Sign in with
Sign up | Sign in
Your question
Solved

General and specific questions about pcb swap

Last response: in Storage
Share
August 22, 2012 4:47:07 AM

Hello all,

I have some failed hard drives that I am thinking about trying to get data off of. I have spent the last week reading up on replacing the printed circuit board and that sounds like the most viable option. I was wondering about your opinions on the process.

One drive is an external Hitachi 640GB drive manufactured around 2009. About one week after its one-year warranty expired, without any physical trauma or electrical mishaps, it suddenly stopped working at all. (It was a hot northeast summer, but nothing out of the ordinary.) Absolutely no sound comes from the device as far as I can tell. You'd think it wasn't drawing any power it's so quiet. But I ripped off the external case it shipped it and connected it to a functioning usb drive bay and the symptoms are the same.

The other drive is an internal WD Raptor from around 2006. I found a really cheap dell system on craigslist a couple months ago and threw the WD in there. It had a drive bay at the very bottom of the case with a nice plastic bracket that didn't even require screws. The other day I wake the system up from sleep and Windows is frozen (not a blue screen but a weird kind of frozen I don't think I've seen since the older 90s windows, the slightly out of focus cement solid frozen) and I hear what I would soon learn is the WD "click of death". I pull the drive out from the bracket at the bottom of the case to make sure that's where the sound is coming from. That would be when I noticed the label on the drive (ont he side I had right up against the floor) saying to keep it clear of any obstructions to permit airflow. So that was probably my fault.

Can someone here identify the problem based on the type of the click, the way some car mechanics can do with car sounds, or does anyone know of a reference? Because I've heard a few different kinds of clicks. In this case, it clicks 4 or 5 times, stops, and you can hear what sounds like a motor die down, then it whirs/starts up and the process repeats.

In these two cases - the drive that simply doesn't react at all and the drive with clicks - does anyone know if a pcb replacement is more or less likely to succeed?

I am curious about this procedure's success rate. For such a commonly encountered problem as dying hard drives I am surprised how little straight information there is online. We all know that a data recovery specialist will say it's hopeless, you'll screw it up even more, etc. But who would be silly enough to take them at their word for it? Ask an antivirus vendor how to clean your system and they'll tell you how you'll could your identity stolen and and could end up and jail unless you pay them hundreds of dollars. Well I've worked in IT too long to spot an industry job. It might cost pros a bit to find a replacement drive, and I guess they have "clean room" overheard, but do you really think they're not going to jack up the price to fleece a group as desperate as their customers - people who've lost tons of memories, their work, etc?

On the other hand I wouldn't exactly trust any of these online PCB vendors. They have no return policies (which is understandable) and most of their customers probably aren't repeat customers (except for the very unlucky few). So even though they make the process out to be simple enough, I have no idea if they're trustworthy.

Does anyone unaffiliated have a sense of the success rate of PCB swaps in situations like the clicking drive (or maybe for a subset of the clicks)? How come this is such an unknown (besides interested parties running up hysteria all over the internet)? So you swap with an identical model, and if there's adaptives, run down to the electronics shop and have them swap the ROM chip. What's the big deal?

Thanks!
a c 311 G Storage
August 23, 2012 10:20:48 AM

A no-sound, no-spin symptom usually indicates a failed PCB.

Clicking usually indicates head or media problems, but in certain cases it may indicate a read channel problem on the PCB. For example, certain WD models had "head mimic" faults due to a problem in the Marvell MCU (88i6745-TFJ1).

For information on board swaps, see the following:
http://www.donordrives.com/blog/pcbswapguide
http://www.donordrives.com/hard-drive-pcb-donor-swap-ma...

For those boards where the "adaptives" are stored within the MCU rather than in a discrete serial flash memory IC, the above vendor provides a free PCB Adaptation service:
http://www.donordrives.com/services
m
0
l
August 23, 2012 4:53:03 PM

Many thanks for the links and the advice. The information was exactly what I was looking for.

I will definitely swap the PCB on the no-sound, no-spin drive.

I guess the clicking drive is unlikely to be fixed by swapping the PCB. It is a WD but there is a different Marvell microcontroller than the problematic one you mention (this has a Marvell 88i8030). But I'll try it since I already have a donor drive so I might as well. Assuming that doesn't work I'll attempt to swap the head assembly. This drive has two platters and I understand a diy platter swap with multiple platters maintaining alignment has nil chance of success whereas a head assembly swap is a shade above nil. Also I am not completely convinced of the absolute necessity of a clean room, I remember a few times when I was a kid I would open those old 40MB or 100MB drives and quickly peek in there and close them back up and didn't notice any issues.

Also, fzabkar, your opinion of freezing clicking drives? I can't tell if this is quackery or legit so I'm just collecting opinions at this point. Thanks again.
m
0
l
Related resources
a c 311 G Storage
August 24, 2012 1:19:20 AM

A straight PCB swap on a recent model WD drive most probably won't work. You would need to transfer the "adaptives". Modern HDDs are calibrated at the factory to take advantage of those heads with better characteristics, eg frequency response. Ironically, swapping PCBs between identical good drives will probably result in both of them clicking. This is because they won't be able to access their reserved System Areas where the bulk of their firmware resides.

The following article should explain it for you:
http://hddscan.com/doc/HDD_Tracks_and_Zones.html

As for freezing, I'm not involved in data recovery, so my opinion is of no value. However, the data recovery professionals ridicule this practice. Some claim that the drives would be damaged by condensation, but at least one pro (who works for Seagate) claims that condensation does not occur. His reasoning is that condensation only happens when the air and the surface are at different temperatures. In fact most, if not all, HDD manufacturers specify a non-operating ambient temperature range that goes as low as -40C, and an operating temperature range that includes 0C.

Hitachi even has an EnduraStar model with an *operating* temperature range of -30C to +85C. It is specified for use in automotive environments.

Here is a thread that may interest you:
http://forum.hddguru.com/geek-squad-t20957.html
m
0
l
September 3, 2012 3:22:53 PM

fzabkar said:
A straight PCB swap on a recent model WD drive most probably won't work. You would need to transfer the "adaptives". Modern HDDs are calibrated at the factory to take advantage of those heads with better characteristics, eg frequency response. Ironically, swapping PCBs between identical good drives will probably result in both of them clicking. This is because they won't be able to access their reserved System Areas where the bulk of their firmware resides.


Thanks, I confirm that's exactly what happened with these two identical Hitachi drives. I saw a guy at a Defcon talk mentioning live swapping the PCBs on SATA HDDs to get around the SA issue, so I tried that, although I confess I can't imagine what the theory behind that would be (and to be fair he said it only solves the issue around a quarter or third of the time). It didn't work, although I was surprised to find Windows recognized a new drive was attached to the same PCB and tried to install drivers.

I then noticed when I connected the drive to a computer's motherboard and booted up the computer's power LED would flash and the computer wouldn't boot - up until this I had been using an external USB drive enclosure. That led me to think there were power issues so I did some googling, which led me to a thread on another forum from a while back suggesting diodes used for voltage protection could be causing the issue. On testing I found no potential difference across those diodes, took them out, and the was very happy to find that the drive then powered up again. Fzabkar actually gave the guidance in that thread and links, so thanks again.

I'm still searching for a diy solution to the clicking WD raptor.
m
0
l

Best solution

a c 311 G Storage
September 3, 2012 9:17:12 PM

I can't help you with your clicking Raptor, but the method involved in a hot-swap is to power up a good donor drive, place it in standby (with the motor spun down), and then transfer its PCB to the patient HDA, with power still applied. The patient drive is then woken from standby. (A live PCB must not be removed or installed while the motor is spinning. This is because interrupting the current in the motor's windings would produce very high back EMFs [voltages] which would kill the MOSFET drivers in the motor controller IC.)

The rationale behind this approach is to load the drive's firmware into SDRAM. The bulk of this firmware resides in the System Area (SA) on the platters. The "adaptive" data in NVRAM on the PCB are required by the drive to locate its SA. If the donor PCB were installed on the patient HDA and powered up from cold, then the donor's non-native adaptives would prevent the patient drive from locating its firmware.

The problem is that the patient drive now has a PCB with firmware, and adaptives, from the donor drive. I'm not a data recovery professional, so I don't know what kind of issues this causes.

BTW, the speaker at Defcon, was he Scott Moulton? FYI, he has numerous useful resources on his web site:

http://myharddrivedied.com/presentations-resources
Share
September 5, 2012 5:27:21 PM

Best answer selected by jason_farnon.
m
0
l
!