I wonder if there's a way to see the voltage from a Generic Zigbee Motion Sensor (which happens to be a Samsung F-IRM-US-2)? I'm asking because I put fresh CR2477's (showing 3.16 - 3.2VDC on my Fluke, I've tried a few) into some units and I rarely get 100% battery. I get 40-62% instead.
I've done Amazon batteries but I don't trust them so then I tried Digikey (direct, not the marketplace) for some Jauch. Those weren't reading 100% so in my further searching I found Mouser has some new "wide temp range" batteries that seem interesting, https://www.mouser.com/ProductDetail/81-CR2477X ... my point here is, "I'm pretty sure the batteries aren't DOA or defective."
The motion sensor works fine at low percentages for weeks/months, but this throws a big monkey wrench in my ability to automatically scan for devices with low battery and alert me.
Battery reporting is.... almost not worth it, it feels like sometimes... Ther are many situations where we see devices report somewhat accurate rwadings, when others report highly inaccurate results.
While I wouldn't discourage you entirely from looking at battery readings, I wouldn't suggest using these as your basis for assessing the health of the battery. rather, looking at the activity for the device, ideally for devices that have some regular reporting metric like temperature. When these devices do not report within a relatively short period (less than a day), then something is wrong.
For those devices that only report sporadic events.... I'm not so sure how best to deal with these myself....
I'm in the "battery values from devices are almost worthless" camp.
I've had battery devices that never, ever, show anything but 100%, for well over a year, until one day they just stop working (and new batteries resolve that). I've had other devices that show 30-90% on day one, work for well over a year, and just die (new batteries again resolves that). I've had devices that seem to track fairly accurately down to 1%, and still others that do so down to about 30%, at which point they die. There's almost no rhyme or reason to it. Devices from the same mfr and same lot often behave wildly different.
My strategy? Change stuff out once per year, usually in Oct or Nov because we live in Maine and batteries don't like cold, and I don't like being cold when I have to change them. Do as you wish, but after 20 years of home automation, I no longer even try, with battery readings.
Yeah I think I’m going to have to join the “proactively every year” camp. I’m trying to manage various locations where I might not be present for many months so I have to predict what the battery will be like.
I used to try to run a spare motion sensor in my garage deep freezer (actually the same model sensor that’s giving me trouble, but a different actual unit). The battery would last maybe 2 weeks so I came up with another solution for freezer monitoring, but I also kinda want to try again with these new batteries!
I sampled the 9 unused CR2477X's that I have, they average 3.171 V (3.158 - 3.194)
The battery in use is showing 2.951 V (and reporting 12% now, lol)
The battery was put into service 2024-04-08 22:17 and measured just now at 2024-04-15 13:31, so it was in service for about 159 hours.
I use Home Assistant and love the built in logs, so here's a snapshot of this device over time (I hovered over the moment I installed the battery on the upper left for timestamp reasons)
