Aha, so the rail you’ve labeled “VDD_CPU_1V1” is the Ram rail (secondary in this case) i was talking about, sorry my bad as it can measure anywhere from 1.05 to 1.15V on any given board. So it’s all good in this case. Off the top of my head all your measurements look good. The rail you’ve got labeled as “VDD_1V05A” I’m guesing the info within the brackets is from the datasheet “from VDD 1V3” ? if yes then it makes sense to not being seeing a voltage here as your missing 1V3/1V35.
So on Mariko boards missing 1V35 is pretty typical, it seems to to suggest something pretty early on in the boot stage is halted. Unfortunately it’s hard to be specific in this case as pretty well anything could potentially halt boot but I’l just throw a few possible or common ones - Fuel guage would probably be most common, CPU/GPU buck reg, M92 sometimes, Bent USB pins sometimes, Bent LCD pins, then we get even rarer but still possble, P13, audio IC etc
looking at your images of the board though everything looks in pretty decent condition, no signs of prior rework afaict, permanent marker is normal from factory… The only thing that stands out to me is over at the other Max IC below the SoC (CPU/GPU reg) it looks as though one of the inductors has chipped possibly indicating impact damage, it could well be this IC itself has suffered damage as a result and is causing the problems or a ball on this IC is floating… or worse some other IC on the board has either of those problems following impact - So I’d start taking a close look in this area to begin with, though I could be wrong and it could be camera//lights playing tricks on me
hard to say but as the damage at this inductor tracks with the damaged backlight ribbon (I imagine it sheared off on impact perhaps (?) ) then I’d say damage to the Max IC or it’s joints is the most likely candidate.
If it were me I’d swap out that Max IC and inductor (although I imagine the inductor is actually fine tbh) just to rule it out…but, It’s important to note I can whip off an IC from a donor and reball it and prep the board etc in a couple of mins and reball the old IC if it was never the issue to begin with for later use. If your just starting out or don’t have the tools (or a preballed IC if you don’t have the tools) then maybe this isn’t the best first step as you could potentially spend hours on this or cause secondary damage. Maybe it would be best to start with the fuel gauge (even though I don’t think that’ll be your issue) as theyre cheap and you can easily find them preballed etc but I think it’s up to you,
I don’t think it will be the M92 IC if you can’t see any signs of damage to the USB port and also because your rail checks earlier make it so much less likely.
You can measure the resistance to ground on the inductor surrounding the Max IC in question but given this is a suspected open scenario I don’t think the odds are stacked in your favour that we’ll see anything revealed specifically on the outputs.
It might also be worth looking very carefully at all these styles of IC’s (glass package style) and just seeing if there is any indicators of harline fractures or chips onn any of them, with particular attention paid to that Max IC below the SoC and the fuel gauge which might help narrow things down, though I imagine you will have looked for this since last time.
No, no random reflows, as if a ball is floating around under there, for example a ball sitting around SYS (just as an example) and some sub 1.2V I/O and a random reflow causes them to touch (and it’s sods law it will happen ) you’d just destroy the SoC the moment you apply power.
Oh, one thing I completely fogot about. Have you got a known good display to test in conjunction with this board? Or do you think your able to repair the backlight ribbon? (I have a post here somewhere going into detail regarding repairing ribbons of this style if your not sure)
As, while it’s unlikely, it might well be outputting a panic colour code on the screen which could further narrow things down. For example, if you were getting an orange screen it would further point towards that Max IC below the SoC or an SoC issue itself is the most likely problem. Just for example
And the plot thickens! I love this fault finding part, thank you so much for the detailed explanations!
Now that you mention it, the ripped backlight ribbon and the chipped inductor are probably correlated, you must be right! The inductor is still conductive, so I don’t think it’s damaged
Never had to fix a ribbon before (I’ve always swapped them in Macbooks for instance), so any resource will be appreciated
Nothing else looks suspicious
I just bought another faulty Switch (because I make bad decisions). Hopefully, the display will be fine, and I will test the board looking for panic colour codes before going crazy with the hot air station.
I’ve never soldered a BGA, let alone reballed one. But I’m eager to learn and will try with preballed ones, starting with the fuel gauge then the max IC
I’ll keep you posted, thanks for your help!
topic here on the subject, different ribbon but same principles
It’s worth repairing them as the replacement LCD panels avaliable for Switch which while cheap are either the manufacturer casts offs much like the enviroment for phone screens which exhibit ghosting, colour patches and other similar issues or theyr’e refurbished with new ribbons bonded or soldered on (in the case of backlight) and that’s if they don’t turn up DOA and then you have the nightmare of getting the panels into the assembly square etc which is not a problem you’ll be used to if your coming from a macbook or ipad background as the fit and finish there is usually spot on and the frame won’t allow you to fit the panel crooked.
Just practice this elsewhere on some old phone or laptop boards prior to trying on the actual device of repair and get to grips with it. after failing the first 10 / 100 times things will start to take shape
Hey there,
So first update: I received the other faulty Switch. It was covered in flux and the USB port had obviously been changed. The ammeter wasn’t even turning on, despite no short showing up elsewhere on the board, so I decided to plug in a charged battery to see what was happening. And it turned on with a 2101-0001 error. So as soon as I receive the new USB ports I ordered, I’m going to swap it. But I have high hopes to fix switch #2
Then I tried to fix the backlight ribbon on Switch #1. I carefully followed your tips.
Scraped some of the plastic
If the port plastic has melted then yeah it needs changing but if the plastic is all intact and “stem” and pins all look good then it might be worthwhile checking continuity with a USBC breakout board prior, if all checks out then I’d swap the M92 first.
Thanks a lot. I just checked with a breakout board, and I’m missing continuity on a few pins here and there. I tested the pins in diode mode, and some like V+ are reading 2.8V, which I have a feeling isn’t right
So that connector is definitely dodgy, I’ll change it first, then I’ll try M92T36. I’ll keep you posted!
Quick update:
I’ve just swapped the USB C port, as it was badly soldered. Never did that before. It came out nice and easy, so I set a new one in place and heated from below. And I melted the plastic inside the connector
Tried my luck a second time with another connector, this time by heating from the top. The inside was immaculate… but I melted the plastic outside!
Third time’s the charm, I thought, and remembered what you’ve said about failing a hundred times… And i did it! Continuity was there, on both sides, the joint were nice and shiny…
I proudly grabbed my ammeter to contemplate my triumph and plugged it in: nothing. What an anticlimax.
So next step: I’m going to change M92, as you’ve completely anticipated
Nice work, glad you got it in the end. Everyone has there own preference on doing the USB, I personally like to wick the ground through hole pads clean, go low and slow with my hot air at the USB for a few reasons (I’ll touch on them in a bit) and I always hit it from the underside at approx 360/380C low to med air, nozzle pointing away from the USB at the section of the board where the pins reside (on the opposing side) While you can potentially melt the USB port at 360C you’d have to try pretty hard to do so at this temp (though every station is different)
Now my reasons for going low an slow - Ram resides very close by, the amount of boards I’ve had to reball Ram as a result of bad rework at the USB port is too many to count. Too high heat concentrated in this area can lead to layer delamination… this is unfortunately another common fialure I’m seeing all to often atm - this results in open lines anywhere from the BQ, fuel gauge or P13 running past this area of the board, sometimes this can be repaired with jumpers but more often than not it can’t
Another common one… and I’ve seen a number of prominent youtubers “teaching”: this, is, reflowing from the top, as they’re unable to get enough heat in as a result of plastic obscuring the inner row they crank the temperature and air right up and aggresively push down on the connector during and following “reflow” - This winds me up for a few reasons as once again, the amound of boards I look at where the board has basically turned into a banana as a result of this is again beyond count Provided the pads are properly prepped and the the board is fully saturated heat wise, there is no need to be doing this.
There is my rant over hoefully this helps you out the next time you do a USB connector replacement
Haha, I like to think of it like a machinist fine tuning his feeds and speeds for the best possible results and finish on some metal stock. Or by failing you learn what not to do for next time
Sounds like you should make some youtube vidoes to show proper smd component soldering techniques, I my self am guilty of the ‘apply slight pressure’ to an ic resulting in excess solder coming out since that was the only way I’ve seen done and until coming across your posts here I thought it was a normal thing to do, though I’ve never left indent marks like I’ve seen others do, I used to let the weight of my tweezers do the work
Yeah your probably fine then. This is one of those areas where you can get away with it in a lot of cases - but,
I’ll just briefly touch on why pressing chips down is not good practice for anyone else. Let’s use M92 as an example - under the plastic is a glass like package (much like the Max IC’s dotted around the Switch board) so by pressing down on the IC (and especially if leaving an indent) then you are in affect compromising the integrity of the IC… What a lot of people don’t realise is damaged chips can still function, for example, you could physically damage a mosfet and it might still work, but it might work at 30% less efficiency or it might just get 50% hotter in operation, well the same is true for ICs such as M92 and others. The other issue is, the landing pads, particularly the central ground pad is basically designed to hold a set volume of solder, this allows some degree of flexibility of the board (remember chips are effectively glass like in nature) which is even more important on handheld devices, so by squeezing the chips down (usually to compensate for bad pad prep or skills) then you’ve basically removed any form of leeway. If anybody wants to see this for themselves, just take an IC like the Max PMIC, pretend it’s not a BGA IC, solder it to some perf board and press the chip down, if it doesn’t crack or chip during this process then try gently twisting the perf board after (which will happen naturally on a handheld device), chances are pretty high the IC will crack, which gives you a very good idea as to whats going on inside other IC’s even if they are plastic encased. This is why in a lot of datasheets they will sometimes go into excruciating detail as to the pad size etc to be used to avoid such issues. Things get even more complex when we talk about BGA ICs.
But yeah, always second guess these youtubers, if all theyr’e doing is fixing obvious physical damage or dead shorts to ground then they probably don’t know as much as they’d have you believe.
As someone mentioned earlier in the thread, it might be worth looking into an eMMC/NAND possibility. I actually just had nearly the exact same situation with a patched Switch going into RCM (with exception of maybe having the power rails in tact). I found a broken trace on the board that went directly from the eMMC module to the SoC. It might be worth comparing the diode readings on the other Switch to make sure that there aren’t broken connections or similar, or seeing if there’s a way to test the eMMC in another unpatched Switch. My understanding is that if the eMMC can’t be read, that even a patched Switch will go into RCM.
Some prior repair had been attempted on the USB port (and probably M92)
what I did in my previous post:
I changed the USB port because it was badly soldered. The ammeter would still remain dark.
What I did tonight:
I changed M92T36 with a new one. It didn’t go smoothly. It did at first, then I accidentally knocked the tiny resistor and the tiny cap at the bottom right hand corner, it took me ages to tidy everything, I was tired, it was late, so not a very pleasant experience
when I plugged the charger in, the ammeter turned on (hurray!)
so I reassembled everything, and plugged it in again… and nothing happened. The ammeter is reading 15V and 0,35A. The APU is getting warm
it’s no longer booting on battery either, looks like I’ve made it worse
Could it be M92 again? Maybe it’s badly soldered? Should I check every pin in diode mode?
My guess it’s soldered badly or one of the components surrounding it. Might be worthwhile double checking the resistor you knocked off is measuring right, and don’t be fooled theres a few resistors in this area that will give deceptive in circuit readings.
Failing that, you could lift the IC straight off the bat and see if the symptoms have returned back to the error screen which would likely confirm bad IC or installation.
Most components are fine, and the diode readings match with the other Switch.
After a close visual inspection, I had a feeling the IC was properly soldered on. I removed it anyway as you suggested. As expected, no power was drawn through USB (which makes sense with no M92 chip), but it wouldn’t work on a charged battery either.
So I’ve made Switch #2 worse somehow… Two things may have caused this:
I’m a moron and damaged something, somewhere, on the board. I’m going to check everything again.
When plugging in the USB charger, because the board wasn’t properly seated in the Switch case, it was inserted with an angle. Don’t know whether there could have been a surge in current, or something similar.
- So I’d start taking a close look in this area to begin with, though I could be wrong and it could be camera//lights playing tricks on me 
hard to say but as the damage at this inductor tracks with the damaged backlight ribbon (I imagine it sheared off on impact perhaps (?) ) then I’d say damage to the Max IC or it’s joints is the most likely candidate.
) you’d just destroy the SoC the moment you apply power.
things will start to take shape
hoefully this helps you out the next time you do a USB connector replacement 
