The Faulty Sensor Phenomenon has brought many elegant connected systems crashing down. As designers, we put so much effort into a robust wireless solution, ultra-efficient data transmission, and a gorgeous, intuitive app that it can be incredibly frustrating when it all falls apart over an unreliable sensor.
While nothing is foolproof, there are a few ways to prolong the life of your IoT sensors that can be designed in from the ground up. These strategies center on reducing power requirements, mitigating environmental impacts, and designing a modular system.
Power
Given how many IoT sensors are battery powered, power management is typically a familiar design concern. While it is always wise to use the lowest possible power transmitter and processor, there are a few other ways to extend that lifetime even more. Some common sensors like Tire Pressure Monitoring Systems (TPMS) use a clever dual-sensor strategy where a low-accuracy and low-power sensors act as an interrupt to the main system that triggers a reading of a high-accuracy, higher-power sensor. TPMS tend to do this with temperature sensors and vibration sensors with a single register threshold sensor tied to a resistor or piezo element respectively, which then triggers a wakeup-and-read event for the microcontroller to read a proper differential temperature sensor or multi-axis accelerometer.
It is also worth evaluating how often you really need each sensor to transmit/receive and if you can extend the interval between transmissions by relying more heavily on event-triggered data. Even if you can ping a sensor every 90 seconds rather than every 60 because you know it will send data as needed, you can see significant power savings over the lifetime of the device.
Environment
It’s a cold, cruel world out there. Sometimes it’s a hot, cruel world with a lot of wind. Your sensors need to be ready for all of it. The smart home revolution may have brought on the first wave of connected sensor networks, but Industrial Internets of Things (IIoT) are popping up everywhere and are placing unprecedented demands on their connected components.
If temperature extremes are a concern, an especially rugged component might be all you need. We have common sensors like accelerometers that can function anywhere between -40oC and +125oC, which might even be a wider range than other components in your device, like the microcontroller or battery.
In that situation, it may be worth physically separating the highly-tolerant sensor from the more sensitive components with a cable. Wireless may be all the range, but a short run of wire within a wireless device can keep you from having to design to unnecessarily aggressive specifications or suffer unexpected failures. An electrolytic capacitor rated to 1000 hours at 25oC may only survive a fraction of that time at 85oC.
Extreme temperatures are difficult enough, but wide temperature swings over short periods of time can cause even more problems. Solder expands and contracts with temperature and becomes more brittle the more shrink/expand cycles it endures. We are used to using solder as an electrical and physical connection for everything but high-strain components on conventional boards, but glue or potting material should be a part of the BOM for any board subjected to repetitive physical or thermal stress.
If the device is not intended to be field-serviceable in any way, completely filling the housing of a device with potting material can do wonders for thermal and shock performance. Even if you intend to have the ability to replace certain components, a dot of glue under a component takes the majority of physical stress of the component’s leads while still letting that component be removed if necessary.
Modularity
If your main source of revenue is a software subscription and your hardware devices are inexpensive, disposable units – ignore this. Based on the previous paragraph, you should already be potting your devices into oblivion and will have no need for modularity. If your hardware is at a price point that the average consumer would consider more of an investment than an impulse purchase, there are two distinct advantages to designing an intentionally modular device – repairs and customization.
Whether you intend to allow users to repair their own devices or just make the refurbishment process easier on yourself, modularity enables a drained battery, blown capacitor, or cracked display to be replaced without undue hassle. If the device is intended to have a replaceable battery, this is a given, but even devices that are intended to spend years on a single coin cell can be salvaged rather than scrapped if individual parts can easily be swapped out.
