Chestnut roasting (Question about charging)


#1

Has anyone been able to completely charge a Librem5 while it is running? Mine always overheats, the LED starts blinking red and the charging stops. Switching it off, letting it cool down and charging while off seems to work. When the red LED goes off (meaning charging complete), one can boot the Librem5. However I never managed to charge beyond 80%. Is the charging supposed to stop at 80% for battery health?


#2

Kyle posted yesterday in the Librem 5 Matrix channel:

a kernel update is coming in another day or so that fixes reboot issues, the long-press power on issue, and has some power managment improvements

Sebastian (developer) also recommended:

plug the charger in while the green led is shining during boot
this should let it charge and be powered on indefinitely


#3

I can confirm that the kernel linux-5.3.0-librem5-h1_5.3.0+librem5.8 works to fix the power button and reboot.


#4

The charge controller on the L5 expects a bigger battery, so with the 2000mAh it will never reach 100%.


#5

Just got OTA update which I think includes this fix (did not check actual kernel yet) but power button now works after couple of seconds.


#6

If this is the case, shouldn’t it report 57% (2000/3500) or so instead of 80%?


#7

You should get in touch with that other guy whose phone freezes, maybe they’ll cancel each other out if taped together and put in a box?

Or would it be more like Schrodinger’s cat? (Would they land frozen or roasted? … ad infinitum.)


#8

In my minimal EE experience (more exposure, really, through a couple college classes) mAh isn’t capacity the way 1 liter is capacity. I can’t remember what exactly it’s a rating of, though.


#9

Mainly because “mAh” (or As as it is correctly normalized: one Coulomb [Q]) doesn’t describe capacity at all, but electrical charge :slight_smile: Capacity is meassured in mAh/V or correctly normalized in As/V which is 1 F (Farad). :stuck_out_tongue:

The topic is marvellously complicated. It’s really hardly comparable to water in a glass.

Two boring examples of how batteries are wonky things. Click me if you like to be bored.

For instance: the amount of energy you can draw from a battery doesn’t rise linearly with the capacity it can hold. It rises far slower. In other words: Double the capacity doesn’t mean that you get to use your device twice as long. But you still need to put double the energy in. Yet in other words: it’d be a glass of 2 l of water, which does hold two times the amount of a 1 l glass, but when you upset the glass, you dont get 2 l out.
Or: The higher the current is which you want to draw, the lower the overall capacity of the battery is. Meaning: you can get more out of a battery by discharging it with a small current (say a TV-remote) than by discharging it with a high current (say a battery powered handwarmer). In the water analogy it would be a jug of water which would fill a glass of water if poured really fast, but would make the same glass overflow if it’s poured slowly.

Boring explaination of how the librem 5 is charged. Click me if you like to be bored.

The charging curcuit in the librem 5 checks when to stop charging by checking the voltage and the current. The actual bulk of the charge is done with a constant (high) current. Then it switches to charging with a constant voltage at a certain point. From then on it meassures the current which is drawn and stops when it goes below a certain threshhold. This is because a battery doesn’t accept limitless amounts of current, due to the highly variable internal resistance. When a battery is full it somewhat resembles a near ideal voltage source with very high source resistance. Due to this resistance only a minute current can flow into the cell - the famous “trickle charge”.

I fail to see how this charging curcuit is supposed to be hardcoded to a certain battery charge. no way.

The fuel gauge doesn’t seem to be limited by any hardware as well.
If there really is an actual limit in any of those parts though than it’s in software in which case it is pretty dismal that purism didn’t patch it once they knew they’d use a different battery.
Even then it should still display 100%, but be rather inacurate at estimating the remaining amount of charge.

Much rather I think that it doesn’t go above 80% because the thermal protection of the charging kicks in and doesn’t allow for further charging. A third method of terminating a charging process (and a very imprecise and antiquated one) is to meassure the temperature of the battery, because it rises exponentially towards the end of the charging process.
The temperature meassurement is a pretty funky part of the circuitry. The NTC isn’t present at all, but they put two resistors there, one of which has zero ohms and the other is … NC? Not Connected? Normally Closed? They even are shorted the invisible NTC to ground by the zero ohm R30. Really, what a shitty shematic on that page :\


#10

That’s “capacitance”. The capacity value we’re looking for here is energy, AsV=J (Joule)


#11

It should, and it does go down to ~60% after a few charging cycles. In my experience, until the fuel gauge gets calibrated enough it calculates remaining charge to be around 2600mAh when full. Right now, after a few cycles it already got down to 2200mAh for me and should eventually settle around 2000mAh.

Keep in mind that taking the battery off resets the calibration.