Hacking Battery Problem After

[Shadow LAG]

Has anyone tried charging the Switch to 100% on the Switch OS, then changing to Linux and charging it a 100% there too? Maybe the Switch OS really does a 100% Soft Shutoff way before the battery charging circuit itself would do.

I appreciate your suggestions; however, if you comb through this thread you can see what has already been tried and this one has been done several times. Also, this method is backwards, realistically you would want to drain the battery in Linux since Horizon shuts down before battery reaches Zero. Then you can charge to 100% outside of RCM in theory. This process is hard on the battery, and unlike Nintendo, I don't want to recommend this to people.

 

[guily6669]

Will try this then. But isn't there thermal paste stuck to the back of it?

Yep, some crap red stuff that looks like bubble gum

You don't have to remove the actual heatsink for the battery. Just removing the IO shield which happens to have thermal paste to the heat pipe. Replacing that TIM won't result in a drastic improvement or anything. Replacing the paste under the actual heatsink might, but that's a more involved process.

But since you are opening the console you might as well change its thermal paste, it makes a drastic difference in temps and maybe help save a bit of battery as the fans will be running slower. And on the dock where it heats up like hell, I think the heating will kinda be fixed with a new good thermal compound on the die...

The backplate wont do much, but if you fill it with thermal paste + thermal pads it still make a considerable difference as the backplate still has a lot of metal that covers the whole Switch even tough its not thick.

This might also help for later if OC gets possible for someone with CFW for some emulators or for Linux... I don't think the original thermals will leave any room for OC.

 

[alkar]

I appreciate your suggestions; however, if you comb through this thread you can see what has already been tried and this one has been done several times. Also, this method is backwards, realistically you would want to drain the battery in Linux since Horizon shuts down before battery reaches Zero. Then you can charge to 100% outside of RCM in theory. This process is hard on the battery, and unlike Nintendo, I don't want to recommend this to people.

People who are claiming a simple charging cycle works don't have the same bug for sure.

You can't even let it discharge from Horizon anyway since it shutdowns every minute or so when under 30%. Letting it discharge in RCM works but it won't fix the issue, it still had the problem this time around 50% !

Only fix for me was to unplug the battery which fixed it instantly.

BTW can any of you guys explain to me what difference it made to actually UNPLUG the battery rather than discharge it ? I mean when it's 0% it should act the same as unplugged no ?

 

[Adubbz]

For reference, I've made an issue on fail0verflow's issue tracker here: https://github.com/fail0verflow/switch-linux/issues/6. If you have any steps to reproduce the issue, or other relevant information I'd strongly recommend you add it. However please don't use it as a place to complain.

 

[aut0mat3d]

BTW can any of you guys explain to me what difference it made to actually UNPLUG the battery rather than discharge it ? I mean when it's 0% it should act the same as unplugged no ?

There is a protection circuit built in which let the battery not fully discharge, which is essential: If you complete discharge a LIIO Battery it would probably get damaged.
So there is always enough power for RTC and other (do not know) stuff

 

[TerraPhantm]

BTW can any of you guys explain to me what difference it made to actually UNPLUG the battery rather than discharge it ? I mean when it's 0% it should act the same as unplugged no ?

No. A discharged battery will still be at 3v or so

 

[Shadow LAG]

People who are claiming a simple charging cycle works don't have the same bug for sure.

You can't even let it discharge from Horizon anyway since it shutdowns every minute or so when under 30%. Letting it discharge in RCM works but it won't fix the issue, it still had the problem this time around 50% !

Only fix for me was to unplug the battery which fixed it instantly.

BTW can any of you guys explain to me what difference it made to actually UNPLUG the battery rather than discharge it ? I mean when it's 0% it should act the same as unplugged no ?

When you unplug the battery and hold the power button, you force the capacitors to discharge and clear the memory. This resets sensors and memory errors. It's not the battery you are effecting, it is the charge on the capacitors. You cannot discharge them normally when you have a battery connected as holding the power button will still pull from battery.

 

[TerraPhantm]

Pressing the power button probably doesn't do much. Just leaving it unplugged for a few seconds should do the trick.

 

[Shadow LAG]

Pressing the power button probably doesn't do much. Just leaving it unplugged for a few seconds should do the trick.

It drains the capacitors. If you just unplug it you kill power, but capacitors can keep charged. This is the difference between a computer with static booting and not booting. You clear memory when you discharge.

 

[TerraPhantm]

It drains the capacitors. If you just unplug it you kill power, but capacitors can keep charged. This is the difference between a computer with static booting and not booting. You clear memory when you discharge.

The capacitors won't take that long to drain. Probably within a fraction of a second of unplugging the battery. Remember, the power button is a soft-switch - meaning that there is something constantly monitoring the state of that button. That circuitry will drain the capacitors on its own when you pull the battery.

 

[aut0mat3d]

The capacitors won't take that long to drain. Probably within a fraction of a second of unplugging the battery.

Not really. It depends what component is sourced with a cap... (rtc, ..)
On a actual project i had to deal with too low power consumption of a uc in deep sleep which was not wanted for a particular reason...

 

[DerProGamer2000]

I noticed that my Switch was at 33% after it crashed while playing around with Arch Linux too and wouldn't turn on anymore until I had it on the Dock and turned it on there.
I think that the Switch OS has some internal calibration where it doesn't expect anything else draining the battery and gets confused after it did so, however it may correct itself after a few tries.

lel i had that too xD i played around with linux a bit and then played xenoblade in handheld it was exactly 33% battery and would not turn on again but after putting it in the dock and letting it fully charge it fixed itself. i lost about 2 hours of play time because no autosave yey. i think its something similar to the 3.0.0 battery bug just not as bad

 

[TerraPhantm]

Not really. It depends what component is sourced with a cap... (rtc, ..)
On a actual project i had to deal with too low power consumption of a uc in deep sleep which was not wanted for a particular reason...

Meh, I'll test mine without actually pressing the power button when fedex drops off my triwing set. I'm confident it'll work all the same.

 

[guily6669]

It drains the capacitors. If you just unplug it you kill power, but capacitors can keep charged. This is the difference between a computer with static booting and not booting. You clear memory when you discharge.

Yeah we talked already about it, just to be safe its better to press power button, but those devices are not using Monster capacity and judging by videos on youtube it seems in just a few seconds of removing the battery connection the Swtich already resets by it self.

Theres really a huge difference from monster capacitors connected to nothing that can kill you even a long time after they are fully charged to small capacitors more just to stabilize a bit the electricity that are connected to a full motherboard passing trough a lot of components that easily drain them considerably fast.

 

[Shadow LAG]

Meh, I'll test mine without actually pressing the power button when fedex drops off my triwing set. I'm confident it'll work all the same.

Let us know how it goes.

 

[Haztel]

Looks like that battery reset did it's magic. Working fine for me. I made some dumb mistakes taking apart the switch but luckily I didn't strip any of screws. I even screwed the wrong screw in the wrong hole but I managed to get it back in perfect condition.

@Shadow LAG So you said you couldn't replicate the issue. Do you suppose that if you do this fix once that you might not have to do it again?

 

[Shadow LAG]

Looks like that battery reset did it's magic. Working fine for me. I made some dumb mistakes taking apart the switch but luckily I didn't strip any of screws. I even screwed the wrong screw in the wrong hole but I managed to get it back in perfect condition.

@Shadow LAG So you said you couldn't replicate the issue. Do you suppose that if you do this fix once that you might not have to do it again?

That's what I'm calling on people for to report back. No one has spoken up. I've never had this bug myself and I've used every RCM exploit available. Also in other good news,



Now you can do the battery pull, cold boot without backlash. Apparently FTC has ruled you have the right to fix your own system.

 

[NightsEkim]

What exactly is the battery bug, from what I read its a bad sync between switches OS and the actual battery life remaining leading to a fast discharge?

 

[Haztel]

What exactly is the battery bug, from what I read its a bad sync between switches OS and the actual battery life remaining leading to a fast discharge?

You're close. Not exactly fast discharging but when the system hits around 15-30% it shuts off because of the system desync thinking the battery is low.

 

[ut0pia]

What exactly is the battery bug, from what I read its a bad sync between switches OS and the actual battery life remaining leading to a fast discharge?

Mine forces shutdown at around 43% charge. So I loose almost half of battery and loose any game progress if not aware. Don't even think about flashing nand in this state.

I will receive my ifixit kit today and will report back if unpluging the battery + discharge capacitor by pressing power at least 20sec fix the issue