Interesting. The device itself still reports 100%… I kind of don’t expect it to move until it’s final month-ish of usage. I’ll stick it on the multimeter to see what voltage it’s putting out when I get a chance.
I haven’t looked at the “battery percentage” circuit much but it looks like a large voltage divider with a FET controlled by software. There is no true fuel gauge so the developers more than likely picked points along the coin cell’s voltage curve to represent a rough percentage. If that’s the case, the percentage won’t translate very well for other batteries.
I suspect you are right, you might not see it move much until the last hurrah. That or you will see it jump from something whacky like 100% to 10%.
With the runtime and an updated battery voltage (and the cell make\model) it might be possible to calculate your actual expected runtime. A multimeter averaged current measurement from a controlled source over an hour or so would be more likely to provide a good estimate of runtime. I’ll try to measure one on a DC supply in the next few days.
Reporting back. I measure an average 0.15mA DC at 3.0V utilized by the sensor. This allows for some reasonably accurate estimates of runtime.
The zigbee microcontroller in the sensor stops functioning at 2.1v (1.05v per cell).
If we take the Energizer Ultimate Lithium AA (Datasheet) we can see that the cell will provide roughly 3500mAh at 1mA. Since we are less than 1mA, it may provide more runtime, but there are no curves to support an estimate at less than 1mA so we will go with the worse case of 3500mAh. This calculates out to 972 days of runtime not including the small amount of self discharge and other likely losses. If you look at the datasheet, you can see that the lithium curve stays well above 1.05V until the very end. Pretty decent performance and energy use.
The way manufacturers publish alkaline curves makes it a bit more difficult to get realistic curves at sub 5mA levels. If you look at a typical Alkaline discharge curve at 1mA or under, you will find that the point at which the battery reaches 1.05V is much earlier than what is in the battery. Even with the unused capacity and higher self discharge, you might very well get a lengthy run-time and probably at a better value per day.
Now I know I’ve arrived late to the party but am I missing something here … I thought the whole idea of the presence sensor is portability on a key fob… so whilst the battery upgrade masks the problem don’t we really need samsung to sort this problem out… I have now contacted support twice and each time they are fobbing me off. My thought is not only the poor choice of battery but the way they poll the device too often… that said often when I measure the batteries I take out they are at 2.8 or 2.9 v so I question the programming to read the levels too. I understand you cant measure current draw capability that easily but given the direct relationship of v to I I think the current draw is also at a good level.
So my thought is to ditch the samsung device and find another… does anyone know any good z wave presence sensors ???
Depends how you want to use it. If you want it in your car, this seems like a good mod.
Obviously there’s a trade off between activity and battery life. But if the device sent out its “I’m here” signal less frequently, then it would be less useful. If I want to unlock my door when my presence sensor arrives home, I don’t want to just stand in front of the door waiting for the next presence sensor ping to the hub.
There don’t seem to be many. This one was announced recently but no one knows when it will be available.
Here is my version
I mounted everything into a 4 battery holder (after taking out 1 divider)
Saves having to buy an additional project box
Does the presence sensor report reasonably accurate battery levels after you’ve done the mods? I suspect it uses voltage, so might not be right?
It works great. Battery is very stable. I use Li-Ion batteries from my post above. They are still 100%.
I modified two. One for my car and one for my dad’s car. For him the garage opens and for my car, the driveway lights come one.
Well… Hopefully it is not reporting 100% when they finally run out!
Nah. I tried that angle. I also put in older batteries prior to the new once and they reported about 75%ish.
Ok ok ok. So I think I’ll go this route.
I’m also planning on using the MQTT bridge and send/receive events on my raspberry Pi using node-red. WIth that I can scan for BTLE ids, and do the same. So, would just need to associate phone BTLE IDs, and/or use BTLE fobs that exist (and are better with battery life).
Now I just feel silly, I think this is the real answer.
Agrred. I didn’t have a dramal tool so I used an outside project box for hardening the project from my kid opening it. Not yet screw driver trained…Thank god.
Where did you get the battery holder?
Mine were bought on ebay in the UK for a stupidly small price:
But I’m sure you can find them elsewhere
I discarded the switch but in hindsight it might have been useful to be able to switch if off
BTW. @lmosenko I didn’t need a Dremel to remove the divider… Just cut the sides and snapped it off.
(And they have a screw to stop Kiddies opening it)
If you’re in the US & have a Radio Shack nearby, they’re on sale for $1.56 right now.
Yeh. RadioShack close all the store in NYC.
I just did this same thing, what battery life are you getting?
I just ordered five of the 4-AA battery holders for nine bucks.
I was messing with using the batteries from the Iris contact sensors, but Cobra’s implementation is exactly what I’ve been seeking.
I haven’t touched the batteries since I did this mod.
All three cars are showing 87% battery