WD BLACK SN850 NVMe SSD not recognised anymore

[mac84]

Hello everybody on the forum,
My WD 1TB SSD since wednesday suddenly stopped to be recognized by my Intel NUC on boot, (no bootable device found). I tried also on two slots of motherboard and external enclosure but problem persists. The contoller chip becomes hot after a minute or two, about 85C, while memory chips around 40C. Any components I can check for shorts or else with my multimeter to diagnose it? Thank You.
Model is WDS100T1X0E-00AFY0 DOM:08SEP2021

[fzabkar]

Measure the resistance between ground and each of the 4 inductors around the PMIC (U6).

[Lardman]

Is that damage, thermal paste , glue residue or just the light on the controller? pmic also looks a little discoloured.

[mac84]

I checked the resistances however readings are not very stable, they fluctuate a little (measured with UNI-T UT60S)
1: 828-838 Ω
2: 320-328 Ω
3: 4.6-6.3 Ω
4: 430-440 Ω
I see something like a solid semitransparent paste around PMIC and DRAM chips, I don't know what it is. However I have another identical SSD in desktop pc so i can take it out to make some comparisons, just need some time to remove graphics card.
@Lardman: The left portion of controller was covered with label

[Lardman]

The goop is conformal coating/underfill - makes things more difficult to work on the board.

You appear to have a short on L3, I'd lift l3 to see which side it's on, given the controller was getting hot you might want to check if the caps around it are on the same rail.

[mac84]

I checked the two capacitors right of L3, both shorted (5 ohms) the other big capacitors no less than 250 ohms. Could I replace those two or something else to check yet? How do I know specs of those caps? Probably the same as other couples next to inductors right?

[fzabkar]

Can you repeat those resistance measurements for your working SSD? Can you also measure the voltages at those points on your working SSD?

L3 seems to be associated with the Vcore supply for the controller. The resistance would normally measure around 100 ohms. As Lardman has pointed out, the discolouration on the controller would suggest that it may be shorted. As he suggested, measure the resistances of the capacitors around the controller to confirm whether they are bypassing the same faulty supply.

[mac84]

Unfortunately after removing label i discovered it has different PMIC and memory chips with differences as you can see... same model just manufactured 6 weeks later, so i don't think we can make comparisons in this case. PMIC has a mirror like surface very reflective, this is the reason is may appear discoloured. I cleaned it with isopropyl alcohol But personally i don't think is damaged, at least visually.

https://i.ibb.co/hWdH23F/IMG-20230605-165746839-HDR.jpg
https://i.ibb.co/WK3ZcdM/IMG-20230605-165916161-HDR.jpg
https://i.ibb.co/Hqkz8jR/IMG-20230605-165929595-HDR.jpg
https://i.ibb.co/hBbThrS/IMG-20230605-171517510-HDR.jpg

[fzabkar]

You can still measure the resistances and voltages. There is one additional inductor, which suggests an additional supply. Are the two controllers identical?

[mac84]

Measured the voltage values connected to PC and resistance unplugged (or I had to measure also connected?). One side of big capacitors is to ground. Found also other shorted caps on top and left of controller, some are small black SMDs so i suppose are inductors right?
Probably bad capacitors are on some rail, maybe even just one blown cap making others looking shorted.
Not being able to compare part number of controller of other ssd as I had to reinstall in only working computer, but I will check when possible

Attachments

[Lardman]

Is it drawing enough current to show heat? You can use freeze spay on the caps to see which might be problematic but this one has the feel of a controller problem. If you've not used it before it's worth practising on another drive with a known short so you can know what to expect.

[mac84]

I'll be damned! My Intel NUC is killing m.2 drives immediately, so it wasn't the WD black SSD which failed.
I put the other drive inside and same message "no bootable drives detected..."
I immediately realized where I am, removed and again it won't be recognized by any different PC.
I started to check visually for damages but nothing to the naked eye, the controller or other chips are not overheating like the first one, but i only had connected for one minute.
I checked for shorts and again one couple of big main caps

[mac84]

I checked for shorts and again one couple of big main caps near the inductor are beeping, as well as some small caps near the controller.
I don't think every cap is really shorted but maybe just one or two which make the entire rail look shorted. Am I wrong?
@Lardman do you think freeze spray is better than thermal camera to identify problematic component?
This was a really bad month, even an 18tb died after two days, it won't even spin anymore, at least i didn't have data on it yet, was preparing to copy from my old 8tb. Already received replacement in 5 days from WD but you can imagine how much trust now I have putting data on these...

[Lardman]

I checked for shorts and again one couple of big main caps near the inductor are beeping, as well as some small caps near the controller.

Same model drive ?

@Lardman do you think freeze spray is better than thermal camera to identify problematic component?

Freeze spray 10usd thermal camera 250usd, you can't really compare them. Although the MG freeze spray I use has now gone up here to 31 gbp a can

Anything on the rail shorted will leave the entire rail reading short - you can normally get an impression of what it is from how hard the short is. You can either use freeze spray, a thermal camera with good enough resolution or sometimes IPA depending on the heat being generated or you can start dividing down the rail to narrow down the location.

It doesn't matter what drive you are using if you have a backup. Ex-enterprise servers are cheap, even fully loaded with drives and there are plenty of open source NAS solutions.

[mac84]

I tried with the freeze spray first on second failed ssd and honestly i didn't even check visually for faulty component because immediately after plugging in the usb enclosure I heard the tone of connected hardware, It was acting OK, checked even smart diagnostics, no problems reported, working like should. Is now still running on main PC. Could the shock from cold spray have triggered some thermal protection or similar? Or maybe leaving it unplugged a couple of days some discharge of static electricity ?
I checked the capacitors are always shorted, so this is maybe wondering just me that it can run like that. I made a backup a couple of files, however the most valuable data is on the first failed ssd, on which I knew I couldn't replicate this magic.
I sprayed on the first ssd and the first component which becomes hot is the PMIC, then 1 sec after the controller starting from the center, then the heat spreads and the frost goes away, no capacitor seems to be the issue using this method. I tried some more times, once I heard the tone of connected hardware but nothing showing up.
On YT I stumbled upon this video https://www.youtube.com/watch?v=RUUI2WRp_cc which is about a model exactly like my first failed SSD, same PMIC. From the comments seem that especially those manufactured in the pandemic period used different chips maybe not the most reliable ones and many SN850 fail even more rapidly than mine, another possibility is that NUC 11th gen (exactly like mine) is stressing somehow the ssd.

[mac84]

Someone with PC-3000 knows if this nvme drive is supported? From what I read there is no support for WD NVMe drives yet.
Next step would be swapping controller from functioning drive? Or anything else before this kind of repair?

[Lardman]

Support for any nvme is patch . They Ace maintain a list of supported drives https://blog.acelab.eu.com/pc-3000-ssd- ... dated.html

If you have an electrical problem controller support isn't the issue.