Hardware My switch doesn't charge, won't turn on, or is otherwise not functioning at all. Where do I start?

[CHTechRepair]

I haven't seen a guide for where to begin if the console has no power at all. I'm writing this as if a Switch is in front of me and I'm looking at it for the first time. Typically the only switches I don't fix are severely liquid damaged ones or CPU/GPU failure as they take too much effort compared to $300 for a completely new unit. They are still entirely repairable and I have repaired several. These aren't the only situations that can occur, tons of things fail on these consoles. Without a schematic things get complicated quickly.


Everything I'll be going over in this thread is not beginner level. Should you figure out what's wrong with your switch you will likely need to find a local or mail-in repair shop confident in their ability to accomplish these repairs. It is very easy to ruin one of these consoles through inexperience and a quick glace at this forum will show you that. Understand by following this guide you are taking huge risk to your device. I do this constantly on way more than Switches; it's not a hobby.


iFixit's teardown is a good photo reference for where the components are and will provide a visual guide. Step 11 and 12 are what you want.


https://www.ifixit.com/Teardown/Nintendo+Switch+Teardown/78263



Where do we start?

The charging port and charger/dock


The most common mode of failure is the port by far. Even a good looking port can have separated traces on the logic board. The first thing I do is visually inspect this area with a bright light and a microscope but this isn't needed. If you see any rogue gold, anything that looks even slightly bent, it's safe to say your problem is here. Next we check the other end, the charger and dock. Are any of the pins bent? Replace these as necessary.


Everything looking good? Well, it can still be the port, but it seems like we have some further electronics issues so let's get down to it.


You'll need a USB-C meter that reads both voltage and amperage out to get any further in diagnosis without taking the console apart. Amazon sells these for next to nothing and you'll get it in just a few days. I would greatly suggest getting one as it will allow for a fast diagnostic even when working with just a board. You should also have a multimeter to test electronics on the board.


Let's plug it up! The charger, regardless if original or aftermarket, will negotiate with the M92T36 chip inside and request the charger send 15.0v here. This is where the meter starts to be very useful. I'm going to include a few specific situations and what I take these readings to mean on initial device intake. This isn't a guarantee, just a "usually x means y". A place to begin. Keep in mind, you may see 14.7v, this is enough you don't have an issue.


PLUG THE PORT IN BOTH DIRECTIONS!

Spoiler: Scenario #1

15.0v

0.13a-0.46a

Battery is charging correctly, starts around 0.12a if the battery is very low in voltage and increases as it charges. Above 0.46a means the device is either on or has a short as it requires fast charging negotiation through firmware.

Spoiler: Scenario #2

15.0v

Up to 0.10a then down to 0.00a. May stay at 0.00a

Battery/Battery Coil (gray coil beside the port/LDI) (Device doesn't think it has a battery)

Spoiler: Scenario #3

15.0v

0.036a (below 0.07a but stable)

Problem before software preventing enough power from building to even charge the battery, typically Battery Charging IC. Test Battery area.

Spoiler: Scenario #4

5.0v

0.00a

May have no display on one side of USB-C. Typically a short on M92T36 preventing negotiation.

Spoiler: Scenario #5

Display on meter doesn't turn on or reads 0s.

No continuity or short on DC-IN rail.

Spoiler: Scenario #6

5.0v or 15.0v

Amperage not stable or similar to any other mentioned scenario

Can likely be pinned on liquid damage.

Let's get a little deeper. You should have one of these situations going on. If you don't this can still help you. I'm going to focus on each one and explain what I would look at from here. We're doing it by number.

Spoiler: Troubleshooting Scenario #1 Warning: Long

Probably the worst situation to be in. Your battery is charging but your device isn't showing a battery charging icon. This means firmware isn't successfully starting and a lot of things have hands in this process. Confirm you have no backlight, sound, or image on the LCD otherwise you have a backlight or LCD driver issue. We need to verify we have no boot. Take the logic board out of the device and plug it in. We should get up to 0.10a constantly on the meter. Normally, when firmware has started, the device will go to sleep when it figures out it has no battery to charge and go to 0.00a. If we're hanging here the CPU is not on. I'm unsure if firmware can fail to this point, but RCM and a device that can recognize an RCM Switch can be used to confirm the CPU isn't responding. Without modification, firmware failure is the end of the device anyway aside from Nintendo or a CPU and eMMC swap. Assuming you have no RCM, inspect the board for any minor liquid damage to give hints where the issue may lie. If there is none, let's walk through what's needed to boot. First, we need to charge the battery, or at least supply 3.7v or so to the rail the Battery Charging IC outputs. This rail powers the PMICs and basic power electronics in the device. It can be measured on the capacitor or coil to the left of the battery charging IC or right of the port. Without a battery it will likely be 4.2v. If you have a charging battery you have this. Things from here on are based on my experience and may be dead wrong. I'll be editing this as I learn more in an effort to put the best info I have out there. From here, we check the basic power electronics. First check the two MAX77621 chips beside the CPU. On the new HAD-CPU boards there is only one IC doing the work of the two. Test all coils and capacitors here. A word of caution: these are some of the lowest impedance rails in the device. Use a diode mode measurement and make sure you don't have a 0.000 reading on both sides of capacitors. They typically only go short in the event of liquid damage. If they were working and being signaled to operate we would have about 0.8v to the CPU. From here we flip the board. Above the Realtek Audio Driver on the back of the board is the area for 3.3v. Test the large coil and capacitors here. If there is no voltage in this area, the small IC may be the issue, or it may not be getting the proper signals again. This is where things get muddy quickly since I don't know where those signals come from, but I've never seen this IC cause a problem. Next we go to MAX77620 just to the right of where we were. This PMIC has dealings in signaling and powering other ICs I believe the list includes the electronics we just tested on top of the CPU and RAM. Most of the things on the board deal with it in some way. Test all the coils and capacitors here before trying to replace this chip. This is the first IC I would replace in this scenario where I can't confirm an issue on any other power electronics. If you've replaced it, and all other electronics seem to be testing good, we move forward into even more complicated territory. Above MAX77620 is the Southbridge/SMC. What all this does is not known since it's entirely proprietary, but if our power electronics are working, it becomes a target. It takes 3.3v in, so the console has to be at that point for it to cause an issue. Again, never seen an issue here without liquid damage, but that doesn't make it impossible. It's easier to work on this than the CPU, so I figured I'd bring some attention to it. We've narrowed ourselves to three potential components if you've ruled everything else out. CPU, and the two RAM modules. Your likely issue is the CPU solder joints breaking if you've made it this far. The CPU can not be replaced without swapping the eMMC too, so realistically only reballing or reflowing will yield any results. This is the most difficult thing to do to a switch board, and should only be done with the understanding that you can't make a dead board any worse, and donors are always useful in the future. If you get new info on your situation you are welcome to respond to this thread and I'll do my best to offer advice!

Spoiler: Troubleshooting Scenario #2

Test battery coil for continuity. (Large gray coil beside charging port) If coil is good: test/replace battery. Same issue: never seen it but focus on battery charging IC between the two components.

Spoiler: Troubleshooting Scenario #3

Usually the battery charging IC causes these failures even without a measurable short. It's good to test the area around the IC and confirm nothing has obviously failed first.

Spoiler: Troubleshooting Scenario #4

It's good to try multiple chargers first to confirm the lack of negotiation isn't the fault of the power supply. Test for shorts around M92T36, you will likely find one. If so replace IC. If you do not, replace the charging port. You will likely find broken traces under the port, these repairs can be very difficult.

Spoiler: Troubleshooting Scenario #5

Before tearing your console down try another power supply. If you have the same behavior, test M92T36 for any nearby short first. Inspect the board for any obvious shorts or damage. Supply 5v to the DC-IN manually with a power supply to test for a short. There is a test point just above the port. Next, replace the charging port, you likely have broken traces if it looked fine externally.

Spoiler: Troubleshooting Scenario #6

Your symptoms indicate a partial short, very common of liquid damage and switches are no strangers to being liquid damaged. You'll need to inspect the entire board to find the source. Most of the time the failed component will get hot and a thermal camera or alcohol will help you locate it. If the board was liquid damaged, take all shields off and carefully clean the board. Look for missing and severely damaged components. Replace as necessary. I use an ultrasonic cleaner to clean these boards as thoroughly as possible. Liquid damaged switches are hit or miss. I have fixed some that live forever, others develop other issues later on. The most common issue is damage under the CPU causing solder joints to break later on.

Good luck to everyone, may your switch turn back on and live a long life!

Edit: cleanup and a new troubleshooting #1 with better info on the electronics needed to boot.

 

[Everlasting1337]

Thanks for the detailed post. Consider using spoilers though.

 

[CHTechRepair]

Thanks for the detailed post. Consider using spoilers though.

Will do, wrote this up on my phone let me get that sorted. Edit: Done, thanks so much for the feedback!

 

[CapTec]

Will do, wrote this up on my phone let me get that sorted. Edit: Done, thanks so much for the feedback!

This is a great write-up.

Lots of good detail here! I have a question about a problem I'm having with a switch that might make a good addition to your guide. If one of the voltage rails is out of spec, e.g. significantly lower than it should be like the 3.3v rail reading less than 1v (0.700v in my case) where should people look?

 

[CHTechRepair]

This is a great write-up.

Lots of good detail here! I have a question about a problem I'm having with a switch that might make a good addition to your guide. If one of the voltage rails is out of spec, e.g. significantly lower than it should be like the 3.3v rail reading less than 1v (0.700v in my case) where should people look?

I'm fairly sure the 3.3v logic/IO rail is switched by MAXX77620. Take a diode mode measurement above P13USB there's a large capacitor on this rail. Red probe on ground, black probe on 3.3v side of capacitor. You want above .300 here, less is going to be a partial short more than likely. Since this goes to nearly everything on the board doing logic, the best method is to inject 3.3v into the rail and look for heat if you do infact have a short. The other option is to simply remove chips until the short is relieved. ICs on this rail include: eMMC, GPU, MAXX77620, P13USB, M92T36, Audio IC, probably Digi IC? never tested since it's part of game card reader. If you don't have a short, replace MAXX77620 it's likely not switching properly. There may be a transistor involved in this or it may be in the IC itself. I've never had to get that deep on 3.3v usually a chip is shorting it.

 

[dicamarques]

This post comes in an excellent time!
https://gbatemp.net/threads/broken-switch-wont-turn-on-but-it-charges.581554

Please have a look in mine and see if you can give any hints! Thanks

 

[CapTec]

I'm fairly sure the 3.3v logic/IO rail is switched by MAXX77620. Take a diode mode measurement above P13USB there's a large capacitor on this rail. Red probe on ground, black probe on 3.3v side of capacitor. You want above .300 here, less is going to be a partial short more than likely. Since this goes to nearly everything on the board doing logic, the best method is to inject 3.3v into the rail and look for heat if you do infact have a short. The other option is to simply remove chips until the short is relieved. ICs on this rail include: eMMC, GPU, MAXX77620, P13USB, M92T36, Audio IC, probably Digi IC? never tested since it's part of game card reader. If you don't have a short, replace MAXX77620 it's likely not switching properly. There may be a transistor involved in this or it may be in the IC itself. I've never had to get that deep on 3.3v usually a chip is shorting it.

Thanks for the reply! I worked it out in the end. It wasn't an IC believe it or not. The inductor next to the RTL chip had failed. I took another from a donor board and voila.

 

[CHTechRepair]

Thanks for the reply! I worked it out in the end. It wasn't an IC believe it or not. The inductor next to the RTL chip had failed. I took another from a donor board and voila.

Good work! This is why it always helps to test around an IC before you go to replace it, sometimes components around it really are the problem you just never know!

Based on this info I wonder if the 3.3v rail actually comes from a transistor above M92 or in that area above Audio IC? Strange spot for a break in that entire rail if MAX77620 actually creates it. I'll have to take that inductor off on a good board and see where 3.3v still exists! Great info here, and another thing to test in the future!

 

[Ra-D-OH-3H]

Could you indicate which coil is it? I have this bug: Missing 5v on pin36 VCONN_IN. I have 1V there.
But no shorts

 

[lpoolm]

Thanks for the detailed write up, will keep this thread saved!
I take it you do repairs then for a living as you seem more knowledgeable then a hobiest?

 

[CHTechRepair]

Could you indicate which coil is it? I have this bug: Missing 5v on pin36 VCONN_IN. I have 1V there.
But no shorts

I personally have never had to troubleshoot any 5v rail aside from DC-IN. It would require boosting since the battery voltage is always below 5v, which leads me to believe it wouldn't be required to start the device as everything logic based is 3.3v or lower. I'm also not familiar with any rail names or schematic diagrams for specific ICs so you'll have to give me a picture of where you're expecting this voltage and I can try to help. There are plenty of 1v rails in the console. I believe the coil CapTec was referring to is above the Realtek Audio Driver on the back on the board. That area is involved in the 3.3v IO circuit.

Edit: Looking at the board right there, it seems that small IC may be a buck converter for the 3.3v rail. If the inductor is good, and you don't have 3.3v with no short you may try replacing this IC. This is likely signaled by MAX77620, similar to MAX77621, so a missing signal may also be the cause of your issue.

Thanks for the detailed write up, will keep this thread saved!
I take it you do repairs then for a living as you seem more knowledgeable then a hobiest?

I do. I mostly do phone, tablet, and laptop repair but recently I've seen more switches. They seem to have a lot of electronics failure at least that's mostly what I see them for. Not a lot of stuff like this out there so I figured I'd share what I do when I see one to try and diagnose without a teardown. There are a lot of common failures, then there are just one off failures. In a short time on here I've learned a lot more about the hardware, even this thread has already brought a new failure to light.

 

[Ra-D-OH-3H]

Today I replaced the M92T36, P13USB and then the same. the console charges the batteries, no reaction,
I replaced them with new ones for confidence. It seems to me that the Tegra is working, because it feels delicate warm
I have to order maxx77620 and 77621.

 

[CapTec]

I had the mother of all faulty Switches to repair, same one I mentioned above with the 3.3v issue.

First thing I noticed was that the USB port was badly damaged, so that had to go. That port took out the M92T36, P13USB AND the BQ chip, so those all had to be replaced, which is pretty common with switches that have damaged USB ports. I also accidentally torched the left Joy-Con socket when replacing the BQ chip so had to take one from a donor board. During all of this the damned battery connector socket literally lifted from the board, fortunately no trace damage so I was able to solder another on without a hitch fortunately.

The MAX77620 had gone funny so had to be replaced. Then I discovered the fuel gauge IC, the chip on the bottom of the board, directly below the battery socket on the underside of the board had shorted, so I replaced that as well.

Then I noticed the inductor I mentioned earlier had gone iffy, so yet another chip replaced. That got me the 3.3v rail back but I still didn't have LCD display output.

THEN it gets worse. Then I realised that a diode next to the LCD buck converter IC had literally blown, so I replaced that and noticed the buck converter had gone bad, as in I noticed a hole in the IC, so had to replace that. Powered it on and magic smoke came off the new buck converter.

So at that point I just took the caps, diodes, inductor and a large resistor from a donor board and replaced the lot around where the buck converter was and placed yet another buck converter on. Sadly I also torched the FPC connector for the LCD, so had to also replace that

Anyways after all that work I brought the poor switch back to life. I even gave it a brand new digitiser and shell because the old ones were pretty battered. Now it runs and looks like it just came out of the factory. The PCB even had a short session in an ultrasonic because the damned thing also had water damage. So even the PCB looks like factory.

And after all that effort? The damned thing has a bloody parental code lock set. You literally couldn't make it up if you tried. Anyways, here's some pictures I took whilst I was still diagnosing the many, many faults. I made these using a mosaic of pictures I took using my microscope. So these are probably the highest resolution switch PCB images you'll find on the web.

It's the most herculean effort I've ever put in to repairing a switch. Maybe all things that could be incorporated into your repair checks though, lol.

 

[Everlasting1337]

@CapTec Hold my beer, still good effort though.

https://d3nevzfk7ii3be.cloudfront.net/igi/MvJEDTUnxMS4Olqs.full (Front)

https://d3nevzfk7ii3be.cloudfront.net/igi/ikTQ6tTjHE4O6VbZ.full (Back)

 

[Ramsaber]

Do you have any suggestions for a perfect working switch and no video out to dock? Replaced m92 p13 port three or four times and b24 just cause.

 

[Ramsaber]

Do you have any suggestions for a perfect working switch and no video out to dock? Replaced m92 p13 port three or four times and b24 just cause.

 

[Ramsaber]

Do you have any suggestions for a perfect working switch and no video out to dock? Replaced m92 p13 port three or four times and b24 just cause.

--------------------- MERGED ---------------------------

Not sure why it made three of these

 

[CapTec]

Do you have any suggestions for a perfect working switch and no video out to dock? Replaced m92 p13 port three or four times and b24 just cause.

--------------------- MERGED ---------------------------

Not sure why it made three of these

Could be one of the EMI filters or one of the diode arrays maybe? Stick your meter in diode mode and place the red probe on a grounding point and touch the black probe to the top of each USB port pin that you can access. Note down the values and the orientation you took the readings in and post back here.

 

[mgabe]

Great write up, going to follow it later on. Curious if you can tell me if the LCD needs to be plugged in for a successful boot?

I bought a used Switch that was advertised as screen died, bought new one. When I got it home and took the back plate off; it’s clear someone went digging with a fork, almost every ribbon cable is scratched, bent, etched, you name it (power and lcd ribbons included).

So just curious if that LCD ribbon needs to be plugged in at all for a successful boot or if I can troubleshoot assuming all I have at my disposal for visuals is docked mode.

Sat down this morning and pulled the power ribbon out, I’m fairly certain it’s damaged and severed so I should probably start by ordering a replacement.

 

[Ramsaber]

Ok I definately don't know how to use this site while leaving it open. Didint see any posts besides mine until today. As for checking if something is working you should be able to set it into the dock. You will get slow blink charge until it makes it fill enough to start. When it does if there is any power on device the dock will show solid green. I am moving close to 50 switches a week currently. I honestly got rid of the last one I had with no video on dock. When I run into it again I will update thank you to everyone on the thread

 

[dicamarques]

@CHTechRepair Small update, after all the testing the switch has the eMMC corrupted (possible from sxOS brick)
Managed to boot it to RCM and hey a payload worked, until I took it a part and noticed that the guy from the store forgot to connect one of the wifi antenas.
Short story, while plugging the LCD cable in, it wasn't completely inside, and I touched it while it was turned on. It showed some green lines of pixels and after that I removed the battery and reinserted the cable.
This time the screen doesn't even show any image.

My question is, how likely is that I just need a new screen and didn't burn anything on the board?

 

[Ramsaber]

@CHTechRepair Small update, after all the testing the switch has the eMMC corrupted (possible from sxOS brick)
Managed to boot it to RCM and hey a payload worked, until I took it a part and noticed that the guy from the store forgot to connect one of the wifi antenas.
Short story, while plugging the LCD cable in, it wasn't completely inside, and I touched it while it was turned on. It showed some green lines of pixels and after that I removed the battery and reinserted the cable.
This time the screen doesn't even show any image.

My question is, how likely is that I just need a new screen and didn't burn anything on the board?

If you want to check for only a screen and can power it on normally I would suggest checking for sound. I was able to get the device to power on and function to get sound with a bad screen. If you can't get it to boot I am not sure what sounds would come from rcm boot. Someone with more knowledge or another device to boot to rcm may be able to help with that