Rusted Hard Drive PCB?

[Jon]

I have a Western Digital 500GB drive and it started clicking recently, needless to say I think I've lost everything on the drive but in an attempt to do my due diligence I've tried a few things to see if I can bring the drive back from the dead such as putting the drive into a sealable sandwich bag and putting it into the freezer for 48 hours, no dice but hey it *was* a long shot. I then removed the PCB and noticed that where the PCB connects to the HD itself the copper plating seems to have rusted, or at least appears to have rusted. Can anybody take a look at these pics and let me know if this is the case? The thing is it doesn't appear to be regular rust, it shimmers in the light a myriad rainbow of colors. IF this is rust, I have another drive that is the same model, I could replace the PCB's and at least see if I'm able to recover the data that way.

[just brew it!]

Hard to be certain, but that does indeed look like some sort of corrosion. Was the system in an area with high humidity?

If the corrosion is the reason for the problems with the drive, you could try gently cleaning the contact points with a pencil eraser. I wouldn't trust it long-term (it may corrode again in fairly short order since you've probably lost some of the protective plating), but it ought to be good enough to get your data off.

Also try running the drive upside-down. I had a WD that developed spindle bearing problems, and would'nt even spin up in the normal orientation. Flipping it over allowed me to spin it up and get my data off.

[Dr_b_]

Putting the hard drive in the freezer is a bad idea, because, even though its in a bag, when you take it out into room temp condensation can form on it and in it (hard drives are not air-tight). Try swapping the PCB with another from the exact same make/model/production run of the same drive.

[Jon]

Hard to be certain, but that does indeed look like some sort of corrosion. Was the system in an area with high humidity?

If the corrosion is the reason for the problems with the drive, you could try gently cleaning the contact points with a pencil eraser. I wouldn't trust it long-term (it may corrode again in fairly short order since you've probably lost some of the protective plating), but it ought to be good enough to get your data off.

Also try running the drive upside-down. I had a WD that developed spindle bearing problems, and would'nt even spin up in the normal orientation. Flipping it over allowed me to spin it up and get my data off.

No humidity, this is Canada winter time

I'll try your suggestions, hopefully it works.

[Jon]

Looks like I'm out of luck, the only remaining option here is replacing PCB's. I'll try it out next week sometime. Thanks for the input!

[lonleyppl]

It's likely that just replacing the PCB won't solve your problem. When you switch PCB's make sure to switch the EEPROM chip as well, so your drive information will transfer. It's a tiny little chip that requires a lot more skill with a soldering iron than I have to remove and reattach. When I did this with two drives 6 months ago, I had a friend from church who regularly solders things this size (and has the equipment to do so) do it for me.

[Welch]

To be honest it looks like more than rust.... the rainbow look your talking about is typical for metallic items that get really hot. You can see this on things like polished exhaust on motorcycles, snowmachines, ect. The connectors to the right in the bottom picture looks a lot like what im talking about. The pin connectors on the left at least in the angle you posted does look like rust.

You'd more than likely be wasting your time with the pcb from the other drive. I've had identical drives before, both bought at the same time with identical firmwares and no dice with transplanting. There are a few websites that guarantee getting you a replacement pcb that will supposedly work with your drive. They have you send in some detailed info about your drive.

Is there anything on the drive worth recovering? If so places like OnTracks can recover pretty much anything even if the spindle is damaged.

[vargis14]

It looks like a old IDE drive.i can see the legacy molex power.I know he is trying to get his data off. I would have to agree with"god its 4:15 am sips coffee"someone who posted before me,It does look like heat caused that.
Being a IDE drive its probably real old and at 500gb probably the biggest IDE drive i have ever seen.But now that i think about it The original sata drives came with a sata power connector and a legacy power connector,since a lot of PSUs did not have the sata power plug when the new sata format was introduced.I have one of the original 10k raptors it had both power connectors.But you can only use one or the other.use both and a poof is possible.They really should put a red warning sticker on them instead of tiny print the same size as everything else on the drive sticker.so yes you guessed it,I think i cooked my old raptor drive.But the bios still sees it once the promise sata/raid driver loads when the computer first posts,but once in the bios its nowhere to be seen.I tried a 80 gig sata drive i know works,Same result ,the sata/raid sees the drive but once i get into the bios its not listed.
I know this because i got my old 940 pin fx-53 running with 1 gig of reg corsair xms ddr400 mem,A bad azz asus geforce mx-440 64 mb agp card.It was that or a skinny small ati rage 128 card.The old girl runs at 2.4ghz stock,Well granny still has some moves left 2665mhz 13x205 with 1.67 volts.Idles at 37-40c loaded 50-60c using a 4800x2 amd stock heatpipe cooler that works very well with some old/new thermaltake white heat paste,that came from thermaltakes first tower cooler with the lightish metallic blue side covers/shrouds and a picture of the empire state building running down the middle of the shroud on both sides.
Its pretty snappy for what it is.With MPHC it plays 720p video at approx 50%cpu usage.I think that because the mx440 64 meg is really doing some HW accel.VLC 2.0 with the same vid is a slideshow.with Google chrome on youtube 720p works fine also but cpu usage moves between 65-80% usage,1080p drops half the frames and pegs usage at 100%.I just might pick up a decent hdcp compliant video card with hdmi out.Newegg has a HIS agp hd4650 1gb card avail for 54$ open box...80 or so for a new one.
Sorry went off topic there Gluck getting your data.

[just brew it!]

A lot of first-gen SATA drives had both SATA and Molex power connectors, since most PSUs didn't have SATA power plugs yet.

I agree the discoloration could be heat-related, in which case the drive may be toast.

[lonleyppl]

It's not guaranteed that the drive is toast...

What you can do is replace the PCB, then switch the EEPROM chip, and it might work.
I had to do this when one of the drives in our HTPC died. Came from HP setup in RAID 1 and had some important info which my dad hadn't backed up. Got a spare drive off the Internet, switched the PCB's and had a friend move the EEPROM chip from the old, burned out PCB (you could see that the controller had burned out), to the new PCB. Still works, but it's since been imaged and had the image applied to a new drive.

[just brew it!]

The worry is that if the PCB got hot enough to get discolored, there may be an underlying mechanical issue that caused it (like say a burned out motor or seized spindle bearings).

[Welch]

The worry is that if the PCB got hot enough to get discolored, there may be an underlying mechanical issue that caused it (like say a burned out motor or seized spindle bearings).

DING DING DING! Cooooorect! What do we have for him Johnny! (Jim Carey Voice), sorry, just had my morning brew (coffee, not beer)

The other route you can do is attempting to clean off those contact with something none-abrasive to try to make them make good contact again. If you look at the PCB in the second picture, the chips themselves (this doesn't mean everything) aren't burnt up at first glance. You can usually tell if SOME chips are burnt up if the white writing on them start to fad out. I'm not sure if this is dependent on how the manufacturer wrote the lettering on (etched, laser, ink.. ect). But I've seen all kinds of the lettering fading or blending in with the black portion of the chip once they get to hot. If you see that its a sure sign of overheating of the chip.

My point is that if the chips themselves are fine, and assuming no other components are damaged, you may be able to clean up those contacts a bit and revive the drive long enough to recover your data. I'd try the pencil eraser method first mentioned here. You can also try other things like rubbing alcohol, vinegar, ect. Last resort (might be one of my first) you can take a soldering iron to the surface if its got a slight bit of solder on top of it and attempt to re-solder the contact. Similar to a "cold" solder, this may bring back that shiny clean surface you want on a contact. If there is no solder on it, you can also try melting a very small amount of solder onto the contact, then removing it. This may apply heat and fresh solder + resin to the contact, sort of refreshing it. I've used that method on other contacts and it works great. But its completely dependent on the material used in that contact, Gold, Silver, Tin, ect.

[DLHM]

Many years ago, (15 or so) I had to go to a architects office to work on one of their servers. When I open the case, it was rusted inside. Basically all exposed metal that I could see had a layer of red rust, or white corrosion on it. The server was in a small closet where they stored there blue printing chemicals, one of which was ammonia. You dont have any chemicals around do ya? Frric cloride for etching may do it.

[CB5000]

Hard to be certain, but that does indeed look like some sort of corrosion. Was the system in an area with high humidity?

If the corrosion is the reason for the problems with the drive, you could try gently cleaning the contact points with a pencil eraser. I wouldn't trust it long-term (it may corrode again in fairly short order since you've probably lost some of the protective plating), but it ought to be good enough to get your data off.

Also try running the drive upside-down. I had a WD that developed spindle bearing problems, and would'nt even spin up in the normal orientation. Flipping it over allowed me to spin it up and get my data off.

No humidity, this is Canada winter time

I'll try your suggestions, hopefully it works.

The pencil eraser might not be abrasive enough to get that corrosion off. I had this kind of corrosion before on some old electronics I was tinkering with on the copper contacts. I had to use some fine sand paper (400 grit) to get that oxidized copper off. A finger nail polisher block works well too since it usually has 4 levels of abrasiveness... you can get those contacts to a mirror finish, and they only cost about 99 cents at CVS or walgreens etc. I usually peel off the individual sheets from the block an use it like sand paper. Usually contacts like that are coated with gold or other highly corrosion resistant metal, but once in a while that coating fails and you get a nice film of oxidized metal that acts as a nice electrical insulator.

I highly doubt that the corrosion was due to heat, as the temperatures you need to reach to cause that kind of film to develop would also partially melt the PCB.

[ludi]

I highly doubt that the corrosion was due to heat, as the temperatures you need to reach to cause that kind of film to develop would also partially melt the PCB.

Do you have a reference for that? My understanding is that the color change (indicating that a thin oxide film has formed) is a combined function of the applied temperature, time duration, and ambient moisture or other chemicals. A reduction in one can sometimes be compensated by an increase in another.

Note that the discoloration only occurs at the screw mounting points and the interface contacts, suggesting that the hard drive chassis was, in fact, the heat source. There's also some evidence of PCB delamination around those screw mounts, which is the first stage of the "melting" you speak of.

[Ryu Connor]

http://youtu.be/ICR-xw1FYlU - PCB Swaps are unlikely to work.

http://youtu.be/ad1uVAB5bNA - The Freezer trick may work, but it further destroys the drive.