This is the second time the XIAO board stopped working for an extended period of time. Rebooting the board, unplugging/plugging the power cable and adapter from the wall, and even pulling the board from the headers, didn’t work to get it back online. I finally pulled the board, plugged it into USB, then made a do-nothing change to the main.py file1 and plugged it back into the headers, which worked perfectly. There were no obvious errors in the board – it just … needed a boring update to the code?
Anyhow, I’d discussed some of the pros and cons of running three Wemos D1 Mini powered sensors2 versus just one XIAO ESP32C3 powered sensors – but didn’t think about what happens when a single one goes down. Although the XIAO was out of commission, the three Wemos board kept going so I didn’t really lose any temperature data – especially since the data was so flat during from 12:30 AM to 9 AM this morning.
Admittedly, the locations for the three sensors aren’t perfectly matched. I have the three XIAO thermistors taped to the entryway flooring, taped to the metal plate just inside the family room, and one dangling in the air above the entryway out of any direct sunlight. The three Wemos AHT10’s are on the entryway floor near the entrance to the family room, well inside the family room on the mantleplace, and the last at the top of the stairs with the sensor in open air at almost the exact same height and area as the original XIAO ambient sensor. I actually wanted the second XIAO sensor over on the mantleplace as well – but didn’t want a run wires across the house that far to do it. I’m at peace with getting different readings from these sensors versus my historical readings because:
- The AHT10’s are supposed to be more accurate – and also measure humidity which I’ll use near my 3D printer later on
- There’s no really good way to tape the AHT10’s to a surface and that doesn’t really tell me much about the “feeling” of warmth inside the house anyhow
- The floor readings were constrained by the amount of wire I had rather than really wanting measurements at those specific locations
Just looking at the charts of the incoming data suggests that the AHT10 sensor data is definitely less noisy than the thermistor data. Then again, it might be due to the way in which I’m taking readings every minute but then broadcasting the data every 5 minutes. I’m sure no one would notice or care if I sampled the data every 5 minutes, but when the peak heating coming in from that one picture window only lasts about 30 minutes, I think I’d like to get a better sense of what’s going on minute-by-minute rather than at 5 minute intervals.
Summer Science - Testing Effects of Window Coverings on Interior Temperatures