ST2000DM001 damaged electronics

[fzabkar]

Nice work!

It appears that the short was within the PWM controller, and that the MCU is probably OK. I would leave the voltage at whatever is set by the two feedback resistors.

The ROM is OK. If it wasn't, then the drive wouldn't spin up. AIUI, the MCU computes the checksums of the various sections of the ROM during the POST.

IIUC, the terminal output suggests that the drive has some kind of internal fault.

As for the corrupt characters, I suspect that may be due to a slight mismatch between the TTL levels of your serial adapter and the drive's I/O voltage (1.8V).

[Jano952]

Well, I identified these corrupt characters, it was just problem between ground of HDD and FT230X (FT was connected to notebook)

It seems to me that this HDD had similiar errors, I ll check what they did
viewtopic.php?f=1&t=27513

[Jano952]

Another log, googling did not help

Ctrl+Z does not work, it does nothing, and I am not sure which are Read and which are Write channel pins (what if i short write channel pins ?)

Code:

Boot 0x40M

Spin Up
TCC-0029[0x000065B4][0x00006A20][0x00006E8C]

Trans.


Rst 0x40M

MC Internal LPC Process

Spin Up

TCC-0029
SIM Error 1002 LBA 0000000000038C07 FD FCFFF3FF
RW Error C44400CD
SIM Error 1002 LBA 0000000000038D33 FD FCFFF3FF
RW Error C44400CD
SIM Error 1002 LBA 0000000000038EF3 FD FCFFF3FF
RW Error C44400CD
SIM Error 1002 LBA 0000000000038C07 FD FCFFF3FF
RW Error C44400CD

(P) SATA Reset


SIM Error 1002 LBA 0000000000034C5C FD FC37830A
RW Error C44400CD PASS
SIM Error 1002 LBA 0000000000039BF8 FD FCFFF3FF
RW Error C44400CD
SIM Error 1002 LBA 0000000000034C5C FD FC37830A
RW Error C44400CDWriteDCOInfo FAILED during Initialization
SIM Error 1002 LBA 0000000000039C14 FD FCFFF3FF
RW Error C44400CD
SIM Error 1002 LBA 0000000000039C16 FD FC37C137
RW Error C44400CD
SIM Error 1002 LBA 0000000000039BFA FD FCFFF3FF
RW Error C44400CD

MCMainPOR: Start:

Check MCMT Version: Current

MCMainPOR: Non-Init Case

Reconstruction: MCMT Reconstruction Start

  Max number of MC segments 0A61

Nonvolatile MCMT sequence number 005BD94C

[RSRS] 079F

Reconstruction: Completed 1:

[MCMTWS]
SIM Error 1002 LBA 000000000003B097 FD FCFFF3FF
RW Error C44400CD

WriteMCTPrimary: EXCEPTION: MCTP Write Fail

Reconstruction: EXCEPTION: MCMTP Write Fail

Reconstruction Fail: Burly Case

MCMainPOR: MCTStateFlags 0000012A  MCStateFlags 000000C1

MCMainPOR: EXCEPTION: POR Failed General

MCMainPOR: Feature Disabled... PASS
SIM Error 1002 LBA 0000000000039BF8 FD FCFFF3FF
RW Error C44400CD
SIM Error 1002 LBA 0000000000034C5C FD FC37830A
RW Error C44400CDWriteDCOInfo FAILED during Initialization[0000000000087B33]v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}

Starting LBA of RW Request=0000000000087B33  Length=00000180

ProcessRWError -Write-  at LBA 0000000000087B33  Sense Code=C44400CD[000000001D2BFDF7]v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}[000000001D2BFDF8]v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}[000000001D2BFDF9]v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}[000000001D2BFDFA]v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}[000000001D2BFDFB]v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}[000000001D2BFDFC]v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}[000000001D2BFDFD]v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}[000000001D2BFDFE]v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}[000000001D2BFDFF]v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}[000000001D2BFE00]v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}v{00000020}... continues to write this

I supply that 1.8V always, might that be problem ? LX7157B had Enable pin, so it might want to stop supplying power and powering it again but it just cant because I bypassed that, I ll try to implement it and I ll see.

[BlackST]

If there was actually WRITING , it's big trouble.
The problem 70/90% is now ALSO internal.

[Jano952]

I hope these pins were really Read (two closest pins to MCU, they are differential pair and had 0,3V while PCB was disconected from HDD, two pins next to them had around 1.5V so I suspected they were Write).
After it boot up i shorted it, terminal hanged and disk clicked, after i removed short it just continued ...

But I am curious, If I bought new PCB before playing with original one, would it work ? I bet that no and I think I saved 50 Dollars

If it is really destroyed, well, i would be missing few C#, Asm projects and few photos and tons of downloadable sh*t .

[BlackST]

You should have asked yourself question BEFORE tinkering, you and only you know what you did. Forget data and the disk and carry on then , as data is not important.

[fzabkar]

But I am curious, If I bought new PCB before playing with original one, would it work ? I bet that no and I think I saved 50 Dollars

You said in your opening post that you shorted the Molex pins together. This would suggest that +12V was applied to the +5V input. If so, then I find it surprising that the 5V TVS diode and its associated zero-ohm resistors survived. Moreover, it is also surprising that the preamp withstood the surge. In short, ISTM that your repairs were OK, and that any internal faults are not due to anything you have done since the surge.

[Jano952]

I was surprised, that nothing else was damaged. PSU probably turned off before voltage reached dangerous values and maybe it was just coincidence... HDD had 2 years and 2 months so warranty was already gone ( Another coincidence ? ).

BlackST, it was not better before... At first disk just clicked, with servo FAILs, and now it aleast do something

Before I make 7200rpm fan from it, I ll try everything I find on net, miracles do happen

[Jano952]

Just to finish it, I removed cover of HDD (in dirty enviroment, but it was already crap, wasnt it ? ), saw heads arent parked, so i spinned platters and parked them manually. After reassembling it clicked few times, terminal show servo fail, but after 5-8 clicks clicking stopped and drive continues doing log above and then something like "reading ?" (like if it were copying sectors from one place to another). Never mind, this is the end for now, dont have tools, dont have experience and years to spend with HDDs Once I ll get bored, I will read again all info and maybe I ll fix it

[BlackST]

HOW do you know about SERVO fail without tracing SERVO ?

You can have multiple symptoms from different reason and viceversa even from only rom problems (but it spins...)

[Jano952]

HOLY SHIT I ACCidentaly removed LM338 feedback divider, voltage fell to 1,25V and disk booted up !!! IT WORKS, BACKUPING LIKE CRAZY

I would write more, but i cant i have to backup most important data, then maybe do image

[fzabkar]

Congratulations!

After you recover your data, maybe you should re-measure those original feedback resistors out-of-circuit. Could that be where you went wrong???

As for why the high Vcore voltage messed up the drive's behaviour, could it be that the read channel section of the MCU is sensitive to variations in this supply? In any case it's a very interesting result.

[Jano952]

Thanks, as far as i remember, it was 40k and 40k, LX has feedback reference is 0.8V, so thats 1,6V... it did not work at 1,6V. Maybe i have accidentaly damaged traces from preamp (I have no idea )
I am amazed that after i opened disk in such a dirty enviroment ( i saw pieces of dust on platters ) it still works well, It even copies at full speed almost 100MB/s !

Here is pic of status in Crystal disk info it has few uncorectble sectors (because of dust). Ignore that BAD HDD, it is just for trash data and it actualy works for 2 years since it went bad

I backed up everything I knew that I want and If i forget something, I ll just power up it again

Anyway, thank you all, you helped me very much, especially you fzabkar

Attachments

[Jano952]

....maybe you should re-measure those original feedback resistors out-of-circuit.

Yes i might desolder them, they are now useless with external power, there was probably paralel resistance, so one resistor seemed lower than it should ...

DIY fix LM338+few resistors and capacitors = cost 2 €

[fzabkar]

Code:

Vcore o--- Rfb1 ---+----o Vfeedback (0.8V)
      |            |
      |            |
    Rload        Rfb2
      |            |
     _|_          _|_
      =            =

If Rload = 0 (as it was), then both resistors would measure the same in-circuit.

[HaQue]

glad you got your data back.
but how do you "accidentaly" remove a LM338 feedback divider?

[Jano952]

Oh god, I have completely forgot that ... I feel like an idiot now. It is miracle MCU survived 1.8V. Well, nobody came smart and experienced I ll desolder Rfb1 and measure it tomorrow.

HDD could work like new even without opening it, but it probably wont because i forgot how to measure resistors and opened it, but who would use repaired disk for important data ?
Never mind, data is saved and I ll probably buy new HDDs to RAID1 and maybe I ll use this for unimportant data, I ll just have to add some 1,2V regulator which will fit to it (LM338 is too big and and eats too much power).

HaQue: I went eat and forgot to disconect HDD from power, when I came back I thought I turned it off (I did not check if it was really disconected). I tried to use different feedback resistors (with different voltages i had different logs), when i removed it, feedback pin was left grounded by second feedback resistor and that caused voltage fell to 1.25V, terminal started writing new messages and windows wrote new hardware detected... Now that was luck !

Here is one photo LM + Flash programmer (It looks weird, but it works). Brown wire is 1.25V, ground is through PC PSU, I cool it with PC fan.

And one screen from crystaldiskmark

It acts like if I never opened it in dirty enviroment (only these few uncorectable sectors in SMART).

[fzabkar]

Ironically, despite all the "bad" things you've done to your drive, its Read Error Rate has improved from 105 to 109.

The Total LBAs Read has increased by 36GiB in the last two CrystlaDiskInfo screenshots, yet the raw error counts have remained the same, as you have said.

(0x8624ee0325 - 0x8620645a23) x 512 bytes = 36.3 GiB

ISTM that you should buy a lottery ticket.

BTW, you might be able to cannibalise a cheap 1V SMPS from an old socket 7 motherboard (by cutting it out with a hacksaw), or you could adapt a 12V car charger from an old mobile phone. Some were based around an MC34063.

[Jano952]

What is the max current MCU can take ? I have found old, completely disassembled HDD ST3500820AS (I have only PCB and body with motor, it was friend´s disk which stopped working long time ago and he wanted to see what is inside, so we disassembled it and because it spinned, we made grinder from it ), it has small ST1S06A ICs, so I might create small PCB and use them. Datasheet says it can supply 1.5A (LX could 3A) but I think that MCU wouldnt eat more than 1A, because these coils on my PCB seems to be too small to be able to transfer more than 1A. I have one Socket7 MB but I dont want to destroy it . Using these small ICs would be nice small workaround, wouldn't it ?

BTW I am going to buy one lottery ticket today

[fzabkar]

You could insert an 0.22 or 0.47 ohm resistor in the supply line to the LM338 and monitor its voltage drop. That should tell you the MCU's current draw. I would do this when the MCU is reading or writing.

Here is a datasheet for a typical coil:

ELL5PS / ELL5RS-A, Panasonic, SMD Choke Coils:
http://industrial.panasonic.com/www-dat ... 000CE7.pdf