What am I up against (clicking drive with amnesia)?

[Jidis]

Hi all!

I had a weird catastrophe knock out my main computer's drive a couple days ago and was wondering if anyone could give me pointers on how to proceed (or if there's no point in trying). I ordered the same model drive soon afterward thinking I could swap the electronics and see if the mechanical half was still functioning. Obviously that was a foolish assumption after reading some of the newbie info, but I had planned not to do anything until I did a bunch more reading anyhow. The drive is a 320 gig SATA WD Caviar SE16 model WD3200AAKS.

Oddly enough, the drive was killed when a CD with a small crack in it hit warp speed and shattered. While trying to force the tray open to clear it, I look up to notice a bluescreen had occurred. I did a power button shutdown after figuring I'd need to remove the DVD drive. From that point forward, the drive spins up, then just makes a light knocking sound a few times. It's totally invisible to the BIOS on two different machines. All I can figure is that the DVD drive slamming to a halt somehow produced a surge or something, or maybe something was shorted by fragments of the CD. It's weird and I really don't have a clue. AFAIK, any crash log dump functions were off, and it should have been sitting idle when it was shut down several minutes after the incident.

The clicking sounds similar to the WD "bad preamp click stop" audio file you guys have in the "Hard Drive Clicking Noise" folder in the files section. I can record it, but I'm trying not to even power the drive up anymore until I've read more and know how to safely proceed.

Some of the info I read mentions resistors and a couple active parts getting blown in these situations. For the record, I have a small bit of rework equipment, and would not object to comparing or swapping some parts when the other drive arrives if that sounds like a reasonable place to start, but I'm obviously a long ways from being able to work with the firmware or communicate with the drive. If that's something that needs to be done and I might be able to learn enough and can afford whatever tools I'd need to do it, I have no problem putting it aside until that point.

What do you guys recommend, and any advice would be greatly appreciated.

Thanks!

[Jidis]

Hi again (and sorry for all this text),

Well, the would-be donor drive showed up yesterday and it's probably the wrong PCB. It looks about the same from the bottom, but it's a 2060-771590-001 rev P2, where the dead drive needs a 2060-701537-002 rev P1 (or at least most of that number). If transferring the firmware is the only way to try a working PCB, I guess I'd have been dead there anyhow, since it sounds like on this generation of WDs it's all trapped in a giant BGA chip that I couldn't have worked with regardless.

I've included a picture where you can see a bit of charring if you guys know what areas a surge usually hits. I did notice some on the pin1 corner of the motor IC, which fzabkar pointed out as a 12v surge over at the HDD Oracle. Some of it didn't show up well in the pictures, but you can see it in the imprints in the foam. I don't know how much of it is just from normal long term heat though. That dark area in the center of the main processor for instance, just looks like some old foam residue, but the print in the foam looks horrible.

Is there any point in trying to check or swap any of the components which are usually damaged in these sort of events, or is a firmware transfer and full board swap what everyone typically does (or the safest solution)?

Assuming you guys are somehow reading and writing to that main chip rather than removing it, are there people who let you send them the two boards, or sell you a good one and perform the transfer for you?

If there's any other information, pictures, or sounds which might help, or anything I can do to narrow things down any further on my end, I'll gladly provide what I can.

*side note*
I know and have read that a lot of the newbie traffic here comes from people in a sudden panic over their recent data loss. I'm definitely there and know you guys probably get really tired of dealing with it, and I apologize for adding to it, but it's an indescribable sick feeling. I've also called myself every name in the book for dragging my feet so long on making a backup, and know that it's my own fault. Ironically, I'm the guy that tells other people that they need to keep their important data on more than one drive. I had also finally just built a new machine with a new SSD to replace all this and was only days away from copying the content to the new drive when this happened. Lastly, none of this DIY thing in my case is about money if that's a usual motive. I'd gladly pay to have someone qualified handle it, but the drive has too much personal data of my own and other people's for me to feel comfortable sending the whole thing off. If the PCB part can be done by someone, that would be great.

Attachments

[data-medics]

I'd say you've got about a 0.02% chance that the PCB has anything to do with anything here. No, that PCB isn't going to work. And, even if you did have the right PCB board number there's still only about a 1% chance that the ROM code stored on the PCB would be close enough to work with your drive. What you probably remember reading about was TVS diodes that sometimes blow on PCBs, however, if that were the case it wouldn't spin up and would short out (and turn off) your PC power supply.

The other 99.98% likelihood, is that the drive was running while you were messing around jostling things inside the computer and the motion caused the heads to crash against the platters. When that happened, it ruined the read/write heads. It's now clicking because it's trying to find the servo tracking tracks on the platters, failing to do so, sending a recalibration....click....attempt again to find servo...click.......attempt again to find servo.....click... and it probably eventually gives up and spins the platters back down. Hard drives can take a lot of movement when they're powered on, but less than a half-inch drop can ruin a desktop drive when it's powered on. They're designed to be fixed in place when operating and have virtually no tolerance to movement during use.

If the data is of value to you (worth $500+) then you should really consider professional recovery. If the heads are badly damaged, such as having missing sliders, then every time the drive is powered on it's sanding away the data more and more. I know you've got "personal" data on there, but let's be honest, so does every single other customer we ever deal with. Trust me, no one in this business cares enough or has enough time to be bothered looking through your pictures. Most of us try to keep our exposure to client data to an absolute minimum. We really don't care to know what sort of odd, kinky fetishes people are into.

[Jidis]

Data Medics,

That's sad to hear, as I don't think sending the whole thing off will end up being an option.

Do the burns you see in the foam or the board look insignificant enough to be normal and not from any sort of electrical event? I don't believe I did anything to jostle it enough after the CD blew up, unless that itself caused enough physical shock. I was also trying the paper clip eject hole thing for a bit around that time which may actually have shorted something. I don't think I was looking at the screen at the exact instant the machine froze and threw the bluescreen.

When it has trouble reading or finding the servo tracks like you say, do the electronics of the drive never end up identifying the device to the host at all?

Much Thanks!

[data-medics]

I've seen plenty of drives where it looked a little cooked like that and the PCB was fine. You might want to try cleaning those corroded contacts next to where it's labeled "J1" with a pencil eraser though, as that can sometimes cause such an issue.

Unfortunately for you, that PCB doesn't have an external ROM chip that can even be transferred to another replacement board. If it did, the chip would be located at those solder pads where it's labeled U12. Yours has the ROM code embedded in the main IC chip. So it'll need to either be read using expensive data recovery equipment or rebuilt from the drive's internal firmware (which also requires expensive data recovery software/equipment).

It's totally normal for a drive to not appear to the system with bad heads. Because the drive can't load it's own firmware, which is mostly stored on the platters, it's unable to boot itself up. Sometimes drives will appear a different way to the system, such as with a small capacity like 3.86Gb, but many drives just don't post at all if the heads are bad.

[Jidis]

Thanks again Data Medics,

Yeah, I noticed the pads on the far side of that connector looked a little singed. I may try to clean them, I'm just worried about even spinning the thing up again without knowing what might be going on mechanically.

Take Care

[fzabkar]

@Jidis, I have written an article which hopefully will help you to understand the nature of your problem.

The hard drive -- a computer-within-a-computer:
http://www.hddoracle.com/viewtopic.php?f=46&t=2600

As for the "singed" connector ...

Oxidisation on Western Digital PCBs:
http://www.hddoracle.com/viewtopic.php?f=86&t=649

[Jidis]

fzabkar,

Thanks, and thanks for the info in the catastrophic failures thread. I'll go read these other two.

Take Care

[Jidis]

fzabkar,

I didn't want to ask you in the thread and look like I was seeking a second opinion, but I don't really know who's who in these hard drive forums yet, and you seem to know a lot of WD-specific info.

Do you also feel that there isn't much point in trying to repair or replace the controller board from my particular drive? It seems coincidental that it's got some burn marks and such in the same areas you point out in the PCB failure thread if it were all just mechanical. Being as sending the full drive out would be sort of a dead end for me as it isn't an option, I'm almost tempted to look into getting a known good PCB up and running (with the transferred firmware) just as a last resort to see if it does any good, but I didn't get any info on whether or not people offer that service. I'm sort of desperate and trying not to think about how much info I may have lost.

Much thanks either way and sorry to bother you on it.

[fzabkar]

@Jidis, given the set of circumstances that you describe, it seems likely that the drive has a media or head fault, or both.

[Jidis]

Well, there's two votes for it then. That sucks.

Still can't believe I actually let this happen. I appreciate everybody's help on it anyhow.

[Jidis]

Much, much appreciated Spildit, for this and the info in the other threads!

It's a lot to take in, as all this stuff is new to me at this point. Some of it also sounds pretty scary, like it would be easy to screw up, but I do have quite a few "junk" drives here. I'll start getting a space and system set up to try stuff on. I'll try not to get my hopes up, but was considering still pursuing the PCB thing anyhow as it didn't seem too expensive, and I could at least say I tried what I could.

Regarding the shock- Yeah, I've bumped and kicked way too many of my machines (even that same one) with no serious consequences, but I guess you never know. Only other time I crashed heads was years ago when I knocked a running external off the top of a tower.

Again, much thanks on all the help!

[fzabkar]

@Jidis, I was originally thinking along the same lines as Spildit. Tornado drives suffer from "head mimic" faults where a bad MCU produces the same symptoms as bad heads. However, in the Tornado case the MCU is an 88i6745-TFJ1 or 88i6545-TFJ1 whereas your Pinclite's MCU is an 88i8845D-BHG2. That other thread should give you some hope, though.

Best of luck.

[fzabkar]

@Jidis, after you have safely backed up the embedded ROM, I wonder of you could measure the voltages in the following photoclips, namely V1, V2 and Vneg at the SMOOTH controller, and pins 1, 4, 6 and the read/write channel pins of the 20-pin HDA contacts. It would be interesting to compare these against a known good PCB.

Attachments

regs.jpg (130.39 KiB) Viewed 15403 times

[Jidis]

Thanks again fzabkar,

Will try. It looks like it might take me a bit to get a second system up though.

Dumb question, but does all that stuff act the same and have the same voltages present with the mechanical half detached?

Take Care

[fzabkar]

... does all that stuff act the same and have the same voltages present with the mechanical half detached?

The Vneg voltage will probably come up and then immediately switch off, but that's normal when the PCB doesn't see the preamp (on the headstack). Otherwise V1 and V2 should be at their normal voltages. The read channel pins will not be at their usual level because they are driven by the preamp, but a comparison with a donor PCB may still be helpful. If you wish to measure the read channel voltages when the drive is running, then these are present at test points TV9 and TV10 on the underside of the PCB.

[Jidis]

Spildit- Are the board failures you guys are referring to typically the result of some sort of event, like in my case, or do they just go out randomly on their own?

Thanks!

<EDIT> Also, thanks for the voltage details fzabkar. Just saw the reply.

[fzabkar]

If there had been an overvoltage event, then the TVS diodes (D4 or D3) would have been shorted.

TVS diode FAQ:
http://www.hddoracle.com/viewtopic.php?f=100&t=86

[Jidis]

fzabkar,

Thanks. I did check that when I looked around and it wasn't shorted. I'm just sort of worried that since my problems all occurred with that CD explosion,etc., it would be a bit too much coincidence to be one of the usual PCB failures. Still plan to follow that anyhow just to be sure.

Take Care

[fzabkar]

It could be a red herring, but the hot spot on the MCU is tantalisingly close to the R/W channel pins. It would be interesting to see whether working HDDs have the same hot spot in the foam. If so, then this might be indicative of thermal problems that result in progressive, rather than sudden, failure.