Black screen after closing switch with HwFly RP2040

Nintendo Questions Nintendo Switch
81

Haha, nope again, just stating it in general. I don’t recall if I touched on it here or in the other thread, but I mentioned this is probably going to become a bigger issue for others in the weeks or months to come, which unfortunately seems to be the case :frowning: and unfortunately, people with not enough experience too it sems :cry: I’ve got no issues helping anyone who is willing to learn :+1:

I don’t have a Mariko rev out of assembly to test but using another topic as reference this seems fine

Good, secondary CPU rail fine :+1:

Fine afair, don’t remember what this rail/line is

All one and the same rail in question (boot CPU rail) - all fine for a mariko rev :+1:

One and the same, fine

perfect :+1:

Damn, almost had all of them :frowning: Second Ram rail (but of course also has SoC I/O implicated) , this is such a shame and I’m not even sure the cause. Pop your meter in continuity and find the caps (possibly plural, I forget how many) on the SoC which correlates with this rail (red) and take an extremely close look at them for any junk around them or solder spatter, or signs of heat etc (I’m really clutching at straws for you here) I might even be tempted to knock these caps off to take them out of the equation (only if your comfortable and sure you aren’t going to make the problem worse)

I mean, failing this, I guess you could pull the main PMIC and see if this clears, then after pull Ram… but I expect it’s going to be the SoC in this case. given your other rail measurements it’s highly unlikely that the EMMC would be causing this rail (in red) to be getting pulled down as a by-product as the EMMC’s two rails are measuring perfectly fine (though might also be worth doing just to rule it out, ie. disconnect the EMMC then measure red again)

Perfect :+1:

All but one :frowning:

I suspect you have two issues, I suspect your EMMC data is corrupt, and I think you have hardware issues. which one caused which fault is hard to say, but you know my feelings on the modchip and the voltage it exposes the switch board to, which again unfortunately tracks with what your measurements I’m afraid

82

pretty much those 00 data is not important since emmchaccgen does not gen empty data but hekate or mmcblkNX dump entire boot0/1 partition as image so that 4MB with empty data at the end IMO.

I also see this kind of ANSI note on a working Switch lite boot0 dump from Hekate.
some kind of log maybe

83

is this only on boards which have had the modchip installed, as then the error kind of makes sense given the modifications to boot paritions

84

YES, but let me quickly dump one clean v2 boot0 straightly from mmcblkNX reader

85

Oooh no, this is so sad!
By removing the EMMC, the measurement is still the same.

Could you highlight with a circle the area where I should look?

photo_2023-09-20_19-52-52
photo_2023-09-20_19-52-521079×692 331 KB

photo_2023-09-20_19-52-58
photo_2023-09-20_19-52-58912×600 292 KB

By pulling the main PMIC and Ram, do you mean by desoldering them? Would a reflow change something?

86

Cannot edit the message anymore but I think you mean the surrounding caps, the ones closer to the IC, right?

87

It’s the caps on the SoC (the area you’ve soldered the modchip flex too, one or more of those caps correlate with the rail red in question) you should also remove that flex btw and take the measurement on red again (just in case) assuming no change, then to find the caps I’m talking about, you’ve got two options, 1: black probe on ground, then red probe at the caps, look for the one/s that’s giving you the 9 ohm reading, option 2: put your meter in continuity, one probe on red (the the inductor by the PMIC), and the other at the caps and look for a virtually 0 ohm continuous path to the caps in question, either / or

yep, the Ram and PMIC would be the only other two major IC’s which have this rail in common (other than the SoC) - I find it largely unlikely it’s going to be some random passive on this rail at fault…

Nope, if anything, you’ll probably only change something for the worse

Hopefully you can tell from what I wrote above, it will be one or more caps on the SoC wafer. If your still having trouble then I’ll highlight, though you’ll still have to narrow down with your meter as I don’t recall them off the top of my head

88

CONFIRM clean boot0 has no such ANSI log at 003F7500 address. That’s said those data belongs to boot0 written from modchip

1 Like
89

Interesting, thanks!

90

I’ve desoldered and clean up remain of flux etc… The reading on read is still the same. There was small amount of thermal paste around caps on right side.

I think I’ve found the caps related to the RED one, you can see from the picture
WIN_20230920_20_51_03_Pro

I think they are, as sometimes the reading for RED changes a bit (9.5 ohms for example) and they have it too. Does it sounds legit?
I have no idea why they are impacted, as I never touched them.

You may notice the cap which is the left one to be soldered on modchip flex is not in a great state, but still gives me same readings as the past.

Let me attach some pictures. I am sorry if the quality is not great, maybe with natural light will be better!

WIN_20230920_22_01_04_Pro
WIN_20230920_22_04_02_Pro
WIN_20230920_22_04_13_Pro
WIN_20230920_22_04_24_Pro

Also my readings on the bottom ones are:
WIN_20230920_22_04_02_Pro

Black left 0.3 Ω, right 14.1 Ω

White top 0.3 Ω, bottom 13.5 Ω
Grey top 0.3 Ω, bottom Ω 13.5 Ω
Red top 0.3 Ω, bottom Ω 13.5 Ω
Orange top 59.8 Ω, bottom Ω 0.3 Ω
Yellow top 59.8 Ω, bottom Ω 0.3 Ω
Green top 59.8 Ω, bottom Ω 0.3 Ω
Light cyano top 59.8 Ω, bottom Ω 0.3 Ω

Blue/ dark cyano left 13.8 Ω, right Ω 0.3 Ω
Pink left 0.4 Ω, right Ω 6.75 kΩ

91

Yeah. What might be worthwhile doing is, checking to see which area / side represents the lowest, what I mean is, maintain a consistent ground throughout (ground pad at the USB for example, and I mean pad here, not the USB steel) and with your other probe measure the inductor in question back near the main PMIC, note the reading to the last decimal, maintain your ground and then take your red probe and measure any of those caps you determined are on this same rail and then note the reading. If it was lowest at the caps on the SoC wafer then most likely the SoC is responsible (your then welcome to try pulling those caps off as a last ditch hope :slight_smile: ) if the reading is lower at the inductor, then the problem is most likely the PMIC.

Worth noting though, there may be one other location worth checking the above too, the caps running down the side of the Ram IC’s, one of them may be on this rail too. If you identify any, you can do the same thing as above, if it’s lowest there, then you know it’s most likely Ram at fault :+1:

Don’t worry about these :slight_smile: they are just basically re-iterating the rails you checked earlier at the much more convenient locations :smiley:

92

Also @minimanimo after you’ve checked those measurments mentioned previous. Can you let me know about the SoC die in your pics, is it just a trick of the light (?) or is there solder on the die, or has liquid metal been used? or is this just some shiny thermal paste? Just asking because it looks odd, but like I say, could just be a trick of the light :slight_smile:

93

Hi @Severence,
I am going to try and check these.

I think is a ram fault: I forgot to mention, but when the modchip was booting (post #68) I’ve tried to run multiple payloads. As suggested, I’ve tried to run usm-loader and is the only payload able too boot and get something on screen. This usm-loader is a piece of code that uses IRam (integrated in the Tegra?). So, if is the only payload to load, then is possible to assume there is an issue with the Ram.
I think our measurements also confirm this. I’ll look for some shorts, but I’m not exactly sure where to look.
Do you think a problem with the ram can be solved?

I think, and hope, it is only the effect of direct light. I took the photos shortly after removing the modchip and cleaning everything with alcohol, so I think this is it

With less light:
WIN_20230926_19_11_43_Pro

94

Quite possibly, though, note there is a fair amound of Ram I/O SoC side, so don’t get your hopes up :slight_smile:

Not really sure what you mean, We have verified one of your Ram rails is good and we’ve verified your other Ram rail is not (ie. the one in red, approx 9 ohm) There’s nothing else to really be checked, it’s either gonna be the PMIC, the Ram or the SoC.

Given your UMS payload findings. Check those caps running down the side of ram and see if any measure approximately 9 ohms to ground using my prior instructions (possibly slightly less ie. 8.8 ohms for example) , if one of those caps is on that rail. and if it’s reading lowest in that area vs at the inductor (red) or the caps on the SoC wafer then you can be fairly confident it’s the Ram at fault :slight_smile: (which is good news)

Yep all good :+1: was just a trick of the light :slight_smile:

Also, refrain from connecting battery / powering up the board until the fault is identified / resolved to prevent further/secondary issues

95

I’ve tested some caps near the M92 IC and actually their resistance to ground is very big. I have measured from a minimum of 3 kΩ to over 100 kΩ. No one is in the scale of Ω, but I’ve tested a few of them, only the ones marked with a color on this picture:

WIN_20230926_19_40_26_Pro

Did I miss any important ones?

96

Bud, We’ve checked all your primary rails at this point, We’ve found your problem, you have a partial short to ground on your secondary ram rail. All we need to do now is narrow down the culprit, either the PMIC, Ram or SoC (as touched on earlier) - All your doing now is chasing your own tail :stuck_out_tongue: M92 IC is USB power management, not relevant in this scenario :+1:

We wanna know if any of these caps (highlited in red)
866961ee8d6100ae360035a7d8f002660477c830
Are on your secondary ram rail, and if yes, which read lowest relative to ground in comparison to the other locations we talked about :slight_smile:

97

Ahahahhahah I’m so sorry, I’m confused and unfortunately English is not my main language (I think you noticed from previous posts).

I hadn’t really noticed them. What is the plan to reach them? Impossible to physically touch them with the current probe, I would have to somehow remove the metal frame or find the connections of those capacitors onto B-side (assuming they go there. Do they go there?).

Thanks, as usual, for the time you are dedicating to me and the patience :grin:

98

no worries :slight_smile:

Not sure, I have a variety of different meter probes and they can all get in there to measure (some at a slight angle) - so you should be able too (meter probes diameter is pretty much standard give or take a hair)

No, don’t do this

This defeats the purpose… we are playing the old childhood game (I forget the name) but you know, you hide the thing/object and have someone else try and find it while telling them, freezing, cold, warm, hot, red hot etc until the item is found (think of the inductor marked red as being one location, the caps on the SoC wafer as another, and the caps down the side of the Ram as another). The reason we are checking at the locations in question is because they are closest to the IC’s in question (and thus can give us a very good indicator of what the culprit is) … If you go trying to find an alternative location to measure then your no longer getting a good indicator and are defeating the point of the “game”.(I’m trying to definitively narrow down the culprit to a high degree so you don’t needlessly have to pull things off the board)

Hopefully that makes sense

If you still don’t think you can get your tips in, then just find some pointy tip extenders on ebay, buy some finer probes, or just grind some probes down yourself :slight_smile:

Hey, no worries :+1:

99

Oh yes, make sense!

Looking closer with the microscope, actually with a thin probe I should be able to do reach them. Unfortunately, my probe measures 1.5 mm in diameter. No matter, it is a good opportunity to buy smaller probes.

Suggestion on the size? I found a couple that measure 0.2 - 0.3 mm at the tip, that should be enough!

100

:thinking: I’m not really sure tbh, I always grind the tips down myself when I get new probes as the upper end brand name ones seem indended for electricians (as opposed for us) and the tips are always a bit “blunt” , so I just twirl them on a sanding disc or diamond wheel to get a keener point on them.

There is some you can get on eBay from China and they are scary pointy (you could sew with them… you wouldn’t wanna drop them on your feet :wink: ) but unfortunately these paticular probes, the leads on them are junk, conductor seems to be aliminium core and it’s only a few strands (so ups the lead resistance)