Hi!
One of my 3.5" PATA drives went belly-up a couple of days ago, so now I'm trying to recover the data. Some of it was backed up, and some was not. :-( The failure seemed to occur "spontaneously and catastrophically" (i.e. the drive, while in use but not heavily loaded, seemed to instantly go from "all fine" to "lots and lots of unreadable sectors"). It did not suffer any physical shock.
As I'm working on it, slowly building up a partial image of the filesystem, I am starting to suspect that the problem likely is that I am completely missing one head (i.e. 100% of the data from one surface of one platter is unreadable, and 100% of the data from the others is fine). For example, almost _exactly_ 1/6 of the sectors are bad on a 3-platter drive, and all of the bad sectors are in evenly separated sets of adjacent LBA blocks spread across the entire range of LBA blocks, the spacing of which would seem to imply a single head/surface if the LBA-to-physical mapping follows a complete track, then jumps from one platter to the next before moving to a new cylinder, which I assume it does.
I don't want to jump to conclusions too quickly, but I'm wondering, does anyone have an idea whether a failure of this type is common? Is there a good source for the technical details on the drive (a Western Digital), such as the exact LBA-to-physical-CHS mapping? Has anyone had success in repairing such a failure?
Most of all, I'm wondering what is the likely source of a failure of this type -- the platter surface (seems unlikely since it really appears to have 100% of the data unreadable), the actuator/arm/head (physically misaligned or something?), a bad connection or broken trace, or a failure internal to the drive controller electronics. I'm completely unfamiliar with the design of modern IDE circuits (e.g. is there a separate amplifier, demodulator, data latch for each head, or can only one head be read at a time, etc.?), so it's hard for me to guess, and I'm frankly surprised at this apparent failure mode.
I'm looking for the failure mode of course as some guidance as to the best way to proceed. For example, if the problem is not the recording surface, there seems no point in taking the massively-repeated-attempted-read approach very far, as no number of repeated read attempts on a given sector will solve a physically broken connection. OTOH, if it seems likely that it _is_ a problem with the physical recording surface (or the actuator/arm/head), there is no point in even attempting to replace the logic board with one from another drive of the same model, and so forth.
Thanks in advance for any advice!
One of my 3.5" PATA drives went belly-up a couple of days ago, so now I'm trying to recover the data. Some of it was backed up, and some was not. :-( The failure seemed to occur "spontaneously and catastrophically" (i.e. the drive, while in use but not heavily loaded, seemed to instantly go from "all fine" to "lots and lots of unreadable sectors"). It did not suffer any physical shock.
As I'm working on it, slowly building up a partial image of the filesystem, I am starting to suspect that the problem likely is that I am completely missing one head (i.e. 100% of the data from one surface of one platter is unreadable, and 100% of the data from the others is fine). For example, almost _exactly_ 1/6 of the sectors are bad on a 3-platter drive, and all of the bad sectors are in evenly separated sets of adjacent LBA blocks spread across the entire range of LBA blocks, the spacing of which would seem to imply a single head/surface if the LBA-to-physical mapping follows a complete track, then jumps from one platter to the next before moving to a new cylinder, which I assume it does.
I don't want to jump to conclusions too quickly, but I'm wondering, does anyone have an idea whether a failure of this type is common? Is there a good source for the technical details on the drive (a Western Digital), such as the exact LBA-to-physical-CHS mapping? Has anyone had success in repairing such a failure?
Most of all, I'm wondering what is the likely source of a failure of this type -- the platter surface (seems unlikely since it really appears to have 100% of the data unreadable), the actuator/arm/head (physically misaligned or something?), a bad connection or broken trace, or a failure internal to the drive controller electronics. I'm completely unfamiliar with the design of modern IDE circuits (e.g. is there a separate amplifier, demodulator, data latch for each head, or can only one head be read at a time, etc.?), so it's hard for me to guess, and I'm frankly surprised at this apparent failure mode.
I'm looking for the failure mode of course as some guidance as to the best way to proceed. For example, if the problem is not the recording surface, there seems no point in taking the massively-repeated-attempted-read approach very far, as no number of repeated read attempts on a given sector will solve a physically broken connection. OTOH, if it seems likely that it _is_ a problem with the physical recording surface (or the actuator/arm/head), there is no point in even attempting to replace the logic board with one from another drive of the same model, and so forth.
Thanks in advance for any advice!
Facebook
Twitter
Instagram