Recover Fried Hard Drive

LockHazer

Commendable
Jan 1, 2017
38
0
1,530
0
Hey all,

Big oopsie done today. Accidentally connected my pc’s hdd with a molex to sata power connector but wired the molex in the opposite direction so when I powered on my pc it fried the hdd.

Rest of Pc still functions but just wanted to retrieve the data on the hdd.

To solve this I’ve stripped a working hdd and took out the circuit. The problem though is that it is the exact same board, but the chips are slightly different. I know I have to switch the bios chips but the donor has 8mb cache whereas my fried one has 16mb. My original hdd also had a 7200rpm ratin gbut the donor is rated to 5900rpm. Would replacing the board then just limit the cache/rpm or not work at all? I can get a compatible board for $10 and switch the bios chip as an alternative.

My original fried hdd MODEL NO. Is ST3500418AS and the board I have is from a ST3500312CS. The board I can buy is ST35000413AS

Two photos of the fried board and donor board here:
https://imgur.com/gallery/w8SB8rv

Any help would be greatly appreciated!

Cheers
 
You can't switch the circuit board on a modern HDD likely.

You USED to be able to do that but modern HDD's have such fine tolerances for the read/write head that they do an auto-calibration on the first run and store the offsets in a chip which is soldered to the circuit board.

OLDER HDD's may work by switching circuit boards but I think they still need to be IDENTICAL.

I would bet my life savings you can't switch the boards for ANY drive if they are different RPMs no matter the age.
 

LockHazer

Commendable
Jan 1, 2017
38
0
1,530
0


How old is old? The hdd's are from 2012 and 2013. I know that the details are stored on the bios chip, the small 8 pin chip on the top left of circuit and I was prepared to de-solder and then re-solder that the the new board and so I believed that would work. On the $10 replacement it is the exact same board but just features 6.0gb/s instead of 3.0gb. Would this method work?
 

ex_bubblehead

Glorious
Moderator
The controller board must be an EXACT match. Close doesn't cut it. However, it's more than likely that your reverse bias fried much more than just the board. It likely fried both the spindle motor and heads. This is when you replace the drive and restore from your most recent backup (you DO have a current backup....Yes?). Your only other option is to break out the checkbook and hire a recovery service to try and recover your data. However, the more you mess with it, the less likely even a recovery service can retrieve anything useful.
 

LockHazer

Commendable
Jan 1, 2017
38
0
1,530
0


Well I did have a backup, but the hard drive was relatively new so hadn't done one since then unfortunately.

If I opened up the casing could I check for burnt spindles?

Can't be bothered with the expense of a recovery service.

Finally, the $10 replacement board has the exact same PCB number but the only difference is that it supports 6.0Gb not 3.0Gb like mine so it could be in the BIOS chip or an extra chip. Sure that wouldn't work if I transplanted the BIOS chip from my old board to it and use that circuit board on my old drive?
 

ex_bubblehead

Glorious
Moderator
1. The controllers MUST be an EXACT match in ALL respects. This includes chip revisions.

2. If you break the seal on the drive, and you do not have the correct tools or environment (class 100 clean room), your chances of even a partial recovery are indistinguishable from zero.

To give you an idea of the tolerances you're dealing with here:

If you were to scale things such that the flying height of the read/write head were 1 inch.

A single strand of hair at the same scale would be over 5ft in diameter. A "small" dust particle would be over 1ft in diameter.

Neither would fit between the head and the platter without bouncing the head off the platter surface and causing damage to both.
 

LockHazer

Commendable
Jan 1, 2017
38
0
1,530
0


Okay so I won't resort to cracking it open just yet then.

Now just having a hard time distinguishing between chips and board compatibility.

My Dead hdd PCB is a 100535704 REV C.
My Donor PCB is also a 100535704 REV C.

So these are the exact same labelled PCBs. I can read on the largest chip on the boards that they are both LSI chips but don't have any number that correlate to each other so seem different.

This would make sense as one was taken from a 6gb/s hdd and the other a 3gb/s hdd and this is the only difference between the 2 drives but still, exact same PCB labelling.

So, what you are saying is that I need a 100535704 REV C board but it has to be from my a hdd the same as my one and not a different hdd even though it has the same 100535704 REV C PCB?

On this replacement website: https://www.hddzone.com/st3500312cs-seagate-pcb-100535704-rev-b-p-805.html
It states that different board revisions ie. Rev A, B, C are compatible with each other. It also says that the same board can have different Main Controllers and that these are compatible.

Go to the bottom of the link to read.

Is this just false then?

Thanks again.
 

LockHazer

Commendable
Jan 1, 2017
38
0
1,530
0
Calm down mate I'm just trying to learn.

Since you seem to know what you are doing, can you offer an explanation as to why the replacement site said that they would be compatible?

Cheers
 

ex_bubblehead

Glorious
Moderator
Simple. For normal use they are (generally) interchangable. However, when talking data recovery an exact match must be used as even the slightest difference can introduce variations in timing (among other things) that will cause the process to fail. This is part of the reason that professional recovery costs what it does. They have the requisite knowledge and tools to do the job the right way.
 

LockHazer

Commendable
Jan 1, 2017
38
0
1,530
0
So would that mean that a blank drive would theoretically work fine with a transplanted board but if it had any data on it previously that would get corrupted?
 

LockHazer

Commendable
Jan 1, 2017
38
0
1,530
0


This link shows a photo of what my board looks like now. https://imgur.com/gallery/w8SB8rv

Sot it seems as if 2 inductors have melted together. Do i just cut these off? The FAQ article you linked states that I shouldn't replace the connection with a wire as it will be a short circuit on the diode so is it ok if I do that to the inductors?

"Warning: Do NOT replace a TVS diode with a piece of wire. By doing so, you will be replacing a short circuit with another short circuit"

The 5V diode doesn't look damaged but should I just remove it and leave the connections not touching?


 

ElectrO_90

Dignified
BANNED
Since you have screwed the hdd up, and you want to try and get the data back and right now you can't.
What have you to lose? Try it, let us all know how you get on.
You can either Pay someone a small fortune and get your data back or you can try and get it back for "free".
Ideally things need to be the same, but no one here actually knows, it's all guess work.

ex_bubblehead said 1. The controllers MUST be an EXACT match in ALL respects. This includes chip revisions.

This maybe so, but who knows. Unless ex_bubblehead is the designer at Seagate and knows this for certain it's still a guess.

The closer the numbers the higher chance of your drive will come back to life, unless of course, you fried the motor.
 

LockHazer

Commendable
Jan 1, 2017
38
0
1,530
0
@ElectrO_90, I will try fzabkar's method first as it seems the most logical and safe. It also shouldn't damage the parts needed for the other method if it does fail but I'll be sure to let you know.
 

LockHazer

Commendable
Jan 1, 2017
38
0
1,530
0


1. I'm not too handy in electronics so what should I be looking for on the multimeter, 0 or a positive reading? And would I be measuring full volts or millivolts or something else. If it is working, then do I leave it in? Would that not mean that it failed to protect the circuit as it should have blown to protect the rest?

2. The diode has a horizontal line on one end and a V so I believe that is the cathode. The 5v rails would just be pin 8 and 9 according to your pinout correct? So would just an ordinary insulated wire connecting to the sata pins already on the motherboard to the cathode work or do I need to bridge the gap where the inductors where? And thus is it safe to have both pin 8 and 9 going to the same cathode end?

If you can be bothered to make up a quick MS Paint sketch that would be much appreciated

Sorry for the hassling, hdd contained some windows boot material and such would like to recover it.

EDIT: Setting the multi meter to 20 V (DC), it has a V with flat line and 3 dots under it. Looked up and a I believe it means DC, I get no voltage through the diode, touching the red and black tips to the cathode and anode of the diode. Meter is working as it shows 1.5 Volts on a D battery.
 

LockHazer

Commendable
Jan 1, 2017
38
0
1,530
0
Actually did some study and found the proper way to test the diodes and what they do. Will multimeter test in morning but I believe it is shorted. If so, I just clip it out and then connect 2 wires to the cathode and the 8 and 9 Sata pins correct?

Should I cut the burnt inductors off as well as they have fused together under the heat? And then as you said fzabkar, clean the charred areas

Fingers crossed the motor isn't burnt...

Edit: tested the TVS on proper diode setting on multimeter, the burnt board gives a 002 reading both ways and the same diode on the donor board gives a reading of 532 one way and just 1 the other way. So seems like the diode is dead

When setting the metre to 200 ohms I get a solid 1 both ways on the good diode and then 1.6 both ways on the bad diode.
 

LockHazer

Commendable
Jan 1, 2017
38
0
1,530
0
Finally got to fixing the hdd yesterday. I de-soldered the 5v diode to pull it out of its socket and flicked off the melted inductors. Cleaned the inductor area with a q-tip and some PCB cleaner so most of the charring went away except for the physical melting of the PCB. Then I added a blob of solder to bridge the 7 and 8 +5v pins and soldered a wire from that blob to the cathode of the diode socket. Had to run the wire on the outside of the PCB so that it wouldn't crush components on the PCB when screwed back into the hdd.

Powered on the hdd in an enclosure, spun up and works perfectly now.

Thanks for the solution Fzabkar and everyone else's help.

Cheers
 

ASK THE COMMUNITY

TRENDING THREADS