Battery level reporting is inconsistent. I have an alert set to let me know when batteries are below 50% since below that, I have seen devices become unresponsive. Over the past few days, my Iris V2 Smart Fob has been reporting battery life of 49%. However, when I checked this morning, it is reading 62%. I replaced this battery not that long ago (Energizer ECR2450BP with a best by date of 03-2029, so the batteries should be good.
I try not to replace batteries too often since that gets quite expensive. However, I also don't want the keyfobs to stop responding. That tends to impact negatively on the WAF when her keyfob stops responding and I leave to go somewhere and the HSM arms for away. Getting an ear full of siren does not make her happy.
Of all the areas in which there is still room for improvement in the HE platform, battery reporting is probably toward the top of my list for things I would like to see done. The Iris platform I had before Lowe's dropped it seemed to do a better job of battery state reporting. A battery should not report 49% and then 62%. I can see going from 62% to 49% over time, but gong from 49% to 62% seems to violate some laws of physics.
If the battery is converting voltage to a percentage, then 49% to 62% could be something as simple as a change in temperature. My garage door sensor bounces all over the place.
That doesn't solve your problem, but perhaps adds a bit of context.
It's up to the device, not HE. HE can only pass along what the device is telling it.
Edit: Battery level is a complex subject. The easiest/cheapest way is to just measure the voltage, and look up the % in a table (or use a formula). Using nice numbers, here's the difficulty:
A 10 V battery may be 100%=10 V, and 20%=8 V, after which it drops off quickly to 0%=5 V where it becomes unusable. Batteries exhibit something called depression. When you draw power from them, the voltage drops, and takes some time to recover. Based just on voltage, the % may drop from 100% when the device has been idle for a while, to 80% just after it has been used. 15 minutes later, it may again read 100%. So, the % will change depending on when it was last used, and it can go up. That's especially noticeable with devices which use coin cells, or draw larger amounts of power when being used.
The same device (keyfob) two different platforms (Iris, Hubitat). One was more consistent in battery life reporting (Iris), the other not so (Hubitat). What is the difference that made Iris more consistent? Do devices report a percentage or a voltage which is then interpreted before being reported? I can see the temperature difference affecting it, but when I am inside in a relatively stable temperature setting and the percentage fluctuates like this, what do I make of it?
I guess I could set the reporting percentage lower, but I don't want to go so low that the devices drop off the network.
I had the same thoughts about how it is the simply what the device reports, and I still believe that will be the case. The thi g that has me puzzled is how a swing of 13% could occur, mathematically I mean.....
I thought the discharge profile of lithium batteries was supposed to be pretty flat until they are nearly discharged. It is then that the voltage drops quickly. I had not seen anything about voltage depression which would be while it is being used if it is happening, In this case, the keyfob is not actively being used. The events list for it does not seem to indicate heavy use which would drain the battery or explain a voltage sag.
Maybe I will switch to one of the Iris V1 keyfobs for presence and suspend use of the Iris V2 Smart Fob to see if it behaves differently when it comes to battery life. I like having the four buttons of the Smart Fob. The V1 fob only has two, so I would lose the ability to have one of them to arm partial for night or full for away since I don't know for sure if I can program for pushed or held on the V1 fob. I'll know shortly though.
Not trying to be facetious but your post title is redundant. i.e. Battery report and inconsistency
Seriously I've never instrumented a battery device and tried to relate reported data to measured data. But I've seen seriously variable results from various battery devices. I had one that reported 0% but continued working. I've seen devices sit at 100% for a long time the dive quickly.
For all my battery devices I use the app "Device Activity Checker" by @bertabcd1234 . It works great for me. I don't have any FOBs so I can't say if this suggestion is useful for them.
Yes, which makes reporting accurate battery level based on voltage very difficult. A battery datasheet will show the discharge curve, and also voltage vs. current (or resistive load).
But of course, the battery is being actively used by the fob's circuitry when it's measuring the battery voltage. And to power a radio and communicate with the hub. So, depending on the distance to the next hop, it's power output, if it has to retry, etc., it may use more or less power.
In any case, HE just uses what the device reports. GIGO.
Battery reporting is a crapshoot at best with these devices. I have mine all set to report at 20%. I figure that's within the error margin and if something hits that just replace it no questions asked. Do not rely on it. Hell, I've seen negative battery numbers...
The only problem with an app to check device activity is that the keyfob is also for presence, not just for button presses. If it were just button presses, it would not be so much of an issue. However, presence does need to work reliably.
I use a smartthings v2 presence fob (no buttons) on our keychains and in my kid's back packs. They've worked well so far in combination with phone presence... (I use combination from HPM). I change out the battery about every 8 mos...
I also use device activity check to identify offline devices. It's a great app for determining if a device has "gone dark"
I know this isn't for everyone but we abandoned presence detection while still on SmartThings. Best thing we ever did. I installed keypads at the entry door and we use that to arm/away if the house is empty. The only false alarms we have is when I forget and open the patio door in the morning before disarming the system. LOL. WAF went up with the keypad.
OTOH, the WAF is high with the presence detectors here. Keypads are in place and functional, but 99+ % of the time, arming and disarming for away is by presence. Only partial/night is done via keyfob, dashboard, or occasionally a keypad.
My Iris V2 motion sensors start acting goofy at around 30%.
I still have a lot of the control contact sensor and glass break sensors that will set at 0 for a long time. I recently started a spreadsheet to see if the "0 time" is a consistent thing.
Okay. I'll change my alert setting to 30% and just plan on changing batteries at that point. The Iris V1 sensors seem to be more accurate in reporting than the V2 sensors when it comes to battery life. I have no V3 sensors, so I can't speak to battery level reporting from them.
The battery type (CR2450 vs. CR2 may have something to do with this as well. CR2 may just have a better capability for consistent voltage than the CR2450.
