WD5000AAKS failure - what am I looking at?

[Wolvenmoon]

Hi all,

I've got a WD5000AAKS-00TMA0 PCB number 2061-701477-800 AB, one night I went to bed and the next day I woke up noticing my system wasn't responding to ping, drove over to get eyes on it, and found it ticking. Repeatedly.

It's visible to operating systems, ish. All it will do is tick when an OS tries to boot up. It's used on Linux OSes exclusively. It doesn't consistently show up in the BIOS. There were no pre-failure warning signs nor SMART issues.

I'm going to describe the picture because pics tend to disappear off of forums leaving the text remaining years later. The PCB itself has discoloration on many of the tinned pads, discoloration on several of the 20-pin connector's pads, discoloration around many of the Marvell chip's pins, and an odd yellowish sheen on some of the screw-mounts reminiscent of automotive-oil tained water. The Smooth chip (northeast corner of the board) has a discolored blotch visible when light's reflecting straight off of it.None of the discoloration comes off when a Q-tip with 91% isopropyl alcohol is used to test it.

There's pretty distinct discoloration in the grommet the likes of which make the discoloration in this video look tame, https://www.youtube.com/watch?v=rRO_eI5 ... u.be&t=555

There's no obviously dead chip, but it looks like my contacts may be oxidized as discussed in this thread: https://forum.hddguru.com/viewtopic.php?f=1&t=19077

I also noticed that the back side has oxidation similar to the pictures on https://community.wd.com/t/wd-hdd-pcb-o ... e-it/13414


Am I heading for a clean room recovery or a PCB swap? Should I try cleaning the contacts with something else before sending it off?

Attachments

Toasted PCB?

[fzabkar]

The HDA contacts look clean to me, but you could gently shine them up with a soft eraser.

Test the spindle motor windings for continuity. The phase-to-phase resistance should be twice the phase-to-common resistance (approx 2 ohms versus 1 ohm).

If you have a spare PCB, try to spin up the drive with it. This will help to narrow down the problem to the PCB or HDA/motor.

Your PCB has no chip at U12, so you will need to find someone to copy the embedded firmware to a donor PCB. I would contact hdd-parts.com (Canada) to see whether they offer such a service.

[Wolvenmoon]

The motor does appear to spin up - the drive does hum as expected, but it'd make sense to me as an electrical engineer that if the motor was messing up the resistances would get higher and there'd be excessive heat generated and such. What pads do I need to test on the motor spindle windings?

My assembled PCB looks like: http://images.techtree.com.s3-ap-southe ... igital.jpg

Which pad's common, which pads are the phases? I did poke around at it a bit, got 1.2-ish ohms on several readings, but I'm not up to date on the HDD anatomy.

I did take an eraser to the contacts. No change in symptoms.

[fzabkar]

Sorry, I assumed that the ticking sound you were describing may have been something like a stiction sound. Clearly that is not the case.

If the motor is spinning up, then you may have the "head mimic" fault described in the Youtube video, in which case a PCB swap would be worth trying.

[Wolvenmoon]

For posterity/future Googlers who'll run into this thread years down the line, do you know which pads are which off the top of your head?

Also, I went and looked things up.

It's definitely not a stiction issue, that has a very distinct sound:
https://www.youtube.com/watch?v=yRFvMRo0A8k
https://www.youtube.com/watch?v=7hSj_uE_BKM

This sounds exactly like:
https://www.youtube.com/watch?v=XwYSdD4fyRU
and https://youtu.be/vbq8vfs1wYY?t=242

So I think the armature head is definitely bouncing back and forth.


If this were your drive and you were looking at either sending the entire drive off to have a PCB swap and further diagnostics on it, or only sending the PCB to have firmware pulled and transferred but the cost was the same either way - what would you pick? (The service I've picked is Outsource Data Recovery via donordrives.com though I also have a ticket open with ACS data recovery).

[fzabkar]

See viewtopic.php?t=25524

I would send the PCB, without the drive, to a company I can trust.

[Wolvenmoon]

Annnnnnd full stop, my train just hit the brakes. Screeeeeeeeeeech!

Who's a company you trust? I've had a few instantly throw out $400 estimates as the bare minimum and I'd really prefer not to pay $400 for someone to flash a chip and image what may be an otherwise functional drive.

Is the reprogramming process going to destroy my original PCB?

[drHDD]

It should be problem with pcb. For this model it's first thing which you need to check. I could sell you board for $20 plus shipping and rom transfer included if you will send me your board.
If you can send drive to me, I can get your data and send it to you by internet. If it's problem with just pcb - $100. If it's heads - $200.

[Wolvenmoon]

Hello! I'm really excited to meet a forum member providing services!

Is it possible for me to pull the ROM data myself? I'm terrified that something will happen during shipping.

[day1data]

Hello! I'm really excited to meet a forum member providing services!

Is it possible for me to pull the ROM data myself? I'm terrified that something will happen during shipping.

On this model of drive the ROM data is embedded inside the Marvel CPU chip you see on the PCB.
So its not possible to extract the ROM data without professional data recovery tools. The ROM however can be re created from the SA (Service Area) contained on the platters.
As long as you are sending the drive to someone professional they should be able to re-create the ROM anyway.

[Wolvenmoon]

Oh darn. It sounds like it has to be opened up regardless. And with a dead PCB that sounds like it'll be a platter swap no matter what, if the Marvell chip is dead. Can you confirm this?

[pcimage]

Oh darn. It sounds like it has to be opened up regardless. And with a dead PCB that sounds like it'll be a platter swap no matter what, if the Marvell chip is dead. Can you confirm this?

No, that's not the case.

All that's needed is for the internal ROM to be read from your PCB (assuming it's working well enough for the ROM to be read, which given that the drive spins should be the case) and written to a known good working PCB of same P/N. Special tools are needed for this.

There is NO need to open the drive or "platter swap" to perform this operation, but if it turns out not to be the PCB then the drive WILL need to be opened to change the heads assembly, but definitely NOT a platter swap!

Good luck

[Wolvenmoon]

Thank you very much!

I'm not claiming expertise in dealing with specialized equipment, but this is stuff I'm curious about on a professional level.

I've not brought this up because I don't know what I'm doing, but I'm a computer science/electrical engineering double major who's done some amount of hardware hacking. I've got a JTAGulator out on my desk and such. I've actually attacked proprietary microcontrollers to try to dump firmware and attach debuggers.

Are we talking about stuff to hook up to a serial port or JTAG? What voltage? I've got four sacrificial WD drives that I'd like to practice on, because this stuff is cool to me. I might want to poke around a bit.

[drHDD]

May be this would help viewtopic.php?f=13&t=20324

[Wolvenmoon]

May be this would help viewtopic.php?f=13&t=20324

AWESOME! Thank you so much.

This is really straightforward. It just requires a steady hand - which due to injury I won't have for another 6-12 months.

Will it irritate DR companies if I've already had a PCB replacement made?

[lcoughey]

This is really straightforward. It just requires a steady hand - which due to injury I won't have for another 6-12 months.

Hope it gets better sooner than that.

Will it irritate DR companies if I've already had a PCB replacement made?

Not usually, as long as we are 100% sure we are getting a PCB with the original ROM on it. There is nothing worse than wasting time fighting with a drive and to find out that part of the reason for the issues is a mismatched ROM. So, be sure to supply both the original and new PCB, marking which one is which.