A Guide to Architecting Batteryless Devices Part 2: Energy Storage

Harvesting energy is only part of the equation for a batteryless device. The energy available from light, heat, motion, RF, or other sources can fluctuate throughout the day or disappear altogether. A sensor still needs to measure, process, and communicate during those periods. 

That makes energy storage a critical part of the system. In Part 2 of our three-part series on batteryless devices, we’ll focus on the role energy storage plays in keeping the system operating reliably when harvested energy doesn’t align with demand. This includes covering wireless transmission power bursts and low-harvest conditions while accounting for factors like self-discharge and temperature extremes. 

In this series, “batteryless” means operating without a replaceable primary battery. However, some designs may use a rechargeable cell or capacitor as an energy buffer. 

Best practices: 

• Size for off-hours. The governing constraint is applications that have off-hours, like a factory that shuts down on weekends. When energy is harvested from equipment or infrastructure when it’s in use, these off-hour times require the system to consume everything from storage. Model the longest off-hour duration, essential functionality for consumption, and minimum rate of charge it needs to support leading up to the off-hours event. 

• Account for self-discharge. For example, supercapacitors lose more stored energy per day with no load than chemistries like lithium. For low-harvest sensors that have low transmit duty cycles, this leakage is impactful to the total energy used for the sensor. Determine the allowable leakage for the system, then choose a storage type and size that provides enough capacity without exceeding this leakage limit. 

• Match chemistry to your operating environment. If the sensor is exposed to temperature extremes, such as in an outdoor parking structure, a rooftop junction box, or an industrial warehouse environment, storage chemistry has a big impact. Equivalent series resistance, self-discharge, and capacity derating vary significantly between chemistries as temperatures fall below -10 °C or rise above 40 °C. 

Common tradeoffs: 

• Supercapacitor vs. LTO. Supercapacitors offer greater cycle life and simpler charging from a fully discharged state. Lithium titanate oxide cells offer higher energy density, lower self-discharge, and a flatter discharge curve. 

• Capacity vs. size. More storage means more resilience to an extended harvesting outage, but more capacity adds size. The answer depends on your worst-case scenario: a sensor needing a 72-hour autonomy period is very different than one that only needs to function when there is energy to harvest. 

Designing a tiny wearable that harvests energy infrequently, yet needs to absorb it rapidly, required careful trade-offs. The high-power harvesting requirement pointed toward a supercapacitor, while the strict size limits favored a compact lithium battery. To resolve the conflict, we took a system-level view (as outlined in Part 1). By pairing the device with an ultra-low-power wireless technology, we drastically reduced the total energy needed per harvest. This smaller energy target fit cleanly within a compact supercapacitor, allowing us to keep the physical size small without sacrificing fast charging performance.

• Usable capacity vs. nominal capacity. LTO and Li-Ion cells discharge down to voltages below the minimum operating voltage of a typical embedded system. The energy between those two voltages is either unusable or requires a more complex power supply circuit, which adds inefficiency. This means the nominal capacity rating for these cells is not equal to the actual usable capacity. This difference is even more with supercapacitors since voltage decays linearly to the capacity used, and all the way down to zero voltage.

Choosing the right energy storage approach requires balancing capacity, size, self-discharge, operating environment, and usable energy. More capacity isn’t automatically better, and nominal specifications don’t always reflect how much energy the sensor can actually use. 

Ultimately, storage should be sized and selected as part of the larger energy model, one that accounts for how energy is harvested and how the device consumes it. In Part 3, we’ll complete the picture by looking at another important part of that energy budget: wireless communication. 

Learn more about how batteryless sensing can help you monitor more assets with less maintenance.