Battery charged but button not responding

[flashcoder]

Okay I did not want to open a new thread but could not find a similar issue.. Maybe I am bad at searching.
So I have had 2 P3 batteries forgotten for like 2 years, I left them half charged back then and forgot. Push the button - no response. Tried to charge - does not seem to take the charge. So I opened them up and charged the batteries directly to get them up to about 15V, with 500mA current. Cells were not badly out of balance, I checked with multimeter. So I connected CP2112 and DJi Battery Killer software - both batteries respond, one of them had PF flag with I cleared up, nothing out of ordinary. But the power button still does not work, leds do not blink.
So now I fully charged them with the stock charger and monitored the charging process through DJi Battery Killer log, it shows all the needed info. Both batteries successfully charged to the full capacity but power buttons still dont respond on both of them and leds do not blink.
So I plugged the battery into P3P and it powered on. Cool but still no leds on the battery and power button does not work..
What the hell is going on with these batteries?
Here is a picture of registers from DJi Battery Killer but I dont think its the registers that are responsible for the LEDs and power button side of the battery.

 

[quaddamage]

Check what controls the LEDS and the button, here:

o-gs wiki: P3 Battery schematics

There may be a pad to reset the uC. If not - I guess you'll have to desolder something to make it reset.

 

[flashcoder]

That is a very helpful link. Yes indeed there is a pad for uC reset. Just used a hookup wire to briefly connect RST with GND and it immediately came to life!
So thank you, this issue is solved. Weird that the uC freezes up like this, it is supposed to have a watchdog automatically resetting it according to the chip documentation but I guess DJI did not implement it. It probably happens alot at least to this revision of P3 batteries because it happened to both of my batteries which I forgot to charge.

 

[quaddamage]

it is supposed to have a watchdog automatically resetting it according to the chip documentation but I guess DJI did not implement it

The watchdog needs to be armed in order to work. If due to low supply voltage the execution freezes before first arming, the watchdog stays inactive.

The uC was most likely rebooting over and over due to low voltage, until at some point it finally found a way to freeze without the reset.

 

[Chef Vortivask]

So I connected CP2112 and DJi Battery Killer software - both batteries respond, one of them had PF flag with I cleared up, nothing out of ordinary.

So I'm encountering my own problem that may have a similar cause to yours. I'm curious as to whether or not you used the Battery Killer software to also reset the chip. What I would do with my batteries that I replaced the cells for is unseal them, clear PF, reset chip, and then seal them. I would think that "reset chip" does what you did manually when you hooked up the GND and RESET pads to each other, but if you had done reset chip in the software and still had your issue, then it would seem that they don't do the same exact thing.

 

[quaddamage]

I would think that "reset chip" does what you did manually when you hooked up the GND and RESET pads to each other

You're confusing the chips.

 

[Chef Vortivask]

You're confusing the chips.

That's very possible since I'm not an expert in these sorts of things. If I'm understanding what the Github page is showing correctly, there is the BMS and microcontroller. The microcontroller is what is responsible for communicating with the BMS and relaying that information to the lights and I would assume also the drone via the data lines. When we hook up the CP2112 board, are we essentially bypassing the microcontroller and communicating with the BMS through that? If that's the case, when we "reset chip", that is resetting the BMS, not the microcontroller?

 

[quaddamage]

When we hook up the CP2112 board, are we essentially bypassing the microcontroller and communicating with the BMS through that? If that's the case, when we "reset chip", that is resetting the BMS, not the microcontroller?

That's all correct.

In newer drones (newer than Mavic Pro), DJI stopped adding the micro-controller, and started modifying the BMS firmware instead to add the specific functionality they wanted.

 

[Chef Vortivask]

That's all correct.

Awesome! So I never reset the uC, just the BMS. I think I should reset the uC because even though it appears to be working, maybe there's some sort of glitch that's keeping it from communicating with the drone.

I don't mean to highjack this post, but since we're on the topic of resetting the uC, this particular battery has the DJI01-ROF board, so the layout and markings are different from the OP's. There are two pads I think are reset, but I'm not sure which one is the right one. I suppose I could try both as long as that won't hurt anything, but I'd like to check first to see if anyone knows which one it is for sure.