TLDR -> 24uW @6uA during sleep

I only received my REV01 PCBA on Thursday, and I was busy Friday night debugging and cleaning up documentation for the Green Powered Challenge.

Now that the bugs are squashed, and documentation posted (in its admittedly unfinished state) let’s have a look at power consumption. After some flailing, I came to a setup for measuring uA’s reliably that played well with the WellWatch.

Power is provided by a supply w/ sense wires connected to a 2 ohm current measurement shunt. The sense wires are key to stop the device from resetting itself due to voltage dips during solenoid switching (this is just an issue for my hacky setup where I don’t have a nearby battery to sponge the bursts of current). One DMM measures the potential across that shunt, and another DMM measures the potential across the 1000uF capacitor (the battery stand-in).

When I first tested the system, it’s sleep current was 98uA (375uW). Not terrible, but not nearly as good as I expected or wanted. Luckily the culprit was fairly easy. The floating UART pin PB3 was hogging power. Driving it low (when not in debug mode) resolved my issue. I tried tuning a couple other parameters, but only really saw substantial benefits from this change.

Fixing the UART issue reduced the current all the way down to 6uA (24uw). The gaps in the chart is where the device wakes up, and exceeds the bounds of this chart.

At this point I was concerned about the noise floor of my setup. In the below screenshot I tested the same setup with the PSU unpowered. This makes me more confident about the 34401A’s mV measurement. We should be able to reliably read above 1uA, though there might be +/- 1uA of slop on that value.