Built to last or built like crap?
Ulf Seijmer
IoT optimist ·?CINO Induo · CTO/co-founder AKKR8 · EUTECH Technology Council | I help companies grow through IoT - ???+46-76-5755751
In the last couple of weeks, I have been involved in many discussions on IoT hardware design. As the IoT market is maturing, and more and more hardware becomes available, it is obvious that all hardware is not built to equal standards.?
Many discussions have involved batteries and battery life for IoT devices, more on that below. One of our partners described battery life as a time that the battery would last, but at that point, the startup that originally designed the product would no longer be around, in other words, if battery life would end up as 5 years instead of 10, there would be no company to blame for it. The other takeaway from that discussion was, that the longer the battery life, the more security measures were circumvented to build a device focused on just a good battery performance.?
The pitfalls…
There are many IoT-hardware suppliers out there. I am representing two of them. One distributing hardware, the other one designing hardware. I will share some pitfalls, so you do not have to make mistakes, no matter if you work with me or someone else.?
1. Skipping CE-testing
I can not stress the importance of CE certification. It is needed to get the stamp of approval for your product's performance from radio and electrical perspective. Of course, it costs money. And yes it is costly. And apart from that, you can not just send your product to the lab. You may need to invest in completely new firmware that is just there for the tests.
I know companies supplying hardware based on a design they purchased via social media. Hardware that has not been CE-tested because of neglect, bad judgment, or whatever it can be. Hardware that could be based on a design that is a copy of someone else's design. Hardware that the supplier probably has no clue if it is performing as expected or if it is dangerous, if it is polluting the spectrum or if it is fit for the environment that it is used in.?
领英推荐
How do companies get away with this? Well, if you purchase a sensor value as a service from a company, you will get a ready-made kit, right? But in these kinds of setups, we tend to worry about CE-testing even less than we normally do.?
2. Battery-life
I recently published an article and a video on batteries and the use of primary batteries, or disposable batteries. I feel that anything that you throw away after using it once is not resonating intending to spare the environment from unnecessary waste. I thought the worst thing anyone could imagine when designing a way to power supply hardware was to think disposable batteries would be the thing. But there are worse ideas. One is to put batteries on the board, which is disposable and soldered to the board, so you have to desolder them to change them. Why save €1 on soldering it in place instead of having a holder for the battery so you can remove it?
Are disposable batteries not useful at all? There is one situation where these outperform rechargeables, it is when you have to use them for many years as they will last for years without self-draining, this is where the rechargeables fall short. Also, and I am talking from experience, the design and safety testing for a product with a built-in charger is quite a large bite in your wallet. Believe me, I have done it. At the same time, I think the environment will last longer if we do not use disposable batteries in the majority of IoT devices, which seems to be the norm today. Dare to challenge your hardware supplier
3. Safety testing
As I just mentioned, testing for electrical safety on a product, something you normally do during the CE-testing is costly, complex, and hard. But it is part of the game. Designing hardware, or being the purchaser of hardware you have to understand that this has to be in place. You should be able to show the correct documents upon request.?
4. Firmware
So, when it comes to firmware in these IoT devices there are all sorts of very crappy code out there. This is super hard to prove or to understand how it is structured, I get it. There are many corners to cut. One is battery life. Battery life will be better if you compromise on security. Skipping end-to-end encrypted transmission increases battery life. So the worse device from a security perspective, the better the battery life.?
I am not presenting you with a solution to these problems. Massive IoT is great and interesting. But as we move ahead, let us jointly understand that a price-competitive product doesn't necessarily have the most expensive firmware or hardware. And even if the software in the device is okay, it could be hampered by processing power and storage space, reducing its capabilities to have a robust and secure ecosystem for firmware.
Please be advised. Ask your supplier questions, and don′t make too fast decisions. Know the pitfalls, and enjoy great IoT hardware.?