I don't belong here - but this is actually being addressed in the hardware side - there are already 'smart' batteries and the tech for battery reporting is coming in the near future. Check out:
Smart Batteries
While those are pretty cool, how exactly would that address a device not reporting the correct battery %? It wouldn't as far as I can tell.
Also, none of those batteries linked are coin sized, or CR2, or CR123a which is what most home automation devices use.
2 Likes
This has been my experience basically 100% of the time since I started using sensors powered by lithium-metal batteries, so I long since gave up on trying to track battery levels and like others settled on a âlast activityâ type solution to get a notification when a device hasnât been heard from in a while.
Looks like those are rechargeable lithium-ion. I think we would all be very happy to get the type of reliable battery level reporting we have come to expect from our laptop and cell phone batteries. But itâs probably not gonna happen given the chemistry of the types of batteries that are used for the vast majority of IoT sensors.
1 Like
I found this discharge for some lithium battery, what is interesting about this chart is it includes low current draw conditions (most don't). Now I'm sure this is not the holy grail of battery discharge curves but it is something we can look at.
If %age were reported, depending on the specific battery 100% may be a range of volts and 0%f the same.
The device reporting would go from somewhere around 90 to 110% in the first 200 hours.
Then drop down to 60 to 40% for the next 20,000 to 30,000 hours.
Then the percent will drop to some low %age maybe even negative before the device gives up the ghost.
I had a device no longer respond. I measured the battery (with a good voltmeter) it was 0.55 volts.
So in my mind %age is impossible to relate to what might really happen. I think voltage, people understand.
I understand your POV - it was a an example of the direction battery hardware is taking. Obviously it doesn't solve old hardware but I think it demonstrates one method we may see in the future of the battery reporting itself. As smart battery tech continues to be driven by the home/auto industry a trickle down is happening that will eventually reach even coin sized batteries IMHO.
1 Like
How does a device, that is running off the Battery, reliably measure the battery? At some point there's not enough voltage to run the device, independent of what the remaining voltage of the battery. With a cliff describing the fall off at the end of life, by the time the device decides to give the battery a check, will there be enough battery left to spit out a message?
I see a LOT of my battery values in the 60% region as the final message they sent over the years.
1 Like
This is part of how Visonic did it with their proprietary radios. Maybe there is a measurable step down in RF quality right before battery death. Granted this could be confused by someone putting a metal watering can next to a sensor 
OK, there's been enough posted here to appreciate the scope of the problems/challenges ...and so do we agree that we'd at least benefit from an HE integrated app that allows you to select all, or some, devices to be alerted to the fact that either:
a) you have a device that has recently had "spotty" (weak RF, incomplete/broken packets,...) reporting compared to the recent past,
and/or
b) you no longer have a device reporting that usually reports within X hours, or X days
--------------10-9-21 Addendum that is Topical ----------------------
Just got a Visonic MP-840 and found the following notes of interest. I read them as... "the kind of things a long time Security System device vendor states (and puts into their ZigBee product) when device reliability (and battery health monitoring in particular) is understood as imperative; to them, to you, and to UL installation certification.
I'm just wondering after all the discussions about "the challenges"; how it's working for Visonic, what standards it meets or breaks in ZigBee or otherwise, and what Hub listens to everything this device can say.
Actually, re-reading it again it's more about the smarts in their device and everything downstream is just that smarter device reporting to the Hub what it knows, more accurately than most it seems, about its' battery state.
We collectively have so many concerns with battery life that maybe we should be holding more vendors to task on making their devices better, even if it costs us a little more.
ELECTRICAL
Internal Battery 3V Lithium battery, type CR-123A. For UL installations, use Panasonic and GP only
Nominal Battery Capacity 1450 mAh
Battery Life (with LED on) 5 years of typical use
Notes:
Inability to connect with wireless network, or wireless link quality no higher than 20% may
significantly reduce the expected battery life. (Intuitively we all know it, but I haven't seen many HA device vendors bring that to your attention.)
Low Voltage Signal A fast low voltage signal is generated at or below 2.9V
Battery Power Test
Performed immediately upon battery insertion and periodically after every several hours.
The power supply is type C in accordance with EN 50131-6 Documentation â Clause 6
Source:
Hard to say if itâs overall beneficial. This exists as community-developed code by a couple people already, which Iâm personally satisfied with.
For Hubitat to take it on, theyâd have to weigh it with other priorities, take on the maintenance and support implications of its limitations causing users to become confused or make inaccurate conclusions about the status of some devices, etc.
Edit: SmartThings had a âheartbeat,â which didnât work very well in my experience, and Iâd be surprised if it does now. Lucky for me, I moved here and donât have to find out 
.
4 Likes
You can suggest there is agreement about this stuff all you want, but I can tell you that this is not something that we have any intention of adding to the hub as a built-in capability. There are several reasons for this, not the least of which is the problem of mistaken expectations (you might have realistic expectations personally, but that does not mean that everyone else will have).
We have provided the means by which this can be done, as @marktheknife has described.
3 Likes
No, I do not agree. There are at least two community apps that provide this function. They work great. I see no need for the Hubitat folks to divert resources to recreate the "wheel"
3 Likes
Thanks to all who picked the baton and ran with this topic. I understand attempting to handle all devices with a generic handler is a big problem. However, I'm frustrated with the HE staffs' refusal to do even the smallest of things that would allow users to handle battery voltage as we see fit with rule machine or a groovy app.
My current and only request.
1. When a device provides battery voltage, report it in tenths of volts as attribute: volts. No device voltage reporting = no volts attribute Leave everything else as is.
With the quality and experience of the staff at HE this should be simple and easy to code. Should the HE staff prefer, and they trust a non HE employee to do the changes, I will sign an NDA and do this free of charge.
BTW. My Sengled contact sensor reported as 0% battery at the top of this thread in March of 2021, still operates at 0% without a battery change.
5 Likes
I'm sorry that you don't understand our decision making process, or agree with what we've decided. But, it is our decision making process. Thanks for your input.
3 Likes
I totally understand NO.
1 Like
This topic has run its course.
