Hey guys,
i recently got a faulty Nintendo Switch OLED with water damage. Only when plugged in, it will show the Nintendo logo but power down after 2 seconds. Also the battery charging indicator shows up.
I dissassebled the switch and cleaned the mainboard with isoprop. There was a bit of corrosion around some chips, but not much. The battery seems to be dead, i measured 1.1 V between the terminals and I think it had a slight buldge. When plugging the board into the original charger without a battery my ampmeter shows 15V 0.02A.
I then tested for shorts around the usual suspects (M92, BQ, MAX77812, MAX77620) but found none (only low impedance around 77812).
Then i followed the RetroSix wiki and injected 4.2 V into vsys without a battery connected and 1A maximum current. When starting the console by shorting the pins from the connector to the power button, the switch draws 110-130mA for roughly 2 seconds and then turns off. The wiki suggests a faulty MAX77812 or MAX77621 but I dont have shorts on MAX77812 and I think the OLED model doesnt have a MAX77621(?).
How should I proceed? How can i check if it really is the MAX77812? The wiki seems only considering non-OLED models so I am not sure if I can trust it.
Any help is appreciated. This forum just is amazing. I learned so much reading through it, but i could not find my exact problem.
I checked the 3V3 rail close to the audio IC and its present. I think for normal switches it comes from the IC labeled ENXX but for the OLED model its a different one.
I then checked the voltages around the coils next to MAX77812 and MAX77620. For the MAX77620 everything seems to be okay but at the MAX77812 I only get 0V8 and 0V6. Problem is, I dont know if this is normal behavior or if im supposed to have something at the other coils.
Also, one random chip on the backside next to the emmc seems to have a missing voltage. Considering the layout it looks like a replacement for the MAX77621 from the V1/V2 models. I get VSYS (4.2 V) on two of the coils, and 0V on the third.
I am really sorry for the bad quality of the pictures. I can take better ones as soon as I get home.
Thank you
Sadly no. I am suspecting the MAX77620 or MAX77812. Im currently investigating in this direction, but I dont have much time atm. Also the emmc could be faulty.
Any chance that the Switch is stucked on the firmware 17.0.0? I had the same missing voltages on my Switch Lite. After using a picofly I discovered that it was stucked on the firmware 17.0.0. After installing FW 18.1.0 the Switch worked again.
Interesting idea. I don’t really know which firmware is installed. Maybe I can ask the seller.
I thought the 17.0.0 brick only occured for modded consoles? Because this one belonged to a little girl and also doesn’t show any sings of a previous installed modchip.
Capacitor has infinite resistance in general, we are talking about the resistance on the path cap seated, so you dont need to remove it in order to measure the resistance
Small update: I desoldered both of the inductors, to disconnect the rail from other components (that could potentially be faulty or shorted). Still, there is no voltage on the coils when powering on (see the picture from my previous post). Which means either the MAX77812 IC is broken, the enable signal is broken or some of the phase configuration is wrong.
I dont think that it is the enable signal, since the other phases (0.6V and 0.8V) are working fine. I looked at the datasheet and found out that the configuration is done via I2C and not with some hardwired pins (that could be faulty/corroded/etc.). This means, if the other phases are working correct, the I2C configuration should be working for all phases.
I would conclude that the chip is faulty. I think I will order a new one and try to replace it.
I think your chasing a red herring down a rabbit hole here these rails (CPU / GPU) won’t be present if your not at the stage in the boot process where they are enabled. Which is pretty typical for boards in the same condition as yours.
Avoid using the retrosix “wiki” and try to forget anything your’ve aqquired from there as it promotes flat out dumb, dangerous and just plain wrong info. together with direct assurance that if you have X problem then Y is the absolute solution (ie. remove this IC) and then the beginner then proceeds removing something that wasn’t at fault etc then muddying the water in the process, makes the problem worse unbeknownst to them and on and on
Did you take any pics of the corrosion prior to cleanup?
I think your chasing a red herring down a rabbit hole here
I was afraid of that
these rails (CPU / GPU) won’t be present if your not at the stage in the boot process where they are enabled.
Yeah, I thought of that but I could not find information to these rails and when they should have power on the internet. Maybe I should look again
Avoid using the retrosix “wiki”
I tried to be very careful with the information I got from there and tried to find other other proofs for its “conclusion”, but maybe I should start all over again
Did you take any pics of the corrosion prior to cleanup?
Sadly, no. The corrosion was mainly at the lower front side of the board, around the MAX77812 and BQ chip and also inside the battery connector. Also there was a bit on the lower back and at random other positions, but overall not very much.
I ordered a modchip to try to boot into hekate, do you think thats reasonable?
Using a modchip is a valid form of diagnosis when in such situations *sometimes but best be sure in your ability to install it as you could make the situation worse in the process, and even unbeknownst to you further muddying the waters.
Fuel gauge is a possible cause for the symptom of not clearing the first boot logo (though it’s a touch rarer than the usual symptom of just a black screen / BSOD etc)
Many things can cause this issue, BQ, M92, Ram, shorted rails, bad PMIC, bad battery, bad CPU/GPU buck, bad EMMC, list goes on unfortunately. Following visual inspection I usually measure my primary rails resistance to ground first things first before anything to either rule it out or hone in on a potential culprit and go from there
these rails (CPU / GPU) won’t be present if your not at the stage in the boot process where they are enabled. Which is pretty typical for boards in the same condition as yours.
