Samsung SSD 840 PRO no longer being recognized by bios

[xga]

Hi Guys,

Hope you can help me!

My main desktop PC has two Samsung SSD 840 PRO 256GB (MZ7PD256) running in RAID 0. Today I had a power failure whilst the computer was on and when the power was restored, the computer could no longer boot. Checking the BIOS, I noticed that only one of the 840 PRO's was being detected. I've removed the failed SSD from the computer and tested it with a USB to SATA adapter on another computer but it's still not being detected.

Apart from taking the SSD to a data recovery centre, do I have any other options I could try to get it working again. Is it possible that the firmware could have become corrupt and I could try flashing it similar to the technique in this post -https://forums.tomshardware.com/threads/samsung-850-pro-ssd-corrupted-firmware-fix-read-first.2193040/ ?

Thanks in advance for any assistance you can offer.

Cheers

[DR-Kiev]

If you wanna try that method, save byte-to-byte image first from donor ssd, just in case.

[xga]

If you wanna try that method, save byte-to-byte image first from donor ssd, just in case.

Thanks for your reply, DR-Kiev however, I wasn't able to get it to take the firmware when I hot swapped over to the faulty SSD.

Is there any easy way to determine if the SSD is powering on? I'm starting to think that some of the components might have been damaged either when the power went out or came back on. I assume that if I measure 5v around the controller IC then that's going to be the best way of telling?

[fzabkar]

FWIW, the voltage test points for the 840 Evo are in @sourcerer's "Missing Manual":

http://www2.futureware.at/~philipp/ssd/TheMissingManual.pdf

If you need specific help with your Pro SSD, please provide detailed photos of both sides of the PCB.

[xga]

FWIW, the voltage test points for the 840 Evo are in @sourcerer's "Missing Manual":

http://www2.futureware.at/~philipp/ssd/TheMissingManual.pdf

If you need specific help with your Pro SSD, please provide detailed photos of both sides of the PCB.

Thank you, fzabkar. I appreciate any advice you can suggest. Photos of both sides of the PCB are below.

[fzabkar]

The voltage test points are marked in the attachments.

Attachments

[outbreak198x]

Hello,

i have SM841 - same as 840pro - can't enter Safe Mode to read out with PC3000, and it stays BSY when entering normal Mode

so i measured the voltages named like in picture, and there are some with 0V - is it correct, or is there another pic where i can see what voltages there should be

V1 = 1,93
V2 = 0
V3 = 2,95
V4 = 1,84
V5 = 0
V6 = 1,16
V7 = 1,05
V8 = 0

[fzabkar]

Check the 3 ENable pins of the RT9991.

https://www.richtek.com/assets/product_file/RT9991/DS9991-02.pdf

I don't have a datasheet for the LM10515SQ, so I don't know what is going on there.

[outbreak198x]

EN1 0V
EN2+3 5V

[fzabkar]

Either the V2 output is shorted to ground, or the RT9991 converter is faulty.

I would also test for shorts at V5 and V8.

[outbreak198x]

hm - it looks like no shorts on all 3 0V voltges. always >30kOhm - so maybe it's really dead - ill check some things - maybe i'll find something out thx for now.

[fzabkar]

I would replace RT9991. Perhaps V5 and V8 are asleep, and V2 needs to come up before they can be woken up??

[outbreak198x]

I would replace RT9991. Perhaps V5 and V8 are asleep, and V2 needs to come up before they can be woken up??

Changed RT9991 - and it's online - thx

[ivo.andonov]

Just registered on the forum to say thank you and fill in this thread with some more info.

Have the same SSD model and board. Suddenly stopped No important data so it was regarded as an engineering challenge to see whether I can bring it back to life.

If was making strange noises from the inductors. No shorts however. V2, V3 and V1 were quite erratic. At some time they would stabilize and no noise, still the SSD would not get recognized. The RAM was getting hot. Turned out the RT9991 was faulty. It was trying to drive the memory at 3+ volts instead of 1.2V. Probably the RAM was latching and so the electronic fuse was doing its job by restarting the board.

I did not know what the actual V2 and V1 readings should be. Powered them (one at a time) externally by gradually increasing the voltage on them until getting 0.8V on the resistor divider feedback to the RT9991. Thus V2 should be 1.2V and V1 should be 1.9V. Removed the RT9991, then applied power to V2 and V2 and powered the SSD (USB-SATA to the PC USB port) and... a kind of magic The SSD is alive.

I do not have a spare RT9991 so that will be the end of live of this SSD. Just dumped the data to be sure it is otherwise working.

Thanks!

[ivo.andonov]

Just registered to the forum to say thanks and share a bit more information about this drive.

So I have one of these SSD 840 by Samsung. The same board as the one from pictures in this topic. Suddenly (but not surprisingly) it stopped working No critical data on it so I took it as an engineering challenge to see whether I can bring it back to life myself.

The V1, V2 and V3 voltages were quite erratic and some noticeable noise was coming out from the inductor(s). At some point the voltages stabilized but the drive did not appear in the OS. The RAM chip was getting hot. There were no shorts on the capacitors.

Though I consulted the other replies here, I did not get a definitive answer what the readings of these voltages are expected to be. So with the device disconnected I gradually applied voltage at V1 and V2 (one at a time) until getting 0.8V reference at the resistor dividers feeding the RT9991. Thus I got 1.2V at V2 and 1.9V at V1.
While having the SSD normally connected I measured V2 maxing at 3+ volts and as it turns they are applied to the RAM. No surprise it was getting hot and probably latching so that the electronic fuse was stepping into action.

So I removed the RT9991. Externally applied 1.2V at V2 and 1.9V at V1 and then connected the drive to the PC (USB-SATA bridge). And... oh, a kind of magic The drive emerged into the OS.

Just dumped the data for the sake of proving it is working. As I do not have any RT9991 lying around, that would be the end of life for this product.

For anyone else that might be running into this - 1.2V and 1.9V sink about 100mA during sequential reading. I limited the bench supply current to 200mA on both channels.

Thanks!

[ivo.andonov]

A short update...

As I hate throwing away hardware that can be fixed I looked around for other power supply options. So I got two (readily available in the town) LMZ21700SILR ($10 both) and pretty easily fitted them on the board. And as they use the same 0.8V ref, no need to fit other resistors. Now the drive is back in the game.
A couple of images of the result.
The 1.2V:

1.2V.jpg (41.9 KiB) Viewed 33178 times

And the 1.9V:

1.9V.jpg (54.2 KiB) Viewed 33178 times

[Zero Alpha]

Very nice work! Did you use an oscilloscope or just a multimeter?

[fzabkar]

Very nice work!

+1

[ivo.andonov]

Very nice work! Did you use an oscilloscope or just a multimeter?

Just a multimeter. No scope though I have several and it might have been interesting to see the bad RT9991 waveforms but I was not that interested at the time

Some more info about the SSD 840. It seems mine suffered from the slow read speed performance of old written data. I rewrote the image I took and the speed was "recovered". Some read results before and after:

Before rewriting

ssd840.1.png (11.17 KiB) Viewed 32473 times

After rewriting

ssd840.2.png (6.84 KiB) Viewed 32473 times

The end of "before" is high as this region was never written.
The tests were performed using the same USB3 SATA bridge, same PC and same USB port.

And this is interesting. The smartctl info:

Code:

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: FAILED!
Drive failure expected in less than 24 hours. SAVE ALL DATA.
See vendor-specific Attribute list for failed Attributes.


And this is the failing attrib:

Code:

184 End-to-End_Error        0x0033   091   091   097    Pre-fail  Always   FAILING_NOW 9

Still, after two weeks, the drive works fine