Samsung 860 PRO SSD No Power

[ddrecovery]

This SSD gets warm, but has no power going to the SMDs. The only power I can detect is at the resisters marked in yellow. Both these read 5v at both sides. The groups of caps marked in red all appear to be shorted. The chip in blue is at the back of the shorted caps. There are a couple of TVS style chips but there is no power to either side so presumably the issue is not there. I presume the whole PCB is getting warm because the power is getting pulled to ground? Let know if you want any chip markings.

Any help diagnosing where the issue lies would be appreciated.

Attachments

[fzabkar]

I would have though that those capacitors under the Power Management IC (PMIC) would be filtering the IC's +5V supply. Moreover, if there is not output from this IC (measure the voltages at the inductors), then the heat must be due to the "RB990" buck regulator nearby. I suspect that this regulator provides the Vcc for the NAND array.

(Edit: This IC requires +5V, so ??? )

I would examine the 10-pin and 8-pin ICs nearest the SATA connector. One is probably an e-fuse (most likely the 10-pin IC, eg STEF4S). Check the Vin and Vout.

PCB protection devices:
http://www.hddoracle.com/viewtopic.php?f=100&t=1615

BTW, it would nice to have the marking codes so that I can update my datasheet database.

[ddrecovery]

I would have though that those capacitors under the Power Management IC (PMIC) would be filtering the IC's +5V supply. Moreover, if there is not output from this IC (measure the voltages at the inductors), then the heat must be due to the "RB990" buck regulator nearby. I suspect that this regulator provides the Vcc for the NAND array.

(Edit: This IC requires +5V, so ??? )

I would examine the 10-pin and 8-pin ICs nearest the SATA connector. One is probably an e-fuse (most likely the 10-pin IC, eg STEF4S). Check the Vin and Vout.

PCB protection devices:
http://www.hddoracle.com/viewtopic.php?f=100&t=1615

BTW, it would nice to have the marking codes so that I can update my datasheet database.

As always thanks for the response.
No voltage at the inductors.

Here are the two IC markings:
8-pin - 13XH 818 A0E5 (no voltage in or out)
10-pin - (See pic) - I cannot find a spec sheet for this chip due to the strange numbering. Not sure if it is AtACS, or the 't' is some kind of specific electronic character. However looking at the inputs and the outputs traces it appears to be correct but that is a guess. I have put freeze spray on the PCB and this 10-pin chip does get extremely hot, so I am guessing this is the issue even though the input and output seems okay?

Attachments

[fzabkar]

The 10-pin chip appears to be an STEF4S e-fuse. Its output is at 0V, so either it is bad, or it is being shorted by something. AFAICT, the only chips requiring 5V would be the "RB990" and the PMIC. Also check the 5V filter capacitors.

TPS22990DMLR, Texas Instruments, 0.6V - 5.5V, 10A, 3.9mOhm Load Switch, marking RB990, WSON-10:
http://www.ti.com/lit/ds/symlink/tps22990.pdf

STEF4S, STMicroelectronics, electronic fuse for 3.3V and 5V lines, marking EF4S:
http://www.st.com/web/en/resource/technical/document/datasheet/DM00095426.pdf

Texas Instruments, 0.6V - 5.7V, 6A, 16mOhm Load Switch, marking 13XH, WSON-8:
http://www.ti.com/lit/ds/symlink/tps22975.pdf

[ddrecovery]

I have ordered a few replacement parts and will get back to you once I have tested.
Thanks for the help.

[fzabkar]

You should check if the load switch IC(s) are connected to the Vout pins of the e-fuse. If the PMIC capacitors are connected to the output of the load switch rather than to the e-fuse, then this would narrow down the culprit. Does the load switch heat up?

There is another 8-pin IC which appears to be a regulator of some kind. It could also be a culprit.

If I were doing this job, I would take some time to identify all the ICS and map out a block diagram. Then we wouldn't be stabbing in the dark.

[ddrecovery]

You should check if the load switch IC(s) are connected to the Vout pins of the e-fuse. If the PMIC capacitors are connected to the output of the load switch rather than to the e-fuse, then this would narrow down the culprit. Does the load switch heat up?

There is another 8-pin IC which appears to be a regulator of some kind. It could also be a culprit.

If I were doing this job, I would take some time to identify all the ICS and map out a block diagram. Then we wouldn't be stabbing in the dark.

I did do some testing after finding the 10-pin IC got very hot earlier. The outputs do go straight to ground and do connect to the caps under the PMIC (also grounded) so that is why I was going to replace it. The other two chips outputs are not grounded. It may not be as scientific a process as yours, but I think we have found the culprit.

[fzabkar]

My concern is that a capacitor may be responsible for the short rather than an IC.

The regulators and load switches could be wired up in either of the following ways:

e-fuse --X--> load switch --Y--> regulator IC

e-fuse --X--> regulator IC --Y--> load switch

You would need to confirm whether there is a short at point X or point Y. Do this with power off to prevent the load switch being from being switched on.

[ddrecovery]

My concern is that a capacitor may be responsible for the short rather than an IC.

The regulators and load switches could be wired up in either of the following ways:

e-fuse --X--> load switch --Y--> regulator IC

e-fuse --X--> regulator IC --Y--> load switch

You would need to confirm whether there is a short at point X or point Y. Do this with power off to prevent the load switch being from being switched on.

Would one cap being shorted cause all three outputs of the load switch to be shorted as well as the three lines under the PMIC?

[fzabkar]

Anything which is wired in parallel with a short will test like a short in-circuit. You would need to desolder each suspect, one by one, and retest for shorts. Sometimes you can detect the culprit with a thermal imager, or a finger test.

[ddrecovery]

Anything which is wired in parallel with a short will test like a short in-circuit. You would need to desolder each suspect, one by one, and retest for shorts. Sometimes you can detect the culprit with a thermal imager, or a finger test.

The only component which gets hot after power up is the IC, all the caps remain cold.

[fzabkar]

If you remove the e-fuse and the short goes away, then you could bridge the fuse (or you could replace it with a picofuse or something similar).