Very weird hard drive issue

[alegator]

I have a pair of WD320GB EIDE hd's, one that I use as the system main drive (WinXP Pro SP3) and the other one is a backup drive (a clone of the main drive that I keep offline). Both drives in excellent health (as reported by HDDlife), never had any issues with them, performing flawlessly and pretty much new (NOV07 drives). I defrag with Raxco PerfectDisk 8.0 once every 2 weeks, and it has done a flawless job at keeping my main drive healthy. About a week ago I cancelled the scheduled defrag job since I was doing some other task, so I manually had PerfectDisk defrag yesterday. Since it takes some time, I left it running, and when I came back a few hours later the PC was unresponsive (the keyboard and mouse frozen) with no signs of anything running except the PC fan (no HD clicking noise that would indicate malfunctioning). So I powered the PC off, let it sit for a couple of minutes and powered it on again, when I discovered that the BIOS wouldn't find any HD present (I even went into the BIOS setup to check this). So I powered it off again and booted from my Winternals CD to diagnose any issues. From the CD's Winternals XP it wouldn't find any HD present either. So I physically removed the HD and replaced it with the backup one (clone performed Feb08) and it would boot normally. So powered it off and physically inserted the main drive (the one not recognized by the BIOS) as secodary drive, with the clone as main drive, and no issues, both drives recognized and could be accessed within XP. Next I went back to the original config, removed the clone drive, inserted the main drive again as boot drive, and everything went fine, booted normally into WinXP. Of course I decided to clone the main drive just in case since the last clioning session was performed as said on Feb08. So what could've caused the BIOS not to recognize the HD in the first place, but later it did? As explained I haven't done any repair job other than power the PC off and on...

[Nelson37]

The cable could have slipped a bit, either the power or data, or gotten some dust or corrosion which was fixed just by removing or replacing the cable. The stiffer SATA cables do seem prone to seperation of the cable just after the head, this can most definitely cause intermittent operation.

The mobo IO could have suffered a momentary failure. Extremely unlikely.

The drive did fail after an extensive defrag. It would have been hot, and remained hot for at least a few minutes. Subsequent testing allowed enough time for major cooling. Fairly new drive or not, other than a bad cable this is the most likely scenario.

If you have extra SATA cables, I would replace them immediately. If you don't have extras, I would go get some. I'm sure you've had a pair of old headphones where one or both speakers would fail, depending on how you tilted your head? Bad cable. May not be visually obvious.

Then run a series of HD stress tests, timed to produce similar heat for similar time as the Defrag. Am assuming a second defrag would not run for very long as it was just done. Wouldn't hurt to repeat the test on both drives.

Check the Warranty on the drive.

[alegator]

Nelson, good analysis. One fact I forgot to mention is that when I physically removed the main drive (the one not seen by the BIOS) it was cold, as if it were offline or unplugged. This made me think that either the power or EIDE plug might have gotten loose or dust, which was corrected when I unplugged and reinserted the drive (as you suggest as one of the most probable scenarios). As said in my post the HD is EIDE, so the cables are not SATA, they are IDE cables. I've had a failure loing time ago when the HD wouldn't be seen by BIOS and it was corrected replacing he IDE cable. I feel this issue is HD unrelated, the HD is silent during operation, runs smoothly and HDDLife as well as WD's own diagnostic software are reporting normal SMART info and healthy status. If anything fails again I can relax knowing that a cloned backup drive is waiting offline. Regards.

[Nelson37]

MAJOR new detail, that means either the power cable slipped or failed, or the HD circuits themselves failed, and since it still works....

Even with the data cable off, it should still be warm.

Had a system like this a while back, finally found a hairline crack on one of the power cables, was just moving them clear of the fan when the HD died. Had put two or three drives in and out, testing on other systems, was driving me nuts. Power usually worked, but if the cable was at just the right anlgle, nada. Each time I'd R & R a drive, it was a crapshoot, though once they ran, they'd continue to run unless I re-arranged something else inside the case, which was usually done with power off.

So I'd add the drive, then two or three other parts, arrange cables, button up the case, and fire it up one more time, HD fail. Back to square one, one at a time, all OK, move PC to other workbench, HD fail. Testbench HD OK, "failed" HD is OK in testbed PC, "failed" HD OK again in original PC, back to square one, failure at some other point. That thing damn near became wall sculpture.

It was only because I happened to move that power cable to the particular angle with power on that the problem became obvious. No visible damage to cable or connector, but I could bend one wire and get a definite on/off switch.

Some testing was done on an alternate power connection, with drive outside the PC. These all worked fine, but when mounted, only one power cable was available. With drive outside the box, the cracked cable was forced into a working angle, I am guessing as it never failed in this configuration.