3 ms·
For actual historical temperature data, see here. [1] It looks like a general range of 5-9 C is the norm, with 12, 15, and 30 C (53, 59, 86 F) spikes over the l
by ColanR 7y ago
For actual historical temperature data, see here. [1] It looks like a general range of 5-9 C is the norm, with 12, 15, and 30 C (53, 59, 86 F) spikes over the last 2 months of data. For that kind of failure rate, I wouldn't depend on it without some failsafe warnings for when the temp had dropped below safe levels.
Neat project though, and I might build one if the reliability improves.
[1] http://house.joeyh.name/historical.html http://house.joeyh.name/historical.html
- joeyh 7y agoAny sharp spike on that graph is a bad sensor read; the 1wire protocol has a pretty weak checksum and you can get bad data that crcs ok when you're reading lots of data. That covers the 30C, 15C. The spikes to 12C or so that then slowly drop back down are generally the fridge being filled with groceries and taking time to cool back down. If you measured your conventional fridge you might see similar events. I did have the arm board hang twice this summer. Once I was on vacation, luckily it froze with the fridge turned off (so it didn't freeze down to -5C which is where the dumb thermostat is set), and the fridge was fine for the 24 hours or so until I got home, creeping up to maybe 10C. I need to find a more reliable embedded computer.