@Kuno :
Thank you very much for these very thorough explanations. (I'll have to re-read them carefully with detailed schematics next to the text to fully understand them – I'm not sure what some of the terms used designate exactly, like “ramp”, “suspension”, “slider” for instance.)
Quote:
Repair process, best guess: Drive is opened in clean room, and actuator assembly and platters are removed.
I thought that removing the platters was an exceptional measure, generally avoided even by the most seasoned professionals, unless there's an absolute necessity to do so (drive immersed in water for instance). Is the sequence of events you described (whether it does correspond to what happened in this particular case or not) a particularly severe kind of failure, despite the rather unimpressive way it happened, externally ? Is it potentially more destructive (or harder to deal with) than, for instance, a drive sustaining a sudden shock while in operation ? (I've recovered almost all the contents from a few drives which had sustained such damage by software-only means, namely ddrescue – for instance after my brother landed a punch on his laptop computer because Internet was too slow and he couldn't watch a Formula 1 race in live streaming...) The screeching noise I mentioned was very brief and faint, and no such noise could be heard the next time I turned it on (which I only did twice so far – the next day after the failure, while the temperature in the room was still very hot, and I gave it another try recently, hoping – foolish I know ! – that it could somehow “come to its senses” in a colder environment), only the regular clicking pattern.
Quote:
New actuator assembly from suitable donor drive is installed, with the suspension assemblies of the heads for any surfaces deemed beyond recovery removed.
In this case, that means that entire platters are not recovered, so the recovery rate must be very low, with many damaged files (especially the larger ones, since, as I've discovered recently on this forum, HDDs record data on platters “in serpentine fashion” (usually in strokes of a few hundreds megabytes as it's been demonstrated
in this thread), not sequentially as it might be more intuitive.
[I hadn't read the last paragraph yet when I wrote this, see below.]
Quote:
All critical system data is read from the drives most intact System Area, and pushed into drive RAM, modified to "hide" any missing heads to allow automatic initialization tests to pass while preserving zone table and translation mapping. This mapping is unique to every drive, and is generated during the drives assembly. It cannot be replaced with data from another drive without causing catastrophic data corruption. Once drive reports state "Ready", changes can be made to the drives RAM state to disable all of the drive self-repair functions which will hinder imaging. Imaging can then be attempted from the drives undamaged platter surfaces in accordance with the zone mapping.
So :
– The “live PCB swap” is just a myth ? Or it doesn't work with “modern” drives ? (I'm aware that even if it could work theoretically, it would imply that all heads are still functional, the theory being that it is supposed to bypass the system area and the self-initialization process.) If the mapping is only related to the P-list (factory-reallocated sectors if I'm not mistaken) as it's been said above, isn't there a possibility to compensate for that corruption by crude software means like shifting portions of the drive's image, after analyzing the P-lists from both the failed unit and the donor unit ? Or is this not practically manageable ? Do
all HDDs have reallocated sectors in the P-list, or is it not too unlikely to be lucky and have two units with an empty P-list ?
– Even if I did manage to swap the head assembly properly with very few dust particles landing on the platters, it is almost certain that it wouldn't work for those reasons ?
Quote:
And after that (assuming the imaging went reasonably well), the data has to be checked to see what (if any) of the original file structure remains, and repairs made where possible. Given that your drive has 2 platters, with all 4 sides in use (best guess, it is possible your drive is a "mule" unit, basically a cut down 1TB drive with 3 platters), and the zones (amount of data drive writes to one surface continuously before moving up the stack)are roughly 100MB, we can say that a single badly damaged surface will destroy any files over 300MB in size. If all surfaces are in reasonable good shape, then a very good recovery is certainly possible. But that possibility gets smaller with every click it makes.
Oh, right, I hadn't read this yet when I wrote the remarks above. That just confirms them. (Although, in the aforementioned thread, the “zones” were shown to be higher, like 263MB for one of the tested models, which would double the threshold.)
@jono-ats
Thanks to you too.
Quote:
So the next step, as Kuno-san noted, is to disassemble the drive in clean room and inspect the heads under a microscope.
So there's no way to say for sure if heads are damaged (and more generally to make an informed diagnosis) prior to opening the case ? If the above scenario is indeed what happened, is the damage likely to be visible to the “naked eye”, or not unless it's tremendously extended ?
Quote:
If the OP wants to try to figure out the cause and extent of the failure (but not really concerned about recovery), then I say, have at it.
People who are not in the business think that people who are are being unreasonable, haughty, greedy -- you name it -- when we discourage client interventions. You haven't seen the abortions that we receive nearly every day from clients who "watched a video" and made their drives [needlessly] unrecoverable.
Well, I
would like to be able to recover some files which I stored there temporarily (and nowhere else – they were recovered from another drive which I then re-purposed and filled entirely, and considering their “track record” I was under the impression that those 640GB units were highly reliable), but they are certainly not worth the cost of a professional recovery (they're more like historical archives, and wouldn't serve any actual purpose). If my main HDD had failed in the same way (a few weeks later I also had
trouble with it, a dreaded Seagate ST2000DM001, but managed to clone it in its entirety), with my most current working documents and pictures and e-mail archives and a gazillion other files gathered over the years, I wouldn't have taken any chance and would have paid the price, however harsh it would have been (although I did have a backup it hadn't been updated in months – bad practice I know, an automated online backup system would be much more efficient but my current measly uploading speed would make it highly impractical).
The problem is that this is an all-or-nothing situation : according to the informations provided here, there is nothing I could do with my current level of knowledge (which is definitely above average I would think, I'm often trying to explain technical aspects of hard disk drives' inner workings to people who see them as “toasters” as you say – although I realize that I'm still far from being an expert in that field) and equipment (quite basic and I couldn't affort the kind of advanced hardware/software interfaces that most regular contributors here use on a daily basis) to get a half decent result, or even just a few usable files, while paying for a full-blown professional data recovery service would be way overkill in that particular case. Considering this, what is the least stupid thing I could do with that drive ?
Just store it in the hope that one day I could become skilled enough (and adequately equipped) to perform the recovery myself with a good degree of confidence ?
@Spildit :
Quote:
If your PCB have an EXTERNAL ROM chip you should be able to read it with a CHEAP external programmer and a SOIC 8 clip like this :
I tried to gather some informations with the two first links you provided, but could not find any mention regarding the price of those devices (external programmer). I haven't yet explored the two other links.
Quote:
Regarding that model just go buy an identical drive, remove the heads of that drive and place them back !!! I'm talking about removing the heads of an exact same drive and place the NATIVE heads back. Now try to read the sectors of the drive. Most likely you will not be able to do so and you did ruin the drive.
Just because of the dust particles, or also because of the required level of assembly accuracy ?
Quote:
And we are not even talking about head compatibility and the fact that heads have their own micro-jogs settings that are stored on ROM module, etc .... We are talking about removing the heads from a drive and use those heads tha are calibrated for that drive on it's own drive. Now if you figure out that it've very hard to do so imagine tryng to do the same but with diferent set of heads that are not calibrated for that specific drive, that might not even be compatible, etc.... Try it. But do yourself a favour and buy a drive to test with first. When you decide that you can do a head swap with some confidence then try to get another drive of the same model and swap the heads from one drive to the other, then practice. After some attempts and some heads destroyed let's say 50 or 100 times you can move to the drive with data that you want to recover. Untill you manage to do the head swap procedure on a drive that you don't care about don't even think in doing it on the drive with the data you want to recover.
Well, that seems kinda daunting... but I guess that everyone here who's speaking with confidence went through this at some point ! I found a used WD6400AAKS in perfect condition for 25€, I'm still wondering if it's worth it to buy it to almost certainly ruin it, and possibly not even gain significant knowledge in the process, which would require that I spend 10 or 100 times that amount... at which point, if I ever become proficient to do it, it would have been cheaper to have it done by an established DR lab, and at which point the hard disk drive technology may be declining to the point where I couldn't even turn the acquired expertise into a sustainable activity !