How to think about your designs if you want to dump batteries

In the last few years, I’ve seen several startups advance their efforts to make well-functioning energy-harvesting chips for real-world applications. And this week, at an event, I watched an employee of chip maker Renesas share how it’s bringing its latest energy-harvesting line of sensors to the mass market. So while energy harvesting is not mainstream yet, I believe we’re just three to five years away from getting rid of batteries in small-scale sensors and radios.

But first, product designers need to change how they think. Or chip designers need to get into the product game and show the product designers how it’s done.

Graeme Clark, product marketing manager at Renesas, hopes to help product companies rethink their designs so they can use energy generated by light, temperature changes, or motion. At a recent presentation on Renesas’ energy-harvesting sensor products, he used the example of a small, 25-centimeter solar cell that could provide about 50 microamps of power, which could then be used by a device that needs 150 microwatts. (For those of us who aren’t electrical engineers, I find it helps to think of amps as the amount of electricity flowing past a device, and watts as a measure of what that electricity can actually do).

In Clark’s example, he noted that with that amount of power you could send a message using Bluetooth every 15 seconds or one using NB-IoT every eight hours or so. But any device running on that solar cell also has to spend power sensing, processing, and then communicating. So designers should figure out how often they want to use power to, for example, take a sensor reading. In the case of outdoor temperature, which doesn’t vary often, taking a reading every 15 minutes is likely fine. But if you’re following a precise industrial process you might want to take a temperature reading every 15 seconds.

And yes, some sensors will require more power than others.

In the case of processing, you might have enough power to perform a simple IF/THEN calculation; for example, if the temperature rises above X, send an alert. Here again, however, designers have to think through how that could work. If a device is likely to send alerts multiple times a day, for example, you might want it to have a more power-efficient radio technology that talks to a powered gateway. This is why many battery-powered IoT sensors in the home use Bluetooth to talk to a gateway or a phone.

A designer has a bit more leeway with a battery capable of storing energy, which is why most of the devices around today come with a variety of batteries. Indeed, most devices that might use energy-harvesting technology in some capacity still have batteries. But Clark sees things changing in the next three to five years as governments make laws that incentivize companies to reduce their reliance on batteries. Not only is it a lot of work and expensive to have someone change the batteries on thousands of connected sensors, but batteries are full of harmful chemicals, which are often mined from parts of the world where most consumers would not want their beloved brands spending money.

Clark expects that the industrial world, especially in an area like asset tracking, will start embracing batteryless sensors and continue pushing the technology to advance. Renesas invests in energy-harvesting technology because it provides both microcontrollers and sensors to companies that sell their products as part of the internet of things.

But Everactive, a startup that has raised roughly $63 million, has taken a different tack. It sells anomaly detection for machines and industrial steam traps that are based on new, batteryless semiconductors, but it focuses on selling the entire solution, not just a chip. And so far, focusing on the whole solution as opposed to a single component has worked out well for the company.

This week, I watched a presentation featuring Everactive’s co-CTO, Ben Calhoun, and U.S. Air Force CTO Frank Konieczny. The Air Force is testing the Everactive steam trap monitoring service, and Konieczny praised the fact that maintenance personnel can easily pop an Everactive device on a machine and not have to worry about keeping the sensor powered. It can get the information it needs without having to either send personnel to monitor it or have someone manage a battery-powered monitoring system.

By selling the monitoring as a service, Everactive can avoid being pigeonholed as a commodity chip provider and build a more power-efficient system tailored to eke out every amp provided by the heat of a steam trap or the motion of a motor. Physical changes in temperature, solar, and motion aren’t the only options for harvesting energy.

Chip design firm Arm is building an entire project aimed at scavenging energy from radio frequencies to power RFID tags with computing power so they can not only communicate but also process a few bits of information — and even store it in memory in case the power supply gives out for a bit. Arm calls the effort Project Triffid, and it’s one of the more exciting things I’ve encountered in this glum year. In a blog post announcing the project, Arm says it wants to build a batteryless device that can work on intermittent power, run software, and maintain its work all while running on about 1 microwatt.

There are half a dozen other startups— such as Ambiq Micro, Atmosic, Wiliot, and more — trying to make batteryless devices capable, resilient, and cheap enough to enable a class of sensing devices that we can embed into our infrastructure, place in factories, and spread around our work environments to help us optimize businesses and improve our quality of life. We’re getting there.