Imagine a central alarm control system, with tiny wireless alarm sensors sprinkled throughout a home or building reporting open doors and windows, glass vibration, smoke, CO2 , etc. Or an HVAC monitoring system, with the sensors sending temperature, humidity, airflow information, with a central processing unit utilizing the information to optimize energy usage.
And these are just mainstream applications- how about spooky spy applications like buried perimeter sensors monitoring vibrations caused by trespassers? Or sensors embedded in roadways detecting traffic, noise, or embedded in bridges reporting structural stress?
A large class of the sensors in these systems are “quiet” much of the time, measuring slow processes on a very low duty cycle or exceptional (interrupt-driven) basis.
These sensors can operate on extremely low power, utilizing the very low power sleep states available in recently available microcontrollers, awakening and consuming short bursts of power only when a measurement is made and processed, or when the radio is sending data.
The microcontrollers can wake in under 20µs, turn on embedded analog and the ADC, make a measurement, provide some processing, and return to sleep mode in a matter of a few hundred microseconds. The averaged power consumption of such units can be less than 1uA (Figure 1).
Figure 1: Sensor load profile for an example sensor making a measurement every 2 seconds, and transmitting the data via Zigbee radio every 10 minutes, on average. The radio may make multiple retransmissions if needed with a minor effect on total average current consumption.
These systems have in common the need for unobtrusive, generally hidden, often inaccessible sensors. And, the sensors are usually tucked away in places where there’s no power- so how to power them? Lately, it’s been fashionable to consider harvested energy.
Tiny solar cells (the size of those in calculators), piezoelectric vibration elements, or thermoelectric generators (TEGs) can sometimes generate enough power, but harnessing this power naturally restricts the choice of location of the sensors. Piezo sources require sustained vibrations from machinery, or (with a little creativity) airflow; TEGs work well with cold or hot pipes, or when attached to some heat generating machinery, for the needed temperature differential.
Solar cells can readily harvest indoor ambient light (the cheap calculator panel I measured generates 2V at 2µA), but what if someone switches off the light? If relying on sunlight, overnight energy storage is required- typically a supercap or small rechargeable cell; but the leakage of these capacitors and batteries may far exceed the current drain of the micropower systems it is supposed to be powering, forcing a larger solar panel and demanding a large value supercap (sustaining a 16 hours overnight charge with 10uA leakage requires a cap in the 100mF range!).
As well, both rechargeable batteries and supercaps contain electrolyte which typically has a shorter service life than a primary battery, defeating harvested power’s promise of unlimited lifetime power.
To read more, go to “A better solution.”