An interview with Sashi Ono, principal embedded systems engineer for Pure Engineering, written by Julian Wise.
Since 2006, Pure Engineering has been at the cutting edge of the Internet of Things (IoT). From low-power, minimal-resource devices to those requiring high speeds and high data rates, we specialize in the extremes, where high-quality talent is typically lacking.
After cutting our teeth through government consulting, we expanded our expertise into the consumer space. Having been in the IoT industry for over a decade, we’ve seen the landscape evolve and noticed a host of common pitfalls to share with you today.
Common Mistakes
The best strategy I can recommend to device designers is this: From the outset, always keep the end product in mind. Unfortunately, most designers lack this foresight, so they settle on unstable code or optimize the product for the wrong market. For both of these issues, experience helps solve these issues.
Open-Source Code
With so much open-source code available on the Internet, I’ve seen product designers grab the most-available, low cost dev boards they can find and copy-paste code to create a prototype. This may work for a prototype, but runs into problems when going to production, such as:
- Unstable code
- Missing key features
- Lacking necessary security features
Taking shortcuts may at first seem cost effective, but if you are expecting to take this prototype to production, expect to spend a magnitude more time to develop production quality code.
Optimize for the Appropriate Market
For different customers, similar requirements sometimes result in vastly different products. For example, government and consumer markets differ across several areas:
Priorities
- The consumer market is driven by cost, availability of parts, ease of assembly, and the FCC approval bands.
- Government customers prioritize performance, so their devices typically have a higher cost (and experience fewer regulatory restrictions).
Chips
- Consumer products often use low-cost, mass-produced chips that run on the available frequency bands.
- A government product is more likely to use higher-end chips that are optimized for size and power, not cost.
Manufacturing Run Size
- Consumer products should be designed for mass-production and manufacturability. All its elements should aim at that end, including the components, code, and hardware.
- Government clients typically only purchase a handful of devices (not a massive manufacturing run) so a device designer should focus on reliability, not manufacturability.
Ultimately, a government device may have greater performance and be one hundred times more expensive than its consumer-facing cousin. Designing a device is not as simple as “making it work.” If a company intends to sell their device, optimizing it at the outset for the right market is key.
The Value of Experience
Inexperienced Designers often don’t know what they don’t know, typically uncovering potential problems by running headfirst into them. At that point, it’s often too late to make major design changes.
With experience, you know what to expect and can avoid common issues. Once you’ve designed a similar device, doing it a second time is faster and more likely to succeed.
The Changing IoT Landscape
At its core, an IoT product is a sensor, a power source, and information that is sent to a server. We’ve focused on these elements since even before IoT was called IoT.
Over the last decade, the IoT landscape has moved from a “wild west,” where anything goes, to a much more standardized environment, especially with respect to communication protocols and security.
Communication Protocols
IoT has only recently gained the luxury of modern chips and industry standards. Previously, interoperability was incredibly challenging, due to a lack of standards. As part of device designs, Pure Engineering was involved with creating protocols that enabled devices to communicate.
Now, with modern standards and protocols,manufacturers can build devices that seamlessly talk over the same network. Going forward, maintaining that interoperability is key. For IoT to keep growing, the market must not fragment.
Security
From our start in government work, Pure Engineering has always been concerned with encryption and security. Many of these processes are now mainstream. Still, IoT devices pose unique, heightened security risks, and some designers might not use best practices, a dangerous combination.
Pure Engineering’s Approach
We call ourselves “Pure Engineering” because our team is exclusively comprised of engineers. Rather than simply making a product, an engineer creates custom, optimized solutions to address a problem. As a team, we focus on core fundamentals:
- Solving problems using core principles
- Designing properly engineered solutions
- Applying engineering over trial-and-error solutions.
We specialize in the full range of skills for behind-the-scenes engineering. Our key skill sets are:
- Architecting
- System-diagramming
- Hardware design
- Choosing the right components
Each and every step is important. If you start with the wrong components, for example, you’ll always build a non-optimal solution. We select from a wide range of components (CPUs, microcontrollers, etc.)--everything from ultra-low power to 8-bit, 32-bit, and 64-bit. We don’t pick a component simply because we’re familiar with it. We pick the best component for the job.
Our team has built hundreds of boards, from high-speed boards with 10 Gigahertz PCB layouts to low-leakage, ultra-low noise boards. After many years of designing systems, we’re familiar with the constraints and can address the trade-offs of each important decision.
We don’t advertise; customers find us. Our government customers keep coming back because they know they can trust our output. Our consumer products are all by referral, a powerful compliment that speaks for itself.
An Example: Ping GPS
Midway through their design process, Ping GPS hired us to help bring their product to market. They’re always saying, “I wish we had found you sooner. If we had started working with Pure Engineering at day one, the process would have been much smoother.”
Ping knew they wanted to make a high-volume product. At the outset (before they contacted us), they didn’t fully understand what that entailed. Too late, they realized that their prototype product was not testable or mass-producible. When planned properly from the outset, this would have saved time and money.
When reviewing the requirements for a device, experience gives you the foresight to know what will be needed at the end. And designing with the end in mind will always produce a superior product.
Check out their Ioterra profile!
