3.5" Seagate Ext HDD arm repair.

[delad123]

Hello all. First may I say thanks for letting me join.
I am an amateur  tech guy age 62 so dont be hard on me. 
My problem is a Seagate 3.5" external harddrive.  The heads are damaged on the arm(see pic). The platters are scratch free . Initial problem was the drive motor was not running.  This has been solved by replacing a diode. But on running it, clicking was heard and on opening, arm was in park and heads were like this.
My question is. Can i get a replacement arm or have to buy a new HD. I have a lot of family data on this drive and would love to be able to retrieve it .
Any help greatly appreciated.

[Twoflower]

I would say: No chance unless you have opened that drive in a clean-room and took great care to not let the head-debris come close to the platters.

The only way I see to get your data back is using your backup (with no backup the data were not important) or use a professional data recovery service. That you opened the drive and contaminated the discs with dust probably not reduced the price for that service and reduced the chance to recover the data.

Some perspective about the dust problem: The heads fly on an air cushion of 5nm over the platters (depends on the drive). A human hair has an diameter of over 100µm. 


Edit: Fixed a typo

[delad123]

Thanks for the information.  This hard drive was the backup drive as the data on it was from an old PC.

[james_s]

Yeah no way, even if you salvage a head assembly from another drive it's highly unlikely that it will ever work, and if you do get it to work it will never be trustworthy. The heads and platters inside a hard drive are extremely fragile high precision assemblies.

[golden_labels]

Unfortunately information on this drive should be considered lost. Recovering data is technically still possible, but requires specialized equipment and expertise. Having a cleanroom at one’s disposition — as mentioned by others — is already putting that beyond what a hobbyist can do. Companies offering such services charge hundreds to thousands of dollars per drive, and there is no price list — one must contact the company to negotiate the cost. Basically out of reach of a typical consumer, who would like to recover family photos. However, I am mentioning that option in case those are not just family photos.

A relevant video from Dave, showing just one aspect of how incredibly precise and complicated mechanism HDDs are:

[MikeK]

I see no problem with attempting a recovery.  If you can find an identical drive you might try swapping the platters into the working drive.  Or swapping the arms?  Decades ago, HDD stiction was a problem and I wasn't the only one who got use out of a couple of drives by opening them up and unsticking them.  The worst scenario is the one you're already in...no recovery.  So why not try?  Yes, the drives are assembled in a clean room.  So?  You're not looking to get anything more than data recovery out of the drive.  Be as clean as you can and don't do something stupid like attempting a fix while eating crackers over it.

[delad123]

Thanks everyone for your answers.  I have fixed stuck arms before and managed to get the data from the drive.
I then would not tempt fate again so I scrapped the old drive.  This is the first time I have ever seen this with the heads.
And for it to not damage the platter.     
I thought I might be able to get the arm on its own.  I will see if I can get the same type of drive as close to the numbers on mine.
Will I still have to change the BIOS on it as I have read about this before.
Thanks again. 

[Ranayna]

Thanks for the information.  This hard drive was the backup drive as the data on it was from an old PC.

So you still have the originals? Then make a new backup ASAP and count yourself lucky.

If you do not have access to the files, and the now broken disk has the only copies, think *very* hard about how to proceed. The chance for success on your own are miniscule.
A professional data revovery company would have almost certainly gotten the data off of the drive. Well hindsight is 20/20, so that does not really help you...
But a professional recovey company likely can still recover some data, but with whatever you did to the drive yourself this will now be significantly more expensive.
If you are absolutely sure that you cannot spend the money, or the data on the disk is still available elsewhere, you can proceed trying yourself. If nothing else it will be a chance to learn

I see no problem with attempting a recovery.  If you can find an identical drive you might try swapping the platters into the working drive.  Or swapping the arms?

I a modern harddrive there is virtually no possibility of removing the stacked platters as a whole. Thinking about it, i don't remember tearing down any harddrive in the last 15 years where that would have been possible. The screws to remove the motor are either under the platters on the inside of the drive, or the motor is glued/welded in and can't be removed at all.
Since you must not loose alignment of the platters in a multi-platter drive, you would need a jig to make absolutely sure that the platters cannot turn against each other once you loosened the screws holding the stack together.
If you swap platters, do not forget to also swap the PCB or at least the EEPROM, since that stores calibration information.

Depending on the drive construction, you may be able to remove the arm with the platters in the drive. But the head construction is so delicate that i can't imagine doing that without bending the tips with the heads, unless you have some specialized tools. Considering the tolerances in a modern drive, you will very likely smash at least one of the heads immediately when turning the drive on.
Also there is the flatflex from the heads to the PCB. Most of the times this is not designed to be removed.
I have no idea if the calibration information in the EEPROM also contains info related to the arms or heads. If it does... well you new arm will not be calibrated and might not be able to read any sensible data at all.

The chances for success are very slim. And thats not even mentioning the cleanroom yet...

[james_s]

I don't think a DIY clean room or glove box is out of the realm of a hobbyist, I mean just yesterday I saw a video of a hobbyist making his own ICs from scratch in his garage. Repairing a hard drive is not impossible, but it does require expertise, patience and some specialized tools.

[golden_labels]

There are people, like The Thought Emporium. But those are exceptions. Rare cases on the long, thin portion of the bell curve. Typically dedicated to a particular goal and wishing to spend enormous effort and some money on their endeavours.

Of course, in binary terms it is not impossible. If you want to setup a 125 L “clean box” and take chances working with the drive in that kind of an environment, you can. Early experiments with devices, now considered requiring extreme precision and conditions, started in places far from perfect. Any company offering data retrieval from damaged storage media was built by people, who before that were not different than us and started from a similar position. So, obviously, not impossible!

But this is not a binary problem. It’s not a question of whether it is not absolutely impossible, but if OP’s chances of that being successful are big enough. If putting effort and pouring money into that makes any sense. Realistically speaking, they would be extremely lucky if that worked.

[Halcyon]

A clean room isn't really required. We used to do head swaps all the time using a small laminar flow cabinet. The filtering was very basic. The success rate was high. I just wouldn't count on that drive working reliably for long periods of time after that.

My suggestion is, unless you've done platter/head swaps before and have the right tools, then leave it to the professionals. There are data recovery companies out there (like Payam) who will do the hard work for you.

With some of the more modern/higher capacity drives, it's not just a matter of swapping out a controller board or platters into a donor drive anymore. There is usually an EEPROM chip that needs to be swapped out as well. Depending on how bad things are, there are specialised tools which will disable one (or more) heads and just read out the good platters. You won't get all the data this way, but you'll at least get some.

[Ranayna]

[...]
With some of the more modern/higher capacity drives, it's not just a matter of swapping out a controller board or platters into a donor drive anymore. There is usually an EEPROM chip that needs to be swapped out as well. [...]

Is having an EEPROM or other non volatile memory somewhere on the PCB really a new thing? I thought that was used since at least SATA, and likely even earlier. I would assume that any drive that supports SMART needs some kind of memory.
Some current drives now have the EEPROM embedded into the main chipset, which makes swapping it a pain.

[golden_labels]

Ranayna:
I was successfully swapping PCBs for data recovery as far as early 2010s and that were already SATA drives. So either no data affecting reading was stored in non-volatile memory or it was stored in something hidden in the enclosure.

Halcyon:
Did you have any openly published reports from such operations? I stopped looking into the topic ca. 2015, but even then it was already perceived as a really troublesome thing. And not even due contamination of the surface, but precision required and how hard HDDs became to disassemble without damaging anything. Would be interesting to see some such experiments with newer drivers.

[Ranayna]

I was successfully swapping PCBs for data recovery as far as early 2010s and that were already SATA drives. So either no data affecting reading was stored in non-volatile memory or it was stored in something hidden in the enclosure.

I find that very interesting. Thank you for that information.
I have to say that most of my knowledge regarding harddrive repair is hearsay. I never had the need (knock on wood) for such a repair for myself.
Broken company drives where we were not able to copy the data off of were always sent to a recovery company.

Considering SMART Values, i would have to assume that the location of pending and reallocated sectors needs to be stored *somewhere*.
But depending on the reason for the PCB swap, that may not be relevant, or might only affect a very small amount of data. I can very well imagine that a couple of "lost sectors" are not easily noticed.

[golden_labels]

Note that I swapped PCBs only to read data from the drive. That is: mount the file system read only, copy data, done. I was not caring about SMART and such: that might have been completely garbage. To be clear, I was also not doing that professionally: only drives belonging to me and friends/neighbours. Less than a dozen of such recoveries in total, with like half of them with drives bought from the same batch.⁽¹⁾

DiTBho was dealing with HDD firmware directly, so perhaps they may provide more detailed information.

⁽¹⁾ Bought together, usually with the intention of being run in parallel (not necessarily RAID) or for the purpose of PCB swapping if one fails.

[Halcyon]

Did you have any openly published reports from such operations? I stopped looking into the topic ca. 2015, but even then it was already perceived as a really troublesome thing. And not even due contamination of the surface, but precision required and how hard HDDs became to disassemble without damaging anything. Would be interesting to see some such experiments with newer drivers.

If you Google Scott Moulton, I think he has some resources and videos out there, but he charges for his training. I might have some old resources of his at home which I can organise to send you if you're interested, just flick me a PM.
You might also be interested in these series of videos, which is taken from his hard drive repair course:



It does require specialist tools, but they are typically fairly affordable for small/home labs.

[golden_labels]

You might also be interested in these series of videos, which is taken from his hard drive repair course:

I was just wondering if you were publishing some of that, for benefit of all of the community, not just me. Some people have devblogs etc. I do not wish to educate myself, in the strict sense, in that direction. More like see where things are nowadays.