbital rocket (Falcon 9 in 2017);
They have also been awarded a contract by NASA to send astronauts to the ISS (in their proposed Dragon 2 capsule) and return them safely to earth.

These achievements undeniably indicate that (despite the long odds) SpaceX has now successfully made the transition from being a fledgling business startup into a full space-launch-provider commercial enterprise. Refer to the graph below which shows the impressive growth of SpaceX’s market share .

Figure 1: Global Commercial Market Share

Below follows the five key traits this blog identifies as having enabled SpaceX to become a successful Startup:

1: Timing Is Everything

As the saying goes, “timing is everything”. Simply put, one of the reasons SpaceX has been successful is because (all things considered - technical, political etc) the timing was right. To give some further context read on:

Since shortly after the end of WWII, the US Government started investing enormous amounts of money into their Aerospace program. In doing so, without a doubt the USA became the #1 leading Aerospace nation in the world. In 1958 the federal government formed a dedicated executive government branch responsible for their civilian space program and aerospace research known as the: National Aeronautics and Space Administration, or “NASA” for short.

The Government (eg: NASA or the US Airforce) would typically commission a piece of work and then engage 3rd party contractors (such as Boeing, Lockheed Martin, Northrop - Grumman etc) to actually design and build the rockets and spacecraft.

The rate of development and early success of NASA was phenomenal.

In the 1950s they developed rocket powered planes that could travel at Mach 6 and into the fringes of space; (The X15 is still the world’s fastest manned plane).
The 1960s saw the Mercury and Gemini Programs - which launched astronauts into orbit, and enabled complex space maneuvers like spacecraft rendezvous and spacewalks.
In July 1969, under their Apollo Program, they landed a man on the moon. The spacecraft was launched by the mighty Saturn 5 rocket booster - still currently the most powerful rocket that ever flew.
In the early 1970s they continued to send men to the moon, and then later established their first Space Station called SkyLab.
The 1980s saw the Space Shuttle, a wonderful but massively costly spacecraft program, which operated until they retired the Orbiters in 2011.

Since the Space Shuttles were retired in 2011, the US has not had their own means to take their own astronauts into orbit. In fact, they pay Russia ~$70M per person to use the rather outdated Soyuz spacecraft to ferry their astronauts to/from the International Space Station.

While NASA is developing a new spacecraft (Orion capsule) and a new colossally-sized Launch Vehicle (called the “Space Launch System (SLS)” - this is still some years away.

Ever since the 1960s, when the crowning glory of landing man on the moon was achieved, NASA’s operating budget and hence rate of development has dropped markedly. Like any government funded agency, they are at the whim of “politics”.

While the appetite for state-funded space exploration was declining, the technology simultaneously continued to advance rapidly. New composite materials & manufacturing techniques, increasing computing power and optimised liquid hydrogen propulsion systems were all becoming available to enable a new breed of spacecraft and launch vehicles. The technology was just waiting for someone to take advantage of it.

The stage was set.

In the early 2000s, Elon Musk saw the timing was right. He seized the opportunity to develop a private Space business - one which was free of direct political influence; one which was disruptive and highly innovative and could leverage the latest emerging technology and also one which was audacious enough to set inspiring long term goals (more on that later).

Had SpaceX been incorporated too early, perhaps the technology would not have been sufficient or perhaps NASA would not have been ready to outsource to novice 3rd party launch providers. Too late to the market and they would have faced the challenge of further competition. Now was the time to strike - and strike they did.

2: Innovation Not Imitation

OK, timing is really important, but perhaps it’s not everything.

When Musk originally had ambitions to send “something” (perhaps a plant or a mouse) to Mars, he didn’t have any rockets. So in 2001, he flew to Russia to try and procure from them some ICBMs (repurposed missiles). In short, Russia wanted too much money and the negotiations failed.

Undeterred, he realised at that moment that he would not try to imitate others - but rather Musk turned to his colleague Jim Cantrell and said simply “ I think we can build a rocket ourselves ”.

Traditionally, the cost to launch medium-size military payloads into orbit has been extremely high. NB: It’s a complex number to unpack but the US Government accountability office has calculated that the average cost of each ULA rocket launch for the US government has risen now to approximately US$422 million ¹ .

Part of the reason of the high cost is that rockets are used only once and then discarded. It was obvious to Musk from the outset that until the cost per launch be drastically reduced - the Space industry would be forever hamstrung.

What was needed, was to bring the cost of a launch down considerably. To do this, they adopted the following innovative approaches:

1. Reuse the rocket:

Instead of letting the first stage rocket booster consume all its fuel during launch and then letting it crash down into the ocean - rendering it useless after a single use, SpaceX has designed a reusable rocket. Essentially, they keep some fuel in the tank and use precision electronic guidance and control systems to vertically fly the rocket back down to earth, firing the rocket engine just the right time/amount to enable a soft landing.

Once the rocket booster has been refurbished, it is ready to fly again, saving considerable costs.

2. Launch often:

SpaceX charge their customers considerably lower fees than their competitors, resulting in less profit per launch. The only way then to make enough money as a business is to launch more often.
In 2017 SpaceX (using their Falcon 9 Rocket) performed more launches than any other launch provider.

Figure 2: Satellites launched in 2017.

SpaceX’s innovation in both their technical and business departments have allowed them to charge the US Government only $90M per launch (and $60M for commercial customers), which is a far cry from the ~$420M per launch that its competitors charge.

As per Elon Musk “ Fundamentally, if you don’t have a compelling product at a compelling price, you don’t have a great company. ”

Not only have these innovations greatly assisted them in becoming successful - they will likely forever change the way the Space Launch Industry operates.

3: Hard Working Team

Yes, you read this title correctly - three key words: “hard” and “work” and “team”.

Musk believes in working hard and long. Personally he works 80-100 hours per week and has been doing that week-in week-out for years. To quote Musk: “Work hard every waking hour. If you do simple math, like somebody else is working 50 hours, and you’re working 100, you’ll get twice as much done in the course of a year as the other person.”

Yet SpaceX is more than just one person. SpaceX now consists of a team of 7000+ highly intelligent employees working perhaps 55-70+ hour weeks. In a recent survey ² SpaceX was listed as the most stressful (and one of the lower paying) places to work. However, it was also considered the most meaningful place to work.

Perhaps there are at least three reasons why SpaceX requires their team to work hard:

Liability:
Firstly, designing and building large Rockets is inherently hard. As well as being complex machines, they are also considered “critical systems”. If they blow up (which they tend to do if everything is not perfect), they pretty-much vapourise everything on board - including cargo and the lives of any souls on board. Given this liability, it is not acceptable for any staff to adopt a “she’ll be right” attitude.
Competition:
Secondly, how do you compete with existing aerospace competitors such as Boeing, who have existed / have over 100 years of experience in the industry? The simple answer is that you beat them by being better - by being innovative and disruptive. However, such innovation doesn’t happen overnight and requires long hours undertaking research and development - including systems development, prototyping, testing, optimising and more testing.
(NB: Incidentally, the “test phase” accounted for as much as 50% of the entire Saturn rocket program - in terms of allotted man-hours and physical resource (Bilstein, p.184.)).
Time is money:
Thirdly, and rather obviously, developing Rockets is an expensive game. Musk invested almost his entire fortune ($160M from when eBay purchased PayPal) into SpaceX (and Tesla Motors). As in any business, generally the longer it takes to commercialise a product, the more investment it will take, and the longer it will be before you start receiving a return on investment.
The first three SpaceX rockets exploded, and the company reached a point where it could barely make its payroll. In fact, they only had enough money for one more launch attempt, which (thankfully) was successful. Profitable NASA launch sub-contracts ensued.

As an aside, it is interesting that new research seems to suggest that working excessive hours often doesn’t make you more productive ³ . However, long hours is Musk’s style and this is unlikely to change anytime soon. Long hours or not, should SpaceX staff have taken a laid-back attitude, it is almost certain that the company wouldn’t exist today let alone be successful.

4: Unifying Company Goal

So why would anybody want to work at SpaceX? Sure, designing and building rockets sounds like a lot of fun - but why on earth would anyone want to endure the copious work hours, continual state of stress and enormous responsibility of launching ridiculously expensive cargo (let alone human beings)?

The answer in part, is that when you work for SpaceX, you are working for a bigger purpose - something not of this earth. Namely, Elon has set a very clear unifying goal for the the company:

Reducing space transportation costs and enabling the colonization of Mars.

Musk believes that humans absolutely must begin to colonise other planets - or risk becoming extinct in the near or far future. The only planet that is (somewhat) viable for colonisation in our Solar System is Mars, and he is highly critical of the fact that in the last ~40 years nobody (NASA or otherwise) has been bold enough to attempt to venture humans beyond the orbit of the moon.

Figure 3. SpaceX has a very clear objective to enable the colonization of Mars.

As a continual reminder to all of their staff and visitors, hanging on a wall of SpaceX’s headquarters in California, are two giant posters of Mars. One, as the barren planet exists today, and the other showing Mars a transformed planet with green landmasses and blue oceans (Vance, 2015, p.4).

It’s a long term goal, it’s a costly goal, it’s an audacious goal, but it’s also a unifying and motivating goal - one which makes clear to each and every staff member why they are ultimately turning up to work every day.

5: Risk Big

Finally, SpaceX (or should I say Musk?), have been willing to take big risks in order to be successful. They continue to do so.

Yes, the timing was right; Yes, Musk and his team of excellent workers were willing to work hard and implement highly innovative solutions and Yes, they also had a unifying company goal - but while all of these traits are important, the entire venture is still fraught with considerable risk. Consider the following:

Financial risk - What if their initial launch #4 was a failure? SpaceX would have simply run out of cash and vanished into oblivion.
Technical risk - What if the whole concept of reusing a rocket booster was fundamentally technically flawed?
Political risk - What if the Government (NASA) chose to not subcontract to new launch provider entrants, thereby leaving no early adopters to continue to fund their operation.
Social risk - What if nobody buys into their vision of colonising Mars and SpaceX thereby losing their unifying company Goal, and perhaps many stakeholders in the process?

There have probably been thousands of “what ifs” that SpaceX and Musk have had to consider since 2002. NB: Before Elon founded the company, apparently all his friends tried to talk him out of it - citing that he was likely to lose his entire multi-million dollar fortune. (They were almost right).

However, risk-big he did and risk-big he continues to do.

In all likelihood, SpaceX would have failed long ago, under the hypothetical leadership of most human beings. Many leaders would been crippled with fear at the magnitude of the risks in front of them - and likely shut the entire operation down. But Musk stood fast and continued to unabatedly believe in his company in the value of his product, and that success would ultimately follow.

Of course, it’s possible that all it will take is a few failed launches (especially if those launches involves astronauts) to end SpaceX. It’s the nature of their industry. They’re aware of this potential risk but instead of being crippled, they continue to be entrepreneurial and forge ahead toward their company goal.

Conclusion

Having read the article, hopefully you have gained some useful startup insights from a company that has “been there, done that”. That is not to say you have to personally agree with or adopt these traits, nor does it mean that if you do - success will be 100% guaranteed.

Nevertheless, it’s almost always invaluable to learn from the insights of others, and to take time to consider to what extent you identify with them.

Questions for the entrepreneur or business start up manager :

How innovative do you consider your business (or potential business) to be?
How hard are you and your team willing to work?
How confident are you in the timing of your product or service into the market.
Do you have a unifying company goal?
Finally, what is your appetite for risk?

Beta Solutions are privileged to work with both startups and established enterprises. If you have a business idea in mind that requires product development, get in touch with us today and we would be more than happy to bounce some ideas around and share with you our own insights.

References

Figure 1. Global Commercial Launch by Market, retrieved from https://www.hq.nasa.gov/legislative/hearings/7-13-17%20HUGHES.pdf
Figure 2. Satellites launched in 2017. Data retrieved from http://spacelaunchreport.com/log2017.html
Superscript 1. Department of Defense Fiscal Year (FY) 2018 Budget Estimates, May 2017 (p. 109). Retrieved from http://www.saffm.hq.af.mil/Portals/84/documents/Air%20Force%20Space%20Procurement%20FY18.pdf?ver=2017-
Superscript 2. Retrieved from https://electrek.co/2016/03/07/tesla-and-spacex-standout-in-tech-employee-survey-for-the-most-stressful-and-lowest-paying-jobs-but-also-most-meaningful/
Superscript 3. Retrieved from http://ftp.iza.org/dp8129.pdf
Image https://www.inverse.com/article/17723-spacex-mars-elon-musk-2025-colony-mission
Bilstein, R.E. Stages to Saturn, (p. 184).
Vance, A. (2015). Elon Musk. (p. 4). London, UK: Penguin Random House UK.
SpaceX images retrieved from http://www.spacex.com/galleries

Kids at Work Day

Thu, 14 Dec 2017 21:05:55 GMT

Author: Brenda Wormgoor, Marketing and Operational Manager

I remember hanging onto my Dad’s every word when he talked about his work and the people he worked with. That’s where he spent all his time away from us and it was a big part of who he was after all. I always marvelled at the pride he took in his work and the products they made. I also remember the way he knew small details about his co-workers’ lives, like the names of their wives and children, what their interests were, and things that mattered to them socially and culturally.

Today, I can still vividly remember the first time he took me to his work. I got to experience first-hand what his work was all about - how they made their products, what the factory looked and smelled like, the atmosphere and ambiance of hard working people going about their jobs diligently, seeing the well-oiled production process and feeling a sense of pride when the perfect final products came off the line. But I think what impressed me the most was the way they all communicated with each other, the professionalism, evidence of mutual respect, and easy-going camaraderie with a bit of humour. It was clear that the way he interacted with his colleagues as their leader, that he practiced what he preached at home.

Hence my excitement when our Technical Director purchased some kid-friendly electronics kit projects and we decided to have an afternoon at work for our children to see what we do. I hoped that this experience would be as significant to my son as it was for me.

It’s ALL about RELATIONSHIPS

At Beta Solutions, “Relationships” is one of our core values. On our website, we define “ relationships ” as:
“We are people people. Our clients and our staff matter to us. We cultivate these relationships and combine our individual strengths to make great things happen.”

To design the most intelligent electronic product solutions for our clients, our employees need to be given the chance to be creative, innovative, and to maximise their protentional in an inspiring and caring environment. We firmly believe that to achieve this, a healthy work-life-balance is a non-negotiable. Management at Beta Solutions strive to create a company that respects everyone’s home lives and their families. We care about our employees’ wellbeing beyond the office, wanting them to be successful in all areas of their lives. It is important that they have time to spend with their families, pursue their passions and causes they care about.

So, when we had our "kids at work” afternoon, it was as much about recognizing and celebrating the importance of our family and work relationships, as it was about introducing our kids to what we do at work.

Highlights of the Day

After sharing lunch and laughter, we moved on the “work” of assembling the electronic products. It was amazing to see how they soaked up the experience, soldering like they have been doing it all their lives.

They relished the opportunity to explore all the equipment (from the 3D printer to the Oscilloscope) in our electronics lab. The wonder and delight in the end products they produced were priceless. To celebrate the new friendships they created, we ended the afternoon with a fun game of table tennis (just like the employees enjoy a competitive game to end off the week every Friday).

he video below is a fun peak at how the afternoon unfolded.

It was heart-warming when my six-year-old son excitedly shared the ladybug with its LED lights he created as his oral presentation topic in class. And when their explore topic for the week at school was about inventions, he proudly shared that his mum worked at a place where they developed new products too.

It was an honour to introduce my son to my work family and their kids and it is a privilege to work at a place where my and my family’s wellbeing are valued as much as the work I do.

Feel free to contact us if you would like to learn more about us.

LPWAN Benefits for IoT Connectivity

Thu, 30 Nov 2017 20:56:10 GMT

Author: Matthew van der Werff, CTO and Senior Engineer

Due to an increasing demand for an “interconnected world”, by the year 2030 ¹ it is predicted that 130,000,000,000 (or if you prefer words: one hundred and thirty billion) devices will be connected to the internet. This is often referred to as the “Internet of Things” or IoT.

Low Power Wide Area Networks (LPWAN) are emerging as a promising and exciting new communications technology which will likely be a key player in helping achieve these tremendously large numbers of connected devices.

The aim of this blog is to provide the reader with a broad overview of LPWAN and to highlight some its significant benefits over existing technologies.

Communications Technologies Defined

Before we look at LPWAN specifically, it would be helpful to provide a basic overview of the general way in which devices connect to the internet.

Simply put - and as readers likely already know - connecting a device to the internet involves sending electronic signals between a Device (“Device”) and The Internet “Server”, in a format that makes sense to both the Device and the Server.

This requires a combination of (i) Electronic Hardware - which physically sends and receives these signals, and (ii) Electronic Software – which controls and orders these signals into an agreed format.

Incidentally, Devices do not connect to the Server directly, but first must connect to a piece of hardware called a “Gateway” which as the name implies, acts as the middleman between the Server and the Device. Multiple Devices can connect to the same Gateway, and there might be multiple Gateways all connecting to the internet.

Figure 1: Device communication/network structure.

So “Communications Technologies” is an umbrella term which refers to both the Hardware and Software which together control the flow of information between end points on a network. Sometimes all that is required to introduce a new communications technology on to the market is a simple software upgrade, while in other cases completely new hardware and software will be required.

Presently, a few different wireless Communications Technologies already exist which can be utilised to connect Devices to the internet, for example:

Cellular
Satellite
Bluetooth
WiFi
Zwave
Zigbee, etc.

LPWAN is a relatively new player to the market but is expected to quickly become a key technology in the IoT sector moving forward.

What is LPWAN?

LPWAN is an abbreviation for “Low Power Wide Area Network”. The term itself implies a generic communications technology as opposed to a fully prescribed communications protocol.

We can break the term down into the following two parts:

Low-Power - designed to enable devices to operate from small batteries and last 5-10 years
Wide-Area-Network – designed to communicate over a long range. Approximately 2 km – 20 km +

Although low power technologies already exist (eg: Bluetooth LE) and although long range technologies already exist (eg: Cellular), until now there has not been a technology which is able combine both of these properties.

Figure 2 ² : Sensor range and battery life comparison.

At this stage you might be inclined to think this technology sounds incredibly amazing, but as the saying goes - “you can’t have your cake and eat it too”. Essentially, in order to achieve both Low Power AND long range a trade-off needs to occur, which is that LPWAN technologies are only able to transmit very small amounts of data.

While this makes it decidedly unsuitable for high bandwidth demands such as audio and video, it is still completely adequate for the plethora of applications requiring low data rates. For example, most water tanks do not need to indicate water levels continuously if there has been no level change.

Within the broad LPWAN technology, there are already a number of more specified technologies which include: LoRa, Sigfox, Weightless-N, and a number of others.

What are the Benefits of LPWAN?

LPWAN provides the following benefits:

Low power usage
The power your Device will consume is highly dependent on how often and how much data is transmitted. However, provided the device does not transmit excessively it is reasonable to expect a battery life of 5-10 years.
NB: Technically, the microprocessor is put into sleep mode when the devices are not operating – drawing very little current.
Low cost / long lasting batteries
Due to the low overall power requirement, these IoT devices can now utilise relatively Low cost batteries – reducing the overall cost of the product.
Technically, in addition to the low overall power requirement, another beneficial property of LPWAN technologies is the low “peak power” requirements. Traditional cellular modems require a very high peak current usage, restricting the battery technologies to more expensive batteries. LPWAN’s relatively low peak power requirements, enables lower cost batteries to be used.
Long distance communications
LPWAN can enable communications at distances of between 3-20 km – which is impressive considering the low power requirements. In reality the actual distance it can communicate is highly contextual. In cities, due to buildings and increased electrical interference, 2-5 km can be expected. In rural environments, with a clear line of sight, 15-30 km is likely achievable. If fact, the Long Range (LoRa) record3 is 702 km, although this was from a weather balloon so we think this is cheating a little.
Low number of Gateways
A positive side-effect of being able to communicate long distances, is that a relatively low number of Gateways are required. This in turn reduces outright infrastructure cost. For example, the entire city of Amsterdam is mostly covered with only around 24 Gateways, which is not bad considering the high density of buildings etc. In rural environments (such as New Zealand farms), it is reasonable to expect 15 km of coverage, and therefore Gateways could be even be shared between farms.
Free radio spectrum
LPWAN operates in the free radio spectrum. This means anyone can transmit and receive on it without requiring a licence or payment for use of this spectrum. This has the advantage of (i) enabling anyone to be a Gateway provider (ii) potentially having zero data costs.
However, one potential downside of using the free spectrum is that a high number of other devices will also operate on the same frequencies – leading to possible “clogging” of the network.

How Does it Work?

If you are interested in a slightly more technical description of how LPWAN works … read on.

How long range is achieved:

Most LPWANs use narrow bandwidth transmissions to achieve long range and some technologies (like LoRa), will also implement some advanced signal processing techniques to further their range.

A basic principle is that the slower you transmit your data the further you can transmit your data.

How slow is LPWAN? Well to put it into perspective, Sigfox can take up to 2 seconds to transmit your tiny 12-byte data packet. A 3G cellular modem would transmit 12 bytes in a fraction of that time. (Data rate of Sigfox is approx 100 bits per second, while 3G cellular can be up to 10,000,000+ bits per second).

How low power is achieved :

Low Power is achieved in LPWAN systems (in part) by specifying that only the device can initiate communication (and not the Server side). The device might communicate 100 times a day or once per day, but it cannot transmit “on demand” from the Server. Therefore, the device does not need to consume considerable electrical power by continuously “listening” to a possible communication from the Server.
The Gateway on the other hand, must always be listening to a possible communication from the Device, but as the Gateways are likely plugged into electrical mains, power consumption of the Gateways is not likely an issue.

This is very different from traditional cellular technologies where the device needs to stay connected to the tower continuously – both to send data and to also listen for more data coming in.

Topology :

A LPWAN network operates with a star topology. There are multiple Gateways that are listening for a device's data. This data is then sent to a network service which then sends a reply back to the Device though the closest Gateway.

The network service also sends the device’s data onto an application Server on the internet.

Figure 3: Standard LPWAN network topology.

What are the Existing Technologies (on the market)?

As previously mentioned, there are several LPWAN technologies already available – including Sigfox, LoRa, NBIoT.

So what technology should you use? Well to that question, at this stage, there is no clear answer as it depends on:

How much data you are wanting to transmit
Where your devices are to be located
Whether you need to add your own network/Gateways in some situations (such as remote farms)
What your target costs are – including the cost of devices and the cost of data (Some of which remain to be determined)
The degree of autonomy you want over your network
Sigfox is more closed
LoRaWan is more open

Overview of the Main Long-Range Communication Technologies

A brief Overview of LoRa Communication Technologies on offer at the moment, can be downloaded here.

Conclusion

LPWAN is an exciting new technology - and we are expecting a number of infrastructure rollouts in 2018 in NZ (eg: Spark and Vodafone) and world wide.

If you are requiring remote sensors that are compatible with low data rates, then LPWAN would certainly be worth considering.

Beta Solutions has experience in developing products containing various communications technologies, including: LoRaWAN, Sigfox, Cellular, Bluetooth, Wifi, and Satellite communication.

If you are after more information, subscribe to our Blog for updates as the next Blog will give further technical details on LoRaWAN. Alternatively you can get in touch via our contact page or give us a call .

References:

IoT News, Number of connected IoT devices to hit 125 billion by 2030, says IHS Markit. Retrieved from https://www.iottechnews.com/news/2017/oct/26/number-connected-iot-devices-125-billion-2030-ihs-markit/
Sensor range and battery life comparison, retrieved from https://kotahi.net/network/
Ground breaking world record! LoRaWAN packet received at 702 km (436 miles) distance (08-09-2017). Retrieved from https://www.thethingsnetwork.org/article/ground-breaking-world-record-lorawan-packet-received-at-702-km-436-miles-distance
References to blog posts in https://iot-daily.com/2015/03/13/sigfox-pros-and-cons/
Let’s talk IoT – NB-IoT/eMTC power saving features: eDRX vs. PSM https://www.youtube.com/watch?v=1j9lYr95qwU
Graphic Illustrations by https://www.vecteezy.com/

A Closer Look at the New Bluetooth Mesh

Thu, 26 Oct 2017 20:44:36 GMT

Author: Morten Kirs, Electronics Engineer

After a few years of being "around the corner" Bluetooth Mesh is finally here. Several product designers and developers have made their own meshes based on Bluetooth LE in the long wait for a standard (indeed we have also developed our own). However, we no longer need to rely on custom requirements. The new specification aims to be another contender in the Smart Home and Industrial Automation IoT markets, which are currently dominated by a plethora of other protocols (e.g. ZigBee, Z-Wave, WiFi, Thread). Shortly after its release earlier this year major players in the Bluetooth Special Interest Group (Bluetooth SIG) had software ready to go. With all the hype we are sure that the mesh will be implemented across various industries and applications. But what exactly is this new mesh, and why would you want it implemented in your products over other wireless protocols?

The New Mesh

Traditional communication is often point to point, whether it be a direct connection from your phone to a wireless headset, or a more complex connection from your laptop to a server which runs through various routers, switches, and adapters. Data often travels through a set route that is determined by central controllers and gateways in the network. Mesh networks tend to move away from the dedicated router approach. Devices themselves are allowed to act as relays of information. When data needs to travel between points in a network, the data will move through other untargeted devices to get to the final destination. This is the world of IoT. Below is a diagram of a mesh network when compared with a more traditional point to point system:

Figure 1. A very simplistic view of a mesh system compared to a more standard point to point protocol. In the mesh nodes will re-transmit data to other nodes that are further away.
Figure 2. Point to point, or point to multi-point systems will often just be a one to one or one to many connection, such as a connection from your wireless mouse, keyboard, or headset to a central laptop.

The Bluetooth Mesh implements a new form of mesh protocol, and it does it well through existing Bluetooth LE hardware. This means it's very likely your Smart Phone will implement some aspects of the new mesh in the near future. You'll likely be able to connect to your Bluetooth Smart Home devices without the need for a central hub. However, from a more technical point of view, this means implementing a Bluetooth Mesh has a fair bit of complexity attached when it comes to laying out a network. With IoT also comes the concern of security which is handled much like the traditional Bluetooth security features that we're used to with "authentication" and "pairing" protocols, among new "provisioning" protocols.
If you're just curious about the new technology and how it fits into the IoT landscape, skip ahead to the "Comparison with Alternative Technologies" section. However, for the more technical readers we'll now get to the nuts and bolts of the mesh, where we'll take a slightly more in depth look at 1. The mesh topology in general, 2. The communication protocol between nodes, and 3. Some of the security features in place.

1. Bluetooth Mesh Topology

The Bluetooth Mesh is a flood mesh. When a message is transmitted across the network, all devices (nodes) that are configured as relays will re-transmit the message. A message will only have a limited number of "hops", or re-transmissions, before it stops. This extends the range of Bluetooth LE to incredible distances, where the specification gives a theoretical limit of up to 127 hops, otherwise known as a message's Time To Live (TTL). The network can be described by the following diagram:

*Figure 2. An example Bluetooth Mesh network topology. It is fairly complex, with various features that could be supported, or not supported by any one node.

Every node can send or receive messages across the mesh, but there are also additional mesh related features that a node could support. The four features are:

Relay Feature
A relay is able to re-transmit messages it receives to the rest of the network through the advertisment bearer.
Proxy Feature
Provides the relay feature, as well as allows messages to be re-transmitted through the GATT bearer. This would allow devices such as phones to communicate with the mesh through a standard Bluetooth LE GATT protocol.
Low-power Feature
A node that operates at significantly low power, requiring a "friend" node to store messages
Friend Feature
A node that stores messages for low-power nodes, and forwards these only when the low-power node requests them.

Proxy nodes include a standard Bluetooth LE GATT service with two Proxy Data attributes. A Smart Phone could connect to any Proxy node to transmit messages to the rest of the network through these attributes via a "Proxy Protocol".

Low-power nodes require friend nodes in the mesh. A "friendship" is negotiated when a friend node sees a friendship request being advertised by a low power node on the network. If a friendship is established a friend will hold messages for the low power node until it asks for them. This allows a low-power node to turn off its receiver a majority of the time and achieve very low power consumption. However, other nodes on the network are expected to receive messages in real time, requiring higher power. This means they would likely require continuous external power, or much more regular battery replacement.

2. Bluetooth Mesh Messaging

The mesh network moves away from the well-known GATT model for Bluetooth LE. It implements unique data structures called "elements". Every element must have one or more models, where the models determine and implement the behaviour of a node, while the states inside them define the current condition of the element.

*Figure 3. Pictorial representation of the new element model. Nodes can have multiple elements with multiple models. States within these models define the current state of a device.

E.g. A lightbulb:

- Will have a generic "lightbulb" element, but could also have temperature elements, and so on.
- The lightbulb element could have an "On/Off" model, and a "Brightness" model.
- Each of these models will have their own unique states, where the On/Off would have a boolean "on" or "off" state, whereas the brightness could have a state value of 0 to 10.

Figure 4. Example lightbulb node.

The mesh also supports composite states, where two or more values can make up a state, allowing you to define data structures. States can also be bound together by a model, where in the example above, setting a brightness could also turn on the lightbulb.

The network communicates via the same client-server architecture Bluetooth developers are used to. Devices acting as clients can read states or write messages to elements on servers. Models will determine whether there is an action taken for a given message, and whether a state change will occur. A publish-subscribe mechanism is in place, where devices can publish messages to certain addresses. If another device is subscribed to messages on that address, it will receive the message. These messages are allowed to be either acknowledged or unacknowledged.
The mesh network communicates via two different messages.

Control Messages:
Mesh related messages related to things like regular heartbeats and friend requests.
Access Messages:
Used to retrieve, set, and report the values of states in server models.

For addressing of messages, there are a few options in place.

Unicast Address:
Every element in the network has a unique unicast address. This achieves basic point-to-point communication
Group Address:
These are addresses that group multiple elements across one or more devices. There are custom dynamic groupings which can be set up by a user on the fly, as well as a few Bluetooth specified group addresses including "All Proxies," "All Friends," "All Relays," and "All Nodes" in the network. It is expected meshes will use group addresses to address things like different rooms or floors of a building.
Virtual Address:
A hashed address of specific Label UUIDs. A Label UUID is a generic 128 bit value associated with one or more elements in a network. These are expected to be vendor specific addresses, where vendors can identify/address their own devices within a network.

Previous to setup, elements will generally have unassigned addresses. Elements with an unassigned address are banned from communicating with the network. When a mesh connection is established, a provisioner will assign appropriate addresses (see the next section for a brief overview of provisioning).

3. Security of the Bluetooth Mesh

Devices can be added to or removed from a mesh via a new process called "provisioning." Devices that can be added to a mesh will act as beacons and advertise their intent to join a network. Nodes acting as provisioners will scan for these beacons and extend an invite to the network. A pairing process similar to standard Bluetooth LE is followed, where keys are exchanged and an authentication protocol is chosen based on the input and output capabilities of the device. The standard example is having a screen on one device display a number that is entered as a passkey on another device.

There are three types of keys present in the network, giving multiple layers of security

Network key
A key shared by nodes on a mesh. This key will be used across all devices on a network to transmit and relay information. A device could store one or more network keys, and therefore be part of one or more subnets.
Application key
Bluetooth Mesh uses a concept of "Separation of Concerns." Essentially the idea is that a device such as a light bulb has no business knowing the content of messages sent to door locks. A separate application key can be used to share information across the network for specific devices. Unrelated relays will still re-transmit these messages, but will have no idea of the content of these messages.
Device key
A key unique to the device and provisioner.

Overall the Bluetooth Mesh is expected to be a very secure network. In addition to the keys mentioned above which keep data private and secure from listeners, the specification also provides protection against:

Replay attacks
A replay attack is an attack where a message that has been previously sent across the network is simply repeated (e.g. An "open door" message is seen by an attacker and is used later when the homeowner has left). A Bluetooth Mesh handles this through the use of sequence numbers. Devices have unique counters associated with them that are transmitted and incremented with each message. For a message to be considered valid by another device, it will need to see an incremented count on the given source address.
Man-in-the-Middle attacks
A man-in-the middle attack occurs when an attacker pretends to be another device. When a user tries to connect to a legitimate device it could instead connect to the attacker, posing a risk to exchanging critical security data. The Bluetooth Mesh has a few guards for this. Authentication is often required to add devices to a network, requiring "Out of Band" keys. A user can validate the correct device is being connected to through the input of a displayed passkey, or through other procedures. Furthermore, the use of public and private keys at the time of making a connection through the "Elliptic Curve Diffie-Hellman" protocol prevents valid encrypted data from being seen by an attacker.
Trashcan Attacks
Network and application keys are stored on devices semi-permanently. When a device is thrown out or sold, it is important these are removed from the network by a provisioner. The provisioner blacklists the device and negotiates new application and network keys to be distributed through a network. As it takes time for low power devices to receive the keys, this creates a period where both old and new keys are considered valid on the network, a time which Bluetooth calls "Phase 2". At the end of this phase all old keys become invalid.

Comparison with Alternative Technologies

Bluetooth Mesh targets the Smart Home lighting market with its first release. Much like it did for many services and characteristics in Bluetooth LE, the Bluetooth SIG has released specifications for exactly how a lightbulb, among other devices, should look and act on a mesh network. This ensures interoperability between devices, and also future-proofs products, as a lightbulb made 20 years from now will likely look like it currently does on the mesh.

However, the Smart Home market is already dominated by ZigBee and Z-Wave products. Wi-Fi products also exist, Google supports its own Thread protocol, and other lesser known protocols can also be found floating around. With all these options already in place it's somewhat difficult to see how the Bluetooth Mesh will fit in. Below is a brief overview of some of the benefits and down-sides of the various technologies

ZigBee

Pros:

Scalability. Very high number of nodes possible in a network.
Power. Very low power consumption.
Support. Well supported, and been in the market for a long time, providing simple and fast development times.

Cons:

Requires a network co-ordinator, creating a single point of failure (Though later versions of Zigbee (3.0+) now support a distributed network model)
Interoperability. Although Zigbee specifies protocols for various devices, these are not strictly followed by all manufacturers. Some devices will simply be incompatible.
Security. Security vulnerabilities due to standards not being strictly followed by all manufacturers. The standards themselves provide robust security measures. However. a single vulnerable device can compromise the whole system.

Z-Wave

Pros:

Less Interference. Operates outside of the standard 2.4GHz band which is already swamped by WiFi, Zigbee, and Bluetooth products.
Interoperability. Requires developers to be part of the Z-Wave Alliance (Confidence that all devices are tested, secure, and interoperable)
Power. Very low power consumption. Routing tables used for efficient communication.

Cons:

Requires a primary controller
Interoperability. Not guaranteed to work across countries due to different operating frequencies and regulations in these countries.
Price. Tends to be more costly than alternative technology
Scalability. Supports 232 nodes, which is low compared to other technologies, but sufficient for most applications
Development costs and time . Requires developers to be part of the Z-Wave Alliance (limits time to market, requires extensive testing)
Data rates . Relatively slow data rates.

WiFi

Pros:

Data rates. High bandwidth; Can stream data at very high rates
Integration . Interfaceable to smart phones, tablets, pc's, etc.

Cons:

Power. Very power hungry, where battery operated devices would require constant replacement
Remotely Hackable. WiFi networks are often directly connected to the internet, creating vulnerabilities for malicious remote access.
Setup Complexity. Initial setup of devices is often complex, requiring custom methods to transfer WiFi passwords and network data previous to joining a network.
Dedicated routers. Requires dedicated routers, repeaters, and other network extenders.

Bluetooth Mesh

Pros:

Scaleable. Up to 32,767 devices
Interoperability. Well-defined models ensure interoperability, and future-proofing
Existing Bluetooth hardware. Bluetooth devices are everywhere - Existing Bluetooth LE (4.x /5.0) device hardware should support some aspects of the mesh, likely only requiring a software update. This includes Smart Phones which should allow for easy mesh setup and device configuring.
Range. Long range between nodes (though range is dependent on hardware)
Data rates. Relatively high data rates. (Though not high enough to support audio streams like Bluetooth BR/EDR).

Cons:

Interoperability. Existing Smart Home hubs do not all currently support Bluetooth LE, reducing interoperability between devices.
Complexity. Fairly complex setup, requiring some understanding of node functionality and supported features
Unknown Risks. An emerging technology. Unknown compatibilities with existing devices, and popularity.
Wasted Power. A flood mesh, meaning transmissions are mostly wasted in larger networks. Relays need to be powered or require regular battery replacement.

Conclusion

The Bluetooth Mesh looks like it will be a solid contender in the IoT market. The protocol provides security and scalability for a range of home and industrial applications. However, the Smart Home market is already largely dominated by ZigBee and Z-wave devices, among other protocols. At a first glance, the complexity of the mesh topology can be off-putting for a new automation system, where a user would likely require some understanding of the various node features and requirements. Nevertheless, perhaps the main advantage for Bluetooth Mesh is that Bluetooth devices are already everywhere, and the mesh can be applied to existing hardware with software updates. Support for the mesh has already started, with modules and software updates already available from major players in the Bluetooth market. It should only be a matter of time until a software update comes along to allow your Smart Phone to implement and control some mesh related features. With all the hype it will be interesting to see how widely this new protocol will be adopted in new as well as existing products.

We'd be excited to hear how this new technology can be used in your Smart Home or other ideas. If you're interested in bringing your product idea to life, please do get in touch via our contact page or give us a call .

References

An example Bluetooth Mesh network topology. Retrieved from https://blog.bluetooth.com/bluetooth-mesh-networking-series-friendship
Pictorial representation of the new element model. Retrieved from https://blog.bluetooth.com/bluetooth-mesh-networking-series-friendship
References to Blog Posts in https://blog.bluetooth.com/category/bluetooth-mesh?_ga=2.97073119.1773091513.1508124072-1445221796.1477622861
Mesh Networking Specifications. Retrieved from https://www.bluetooth.com/specifications/mesh-specifications
Z-wave vs Zigbee vs Bluetooth vs Wifi: Which Smart Home Technology is Best For Your Situation? Retrieved from https://inovelli.com/z-wave-vs-zigbee-vs-bluetooth-vs-wifi-smart-home-technology/
Z Wave Vs ZigBee: Which Is Better For Your Smart Home? Retrieved from https://thesmartcave.com/z-wave-vs-zigbee-home-automation/
What’s The Difference Between ZigBee And Z-Wave? Retrieved from http://www.electronicdesign.com/communications/what-s-difference-between-zigbee-and-z-wave
ZigBee, Z-Wave, Thread and WeMo: What's the Difference? Retrieved from https://www.tomsguide.com/us/smart-home-wireless-network-primer,news-21085.html

Electronic Product Compliance - Answers to Frequently Asked Questions

Wed, 04 Oct 2017 20:21:12 GMT

Author: Jason Cleland, Electronics Engineer

What is compliance?

Inventing and developing the next big electronic product is a significant step but is not the last step before that product can go to market. New Zealand, and most overseas markets have laws and regulations which regulate what can be sold in that market. These regulations specify all the legal requirements for which a product must comply with.

Any given product can fall within the scope of multiple regulations which each govern its sale. The two most common regulations for electronic products in New Zealand are:

Electricity (Safety) Regulations 2010
Radiocommunications Regulations 2001

Compliance with regulations often means complying with relevant Standards. Standards are documents which set out the specifications, procedures and guidelines for products within their scope. Standards are used extensively during the design process. As examples, AS/NZS 3100 specifies the general requirements for electrical equipment and AS/NZS 60950 specifies the requirements for information technology equipment (E.g. laptops, and phones). Once the design is finished, the product is then tested to show compliance. Higher risk products require that they be tested by testing laboratories which are accreditted.

Products sold in overseas markets might have to comply with FCC, CE, UL, or other regulations and standards.

Who must comply?

Anyone who makes or supplies electrical, electronic, or radio products to the New Zealand market must ensure their products comply with the regulations.

New Zealand's Electrical Equipment Safety System (EESS) and the Radio Spectrum Management’s (RSM) product standard framework is based on the principle of supplier self-declaration. Suppliers must take responsibility for the products they place on the market, whether they are imported or domestically produced.

Why should you concern yourself with compliance?

Complying with the requirements set in these regulations can sometimes be a time-consuming and expensive process. Given that, it can be easy to think that we would be better off without them - as products could be put to market faster, which can be critical for a new product's success. However, it needs to be remembered that the purpose of these regulations is to protect the consumer.

Undertaking compliance goes a long way to ensuring that your product will not cause harm to other products, the environment, and most importantly to people. Hence, investing in compliance upfront minimises your risk and liability.

How can I know what I must do?

Compliance can be a tremendously complex area, often requiring one to be familiar with hundreds (or thousands) of pages of regulations. Therefore, it is often prudent to use the services of a professional electronics design consulting firm to take the burden off you.

To ensure our clients' products comply with regulations, we at Beta Solutions, work with the testing laboratories themselves to identify all the requirements at the start of the design cycle. This makes it possible for our engineers to consider the implications for compliance with every decision. We aim to design our clients' electronic products so that they pass the compliance process first time. Beta Solutions has years of experience with complying with standards which enables us to create products which are safe and do not cause interference.

What is the compliance process?

The process of conformity varies across different product categories and markets but generally follows the steps below:

Identify the markets which the product will be sold in.
Investigate the regulations for each market and identify all the requirements for your product.
Assemble a compliance folder with evidence that shows that your product complies with all requirements (E.g. Technical test reports).
If required, submit this evidence to an approved body who will review it and generate a compliance certificate.
If required, register your product in the relevant database.
Put the compliance mark(s) on your product.
Your product is ready for sale.

How can compliance costs be minimised?

Compliance costs have a roughly proportional relationship with the level of risk inherent to the product. For example, products which have voltages or energies which have the potential to harm a user need to be designed and tested to a higher standard. Likewise, products which have features which could radiate (intentionally or unintentionally) harmful radio energy also need to be tested to an agreed standard so as not to adversely affect other devices.

Here are three ways to reduce compliance costs by reducing risk:

Think Simple .

System complexity is also roughly proportional to compliance costs. Products which incorporate more features commonly require more testing. For instance, a product which has a wireless radio, is powered by mains, has live parts which the user can touch, and operates in an explosive environment must be compliant with IECEx, EMC, Radio, and Electricity Safety schemes and will require extensive compliance testing.

Implementing a “minimal viable approach” can reduce complexity/risk and therefore compliance costs. For example, if it is not absolutely necessary for a product to have a wireless radio then it might be decided to omit this feature – at least for the initial product release. Additional features can very likely be added at a later date for future product versions.

Design Modular.

One method to reduce compliance costs in electronic product design is to use “commercial off the shelf” (COTS) pre-compliant modules.

For example, it is common practice for manufacturers to supply their products with an external mains power brick. This power supply, which can be reused across their product line, only needs to be tested once. Often already compliant power supplies are sourced from another manufacturer so that only the product needs to be tested. The device itself then contains no dangerous voltages and design time, and testing can be significantly reduced. Implementing COTS radio modules are often another way to reduce, or completely negate, the need for radio compliance testing.

Think Compliant.

The best time to start considering compliance is right at the very early stages of development. The implications of compliance issues arising in the later stages of a project could result in (i) costly redesigns (ii) delays (iii) a less than successful project. Hence, it is critical to consider the compliance requirements early on during the design process to reduce potential costs.

Conclusions

The laws in New Zealand, and many other countries, call for electronic products to comply with the requirements set out in strict safety, EMC, and radio standards. These requirements need to be considered at every stage during the development process.

The legal requirement for compliance is here to stay. Every electronic product sold in New Zealand must comply. Embrace it now to minimise liability and ensure that your new product is a success.

Compliance is a complex area, prehaps the best thing you can do is to use the services of a professional electronics design consulting firm to take the burden off you. At Beta Solutions, we have considerable experience in designing compliant products for our clients.

If you would like to know more about compliance in electronics product development, feel free to get in touch via our contact page or give us a call .

References

Electricity (Safety) Regulations 2010 - http://www.legislation.govt.nz/regulation/public/2010/0036/latest/DLM2763501.html
Radiocommunications Regulations 2001 - http://www.legislation.govt.nz/regulation/public/2001/0240/latest/DLM71513.html
Compliance Engineering. "Regulatory Compliance Mark (RCM)." Retrieved from http://www.compeng.com.au/wp-content/uploads/2014/10/Regulatory-Compliance-Mark2.pdf
Electrical Regulatory Authorities Council. "Australian/New Zealand Electrical Equipment Safety System . Equipment Safety Rules". (Version 1.1). Retrieved from http://www.erac.gov.au/images/Downloads/Equipment%20Safety%20Rules.pdf
Energy Safety. (23 June 2015). "What we do (and don’t do)." Retrieved from http://www.energysafety.govt.nz/about/who-is-energy-safety/what-we-do-and-dont-do
Energy Safety. (28 May 2012). "Compliance." Retrieved from http://www.energysafety.govt.nz/appliances-fittings/electrical-appliances-fittings/core-requirements/compliance
Percept. (2017). "What are Safety Certifications for electronic products?" Retrieved from http://www.percept.com/regulatorycompliance_safetycertifications
Radio Spectrum Management. (5 April 2017). "Compliance standards for EMC & Radio." Retrieved from https://www.rsm.govt.nz/compliance/supplier-requirements/compliance-standards-for-emc-radio
Standards New Zealand. (January, 2017). "Why standards? An introduction to the benefits of standards." Retrieved from https://www.standards.govt.nz/assets/About-us/Resources/Why-Standards-brochure-2016-v4-Web.pdf

New Zealand Trialling Centimetre Level GNSS

Thu, 07 Sep 2017 21:16:18 GMT

Author: Phillip Abplanalp, Hardware Engineer

Since the Global Positing System (GPS) was launched by the United States in the late 1970s, modern civilisation has developed an ever increasing dependence on global navigation systems. Think of the many GPS equipped devices and applications which we use on a daily basis, such as modern cars and smart phones with applications like turn by turn navigation and everyone’s favourite phone app – Pokémon go! Such applications do not require high levels of positional accuracy and therefore can utilize a single satellite receiver. However, numerous industrial applications such as land surveying, precision agriculture as well as emerging fields like precision drone control and autonomous driving, all demand much higher levels of positioning accuracy than what has been achievable by single receiver positioning in the past.

For this reason (among others) Land Information New Zealand (LINZ) announced in 2016 that it was collaborating with Geoscience Australia in a two year trial of some of the latest global navigation technologies - known as Satellite Based Augmented System (SBAS) and Precise Point Position (PPP).

Before we take a look into these new technologies we will review the methods which are presently used to enhance the positional accuracy of Global Navigation Satellite Systems (GNSS). This will highlight the advantages these new technologies will bring and see why these new technologies could open a number of new product avenues.

Standard Single Receiver Positioning

Many readers will be aware of the basics of standard GNSS positioning, but we will briefly recap this. GNSS consist of over 80 orbiting satellites (which includes GPS (American System), GLONASS (Russian system), Galileo (European System) and BeiDou (Chinese system)), the key components of these satellites are, extremely accurate atomic clocks and multiple transmitters. To allow a GNSS receiver (ground based device) to be able to calculate its position, it needs to know exactly where each satellite is in space at the precise time the GNSS signals are broadcasted from each satellite. For this reason, each satellite transmits among other things, time (at which the signal is sent), date, satellite health (whether it is currently in operation) and a unique ID known as the Pseudo-random Code Number (PRN). Since the speed at which the signal propagates through the atmosphere is known, the receiver can calculate the distance to that satellite. With a minimum of four viewable satellites and using both the known position of the satellites in space as well as the distance to each of the satellites, the receiver can resolve its position anywhere on earth using trilateration. A simple diagram of this can be seen below – Figure 1.

Figure 1 – GNSS receiver with GNSS satellites transmitting their information needed for receiver to calculate its position [2]

Whilst this system is effective and extremely powerful, allowing a single receiver to calculate its position anywhere on earth (assuming line of sight to satellite) there are significant positional errors, ranging from 1 – 10 m. These are largely due to the following effects: satellite clock drift, satellite orbital error and atmospheric effects from both the Ionosphere and troposphere with both affecting the signal propagation to the receiver. See the table below showing the contribution of these errors.

Figure 2 – Contributing factors to GNSS positional error [2]

Relative Position Correction or Differential GNSS (DGNSS)

Due to the errors associated with a single receiver (as outlined above), many industries have adopted methods of eliminating or reducing these errors. One of the common methods which was developed to overcome these downsides is known as Differential GNSS (DGNSS). DGNSS typically works by having a reference station in a known fixed location. The reference station includes both a GNSS receiver and some form of radio link (such as 433 MHz or GSM) connecting the reference station to the “rover”. Since the reference station has a known location any difference in the calculated position and its actual position can be considered as an error. This error or difference is then sent out to the rover via the radio link, allowing for the error to be subtracted at the rover – see illustration below. The typical accuracy using such a system will usually be around 1 - 2 m.

Note: This error is largely attributed to the system being unable to resolve higher levels of accuracy due to the inherently low resolution/frequency of the GNSS “binary code phase”.

Figure 3 – Diagram showing key features of DGNSS

To further improve the positional accuracy, a method of Real Time Kinematic (RTK) was developed in the mid-90s. Like DGNSS, RTK also makes use of a reference station to send correction data out to the rover. However, in essence this system overcomes the error introduced by low frequency/resolution of the binary code phase (~1Mhz), by utilizing the much higher carrier phase frequency (1575 MHz) to calculate its distance to a satellite – see Figure 4. This results in a substantial improvement, typically resulting in positional accuracy of 1 - 2 cm.

Figure 4 – Signal modulation used by GNSS satellites – note carrier wave is convolved with Binary Code to generate modulated signal

There are however a number of draw backs to these systems, not least the requirement of the local infrastructure, where a local reference station is required. This drives up the installation, product and maintenance cost, with traditional RTK systems (including receivers and reference stations) having been in the tens of thousands of dollars. For this reason, the applications have been limited to high value applications such as cropping tractors and surveying equipment. In addition to the high infrastructure cost, RTK also has limitations to its baseline (the distance between the rover and the reference station) – which is typically limited to 20 - 30 km. The limited “baseline” distance is a due the varying atmospheric (ionospheric and tropospheric) effects - which the reference station is accounting for - eventually have little correlation to what the rover is observing.

Satellite Base Augmented System (SBAS)

Much like DGNSS, the Satellite Based Augmented System (SBAS) also makes use of reference stations to provide correction data to the rover. However, rather than using a single reference station, SBAS makes use of several reference stations spread over the entire operational area (New Zealand’s trial makes use of 5 stations). By using several reference stations a map of the atmospheric effects is generated and this allows the system to cover large areas. Along with the atmospheric effects, the reference stations also allow for correction data of satellite orbital error and satellite clock errors to be accounted for. The correction information is then beamed to a geostationary satellite, which in turn broadcasts the correction data back to earth, allowing the rover to make a local correction. Since this correction data is sent over the existing GPS L1 channel (the original civilian channel - see figure 6) this can be received by existing GNSS receivers without the need for any additional antennas or hardware. This means that the existing single receiver GNSS devices could make use of this correction data with just a firmware update. Systems similar to SBAS have been available in the USA (Wide Area Augmentation System WAAS) and Europe for a number of years and allows for a positional accuracy of 1 - 2 m

Figure 5 – Diagram of SBAS core features [5]

In addition to standard SBAS, the LINZ and Australian Geoscience trial will also be testing the second generation SBAS system which uses the latest Dual Frequency Multi Constellation (DFMC) setup. DFMC uses both the GPS L1 and the most recently added L5 channel (see figure 6), by making use of both these channels it is expected to reduce the positional error to less than 1 m (the exact error is still unknown as this will be the first trial of its kind). Whilst the standard SBAS has been available in New Zealand since the 1st of June 2017, the SBAS DFMC trial is forecasted to start later this year. However, due to this being the first SBAS DFMC trial, there is presently no “Off the Shelf” hardware available to support it, furthermore there are still only a handful of satellites which support the GPS L5 channel and therefore may have limited use initially.

Figure 6 – Present GPS channels [2]

Precise Point Positioning PPP

Precise Point Positioning is an exciting technology which can achieve positional accuracy similar to RTK but using just a single receiver. Although there are no strict standards on PPP operation, the principal of PPP is essentially to use extremely precise satellite clock and orbital correction data which is collected and processed through a number of reference stations positioned globally (rather than locally for SBAS). The satellite correction data for the New Zealand trial will be broadcast along with the SBAS data over the L1, L2 and L5 channels. Along with the precision satellite correction data, dual frequency PPP devices takes advantage of the fact that ionospheric propagation delays are directly proportional to the frequency of the signal. This means that by using a dual frequency receiver, i.e. reading both L1, L2 and L5 (see figure 6,) the receiver has the ability to almost entirely remove the ionospheric errors.

The big advantage of dual frequency PPP is that its baseline (operational area) is global, this is because its atmospheric adjustments are no longer based on local reference stations but rather frequency dependent. Of the four positional technologies discussed, PPP shows the best baseline vs accuracy - see Figure 7 below. However, the drawback of PPP is the long convergence time (time to gain an accurate initial positional reading). This convergence time can be anywhere from 20 minutes up to several hours and this therefore makes PPP (on its own) less attractive in real time applications.

Figure 7 - Positional accuracy vs Baseline of different GNSS correction methods [9]

Conclusion

With the combined technology of both PPP and SBAS, the positional accuracy is projected to be around 5 cm for dual frequency (L1 and L5) single GNSS receivers and sub meter accuracy for low cost single frequency GNSS receivers, such as Ublox NEO-7P. Whilst positional systems providing 5 cm or greater positional accuracy has been available for the last 20 years through the use of RTK systems, the cost of such systems have always made it uneconomical for a number of applications, with the infrastructural cost of such system being in the tens of thousands of dollars. The end effect is that by using SBAS and PPP extremely accurate positioning can now be achieved at a fraction of both the product design and manufacturing costs. Furthermore, unlike local DGPS or RTK systems, the onus of the infrastructural cost and maintenance is no longer with the consumer or product user, but rather with government. For these reasons, this looks to be an extremely exciting technology which could have big implications in a number of industries such as Unmanned Aerial Vehicles (UAV), autonomous cars, precision agriculture and surveying.

Although the SBAS and PPP services are still in their trial phase and will likely end early 2019, we at Beta Solutions are excited about the future such technologies pose. These technologies will allow us to further enhance our GNSS product design solutions and will give our clients a leading edge in their respective industries, thus we will be following the developments of this trial closely. If you’d like to find out more about how SBAS or PPP enhanced GNSS could benefit your business, please do get in touch via our contact page .

References:

It’s here! Announcing the launch of our new website.

Wed, 14 Jun 2017 21:02:26 GMT

Author: Brenda Wormgoor, Marketing and Operational Manage r

Beta Solutions is excited to announce that our new and refreshed website, which has been many months in planning, is live. The updated site can be found at the same place: www.betasolutions.co.nz

The main objectives were focused on aesthetics, simplifying our content, and increasing the visibility of our full spectrum of services. The new design also allows for streamlined menus, clear navigation, and a responsive layout for all platforms. We trust the improved structure of our content will let you get more from a quick read. There’s a whole host of smaller but impactful changes, including new graphics, a gallery’s worth of images, and a few updates that have made the site easier to use.

We are especially proud of the:
Interactive 3D model, designed by our engineers,
Getting Started Form on the Contact page,
Tour of our facilities slideshow,
Video that shows you how we work,
New blog, and
Case studies our clients allowed us to showcase.

The new website has been a collaborative project and we would like to thank our staff for their work and input into the content and eTheory for the development.

We hope you like the changes and would love to hear any feedback you may have - email us on contact@betasolutions.co.nz

Test Driven Development

Sun, 28 May 2017 20:56:56 GMT

Author: Morten Kirs, Electronics Engineer

There are a multitude of coding standards and methods available for writing modern day firmware and software. One of the more popular approaches that has been floating around for us and in the wider coding community is Test Driven Development (TDD). While this is more commonly used in software development, various tools allow the process to be implemented in firmware as well. A recent project has allowed us to revisit this approach in more detail, where we've implemented a few chip drivers using TDD on the CppUTest platform.

What is Test Driven Development?

When first learning to write code, it is very common to write a number of lines, and eventually test the code to see if what you wrote works. This process often sticks with us, where we write more and more complex and lengthy code only to test once it has been written. TDD is a shift in this mentality, where tests always come first.

The basic process of TDD:

Refactoring is an important component of the process as well, and simple to do. If you see a simpler method of doing things, the code should likely be changed to use this method. As most of the code would be protected by a multitude of tests any refactoring errors should immediately pop up in test failures. The code will shape itself as you progress, often deviating from an original design to make things simpler or more efficient. By following the process, and refactoring as you go to keep things neat, the idea is that eventually your code will be complete, readable, and fully tested.

For us the test process starts with simple unit tests which slowly grow into more application specific system tests as the drivers themselves grow. Making sure that these tests fail before writing the code seems like a trivial thing at first. However, there have been minor instances where we've caught ourselves not initialising a variable properly, leading to a surprise test pass with no application code actually being written. Test failure in a way is a small test on its own.

By following the process we end up with some basic chip drivers that can easily be expanded on in the future. While we believe that refactoring is an important part of the design process, getting together a full plan of what the code will look like should minimise the need for this. The danger with this in TDD is that you are tempted to write an extra few lines of code even though a test currently does not cover what you have written. Coming from a process of "Test Eventually Development", ensuring this doesn't happen requires a fair bit of discipline.

Agile Development and Feature Creep

We live in a world of agile development. Often times we are given a loose specification and it is up to us to see what we come up with. This loose spec becomes more and more defined over time, where the client will often request additional features during the project's lifetime, even after deployment or the start of production. Getting defect free code out fast often becomes the norm. Attempting to fit additional features into a plan is not always simple. However, TDD is definitely suited for this. An additional feature simply becomes another test or set of tests that needs to pass. Even if code needs to be refactored in order to get the feature into the plan cleanly, you can be fairly sure the rest of the code is unaffected as long as no other tests have failed.

Debugging vs. TDD

From small $2 LED drivers to the $2 billion Mars Rover, almost all firmware and software will have hidden bugs. Whether these pop up during development or later on in the life-cycle of a project, they will often need to be fixed. Debugging and TDD have very different amounts of effort involved. Debugging often involves inserting break points, miscellaneous print debug statements, and temporary variables. It is often a lengthy process, involving looking at large sections of code, where most of the effort is eventually wasted. All of these features are eventually removed when the bug is found, and any new programmer investigating further issues will need to do so with a fresh set of eyes. In contrast, TDD will have code surrounded by a suite of tests which can be run on hardware or software. In firmware, often times the bug can simply be a hardware fault, and TDD can point this out quite quickly when the tests are run on the actual hardware. TDD tests are often focused and can pinpoint the problem so that only a small portion of the code needs to be examined.

The problem is that TDD tests won't always cover the system as a whole. In situations where interrupts can disrupt processes at any stage, it may be necessary to move to the standard debugging steps to find the cause of the problem. However, at the least TDD gives some confidence once the fix is in place that other functionality is unaffected. Tests can be added to cover any further edge cases found through debugging.

Conclusions

Test Driven Development can be a powerful method of programming in agile projects. It can save time in debugging and additional feature development due to the thorough testing that underlies the whole process. We will likely continue to use this method in the future for quite a few projects, especially in the case of reusable drivers which we can continually improve on.

References:

Test Driven Development for Embedded C (Pragmatic Programmers) by James W. Grenning
https://www.versionone.com/agile-101/agile-software-programming-best-practices/test-first-programming/
https://www.agilealliance.org/glossary/tdd/

Beta Solutions Move to Palmerston North City Centre

Thu, 30 Mar 2017 19:52:44 GMT

Author: Brenda Wormgoor, Marketing and Operational Manager

After eight years of continuous growth, Beta Solutions Limited recently moved its design team to bigger, brighter, custom refurbished premises in the heart of Palmerston North.

The building dates back to 1925 and we worked closely with the landlord to ensure we respect its heritage whilst merging it with our high-tech needs. There is an artistic eclectic balance that permeates throughout the open plan office - like the steel beams resting above glass walls, industrial steel trays that light up the apex, as well as treasured original timber, brick walls and concrete pillars.

The idea was to create an environment that enables us to best meet our clients’ needs and augment the Beta Solutions brand values. As complex problems are best solved collaboratively, it was important to create a space that will facilitate the innovative design work whilst providing a professional electronics and mechanical lab for development, prototyping and small scale manufacturing.

Work time is serious and sometimes you can hear a pin drop while everyone is dedicated to their projects, but come break times the team at BSL play card games, challenge each other at table tennis or takes the frisbee to the nearest open patch of grass… That is why we were mindful in creating various sub spaces within the open plan office to ensure we can work hard and play hard together.

By positioning the cafe style kitchen and lounge in the front, by the windows, we were able to maximise the views from the six fantastic windows and bring the greenery from the City Square and the bustle from the streets inside. We even have some historical photographs on the walls to establish a sense of place and celebrate the history of our city and the small part we get to contribute to its future. A five meter long window seat allows for a quiet space to think, contemplate, plan and design from. Or the kitchen bar proves popular if you need a coffee fix to enhance creative thoughts. Thus within the open floor plan, there are quiet places to concentrate or social places to collaborate in - an environment that fosters creative thinking and functional work.

Ceiling height acoustically rated glass walls in the meeting rooms allow one to enjoy the office space and City Square views whilst ensuring privacy when discussing new ideas or design projects in progress. The two meeting rooms are separated by a soundproof door for multiple meetings or by opening the door a versatile spacious meeting or training area is created. Perfect places for clients and employees to share ideas and work together towards creating successful products.

The workstations in the open plan office area are in a formation that facilitates discussions and problem solving. Each Engineer’s workstation is equipped with essential technology and tools they require to innovate, develop, design and deliver from. Wheeled shelving units with electronic and software equipment can easily be moved between desks and the electronics and mechanical lab.

And then there is the lab, with its high tech tools, workbenches and components stored and organised with engineering precision and efficiency. Our lab includes both our specialised electronics area - complete with anti-static mats, soldering equipment and diagnostics instrumentation and also a dedicated mechanical & materials section - which contains more essential tools & drills and our indispensable 3D printer.

A dedicated social area houses a table tennis table, dart board and golf putting mat. Designed as much to invigorate and build work relationships as it is to stimulate creative problem solving. Great ideas or innovative solutions often come to mind during either our proven design processes or competitive social contests.

All in all, it is proving to be a perfect place with various dedicated spaces for clients and employees to share ideas and work together.

You are welcome to drop by and see what we are up to, enjoy the views from our windows over a cup of coffee or challenge our reigning table tennis champion.

beta_solutions-v2

3 Habits To Communicate Better - In Slack, Teams and Email

Habit 1: Address Your Messages to a Single Person

Habit 2: Number multi-part questions

Habit 3: Respond Even If You Don’t Have the Answer

❌ Bad Example

✅ Good example

TLDR:

Found this post useful?

References

How to Effectively Run After-Action Reviews (AARs)

What is an AAR?

Why Does It Matter?

How To Conduct AARs

Timing

Led by the Project Manager

The Presentation

Team Participation and Open Dialogue

Common Challenges and How To Address Them

Key Takeaways for Implementing AARs in Your Team

﻿ References:

Modern Electronic Hardware Design: An overview

The Beta Solutions Hardware Design Flow

Phase 0 - Pre-engagement Evaluation

Phase 1 - Discovery & Specification

Phase 2 - Detailed Design & Testing

Schematic Design

PCB Placement & Routing

Prototyping & Testing

Overview of Phases 3, 4, and 5

Why Choose Beta Solutions

An Introduction to Firmware

What is Firmware?

How does Firmware relate to Hardware?

How is Firmware different from Software?

How is Firmware developed?

Want to learn more about Firmware?

What next?

One year developing Firmware with AI

Introduction

How has AI changed the way we develop firmware?

Where does AI struggle?

Looking Towards the Future of AI in Firmware Development ﻿

Conclusion

The Art of the Launch: How Real-Time Diagnostics Can Secure Product Success

How to Nail Your Products Release.

Example Scenario: Navigating customer issues the outdated way.

Introduction

The Value of Real-time Diagnostics:

What's Memfault?

Memfault's Backend

Example Scenario: Navigating Customer Issues With Real Time Diagnostic Tools

Conclusion

How Legendary All Blacks Utilise Real-time Player Tracking Technology

Decoding the Tech Behind Rugby World Cup Player Analytics

Overview

The Heart of Sports Tracker: Sensors

The Wearable Factor: Safety, Comfort, and Durability

Battery Life: Making Energy Efficiency a Priority

Live Data Retrieval: Ensuring Timely Feedback

What Metrics and How?

"One of the metrics we look at is high MET, and essentially, what we’re trying to understand is the acceleration/deceleration volume, or intensity." – Nic Gill, NZ All Blacks , Head of Strength and Conditioning [1].

Hitting the Finish Line

References

Advancing the future with today’s innovative Satellite solutions

Technologies for Satellite IoT

LEO Satellites

GEO Satellites

MEO Satellites

What are Micro Satellites?

Conclusion

References:

Unprecedented NZ Motorway Radar

How does an ITS Radar System work?

Conclusion

References:

Global Chip Shortage - How Beta Solutions has adapted

﻿ Introduction

How have we adapted to the challenges faced by the global chip shortage?

Design Process - Traditional approach pre-shortage

References:

Looking Towards the Future of AI in Firmware Development

Example Scenario:
Navigating customer issues the outdated way.

Example Scenario:
Navigating Customer Issues With Real Time Diagnostic Tools

"One of the metrics we look at is high MET, and essentially, what we’re trying to understand is the acceleration/deceleration volume, or intensity."
– Nic Gill, NZ All Blacks , Head of Strength and Conditioning [1].

Introduction

3. Hyper-accurate Positioning

5. Multi-skilled AI