HD Keeps Spinning Down: How to ID?

[Piyono]

One of my 3 internal hard drives has taken to clicking off, spinning down and whirring back on again. This phenomenon does not affect my computing, so I assume it's not my system drive. Still, I'm concerned that one of my drives may be failing. SMART tests show nothing out of the ordinary (of course).
If the problem happened every few minutes I'd happily sit myself down in front of my opened computer case and try to determine by ear which drive was afflicted. Unfortunately it happens with no fixed schedule making that option very impractical. Is there software that I can use to help ID the problem drive?

Piyono

 

[jtr1962]

It may not be a drive problem at all. About two or three years ago my main PC was exhibiting similar symptoms. First one of the drives was periodically spinning down. Eventually all three of them started to. The number of reallocated sectors started climbing on two out of the three. As things got worse, my machine locked up when I was burning CDs. I thought maybe the motherboard was going but as an electonics engineer I'm keenly aware that a poor power supply can cause exactly the type of problem I was experiencing. The computer at the time was about 5 years old. Both me and the guy who gave it to me used it constantly. I figured the power supply was going so I bought a Seasonic. The machine worked perfectly after that (and still works perfectly). It's also much quieter.

My recommendation is if you have a known good supply in another machine swap it with the one in this machine and see if the problem continues. If this cures the problem, then buy a new supply.

 

[Piyono]

I actually did upgrade the power supply on this machine to an Enermax model a couple of years ago when the AOpen power supply couldn't handle the load. This is a 420W model and it's powering an Athlon XP 2600+, 3 HDs, a Radeon 7500, a PCI modem and a PCI sound card. Not a heavy load. And even if it were, why is only one drive exhibiting these symptoms?

Piyono

 

[timwhit]

I had a WD drive that started doing this a couple months ago. I don't know if the drive is bad or not. I replaced it with a Samsung drive and I haven't had the problem since. I have 5 drives in my system and the WD is the only one that was having a problem.

I don't know if I would trust this drive in the future or not. I thought it might be my power supply, but if another drive is fine in the system, it would lead me to believe that there is a problem with the WD drive and not the power supply.

 

[ddrueding]

...or at least that the WD drive is more power-hungary than the Samsung.

 

[timwhit]

I thought of that too, but why only that one drive spinning down and not any of the other ones?

 

[Handruin]

I had a WD 18GB Expert drive that did that years ago (I've since stopped using it). I never figured out why it did it. It wasn't the power supply because I moved it between 3 different systems (with 3 different power supplies). It was not over heating because the drive was always mild to the touch if not room temperature. I have no idea what causes that to happen. (and all my machines have always been on a UPS, so I don't think it would be from my power.

 

[P5-133XL]

The only time I've experianced it was on a machine with many drives. So many, that I had to use molex splitters. How ever, it wasn't just one drive, but a random drive that would power down and back up. This was a problem because it kept breaking my Raid-5 array. I found that the problem was an intermittant connection caused by the molex splitters: I could reproduce the problem by wiggling the power wires. I kept replacing splitters, but that didn't help: The new splitters did the exact some thing.

Eventually I fixed the problem by getting a modular PS that had enough molex connectors that I didn't have to use any splitters.

 

[timwhit]

It could have been powered with a splitter. I replaced my case since then so everything is connected differently now.

I don't really need the space right now or want to lose upto 250GB of data is the drive is bad, so I think I will just leave it in a box for the time being.

 

[Mercutio]

This is me coming to say "I told you so!" to timwhit.

I told you so, timwhit.

That is all.

 

[Piyono]

The drive in question is a Samsung SP2504C. It and a second SATA drive — a WD3200KS — share a drive cage on the floor of my PC case, where they are mounted a few millimeters apart, the cooler (ostensibly) Samsung positioned below the hotter WD. There is no airflow to these drives. My S.M.A.R.T. monitor pegs their temps at about 48°C, which, I think, is well within spec.

I was going to run Samsung's Hutil diagnostic software but I'm out of room on my backup drive and don't have a current backup of the data on the Samsung so that will have to wait till tomorrow or so, when I can pick up a new HDD. I'm a bit paranoid about losing data and the Hutil page makes a strong statement about possible data loss while running the utility so I'm not taking any chances.

In the meantime I pulled the drives out of the cage and onto the floor of my bedroom where they're running much cooler. I've been sitting here for an hour or so now and haven't heard that stress-inducing *click-whirr* yet.

Who knows... maybe it was caused by overheating? I'll keep y'all posted.

Oh, another something to note: Yesterday while backing up from the Samsung to an external drive on USB 2 I was experiencing transfer rates of like 10MB/s, which was mildly disconcerting. I'd try again now with the SATA drives in the open air but the backup drive is full...


Piyono

 

[Piyono]

Oh, OK, ANOTHER really curious phenomenon:
I store my music on the Samsung. Lately when I've been playing music off that drive it would play back sounding like an old, abused tape machine; it would play too slow, dropping the pitch, and I'd get these weird "wow and flutter" kinds of effects. Very, very strange.

Piyono

 

[paugie]

My S.M.A.R.T. monitor pegs their temps at about 48°C, which, I think, is well within spec.

I wouldn't say that. I think a sustained 48deg would be courting disaster.

 

[ddrueding]

My HDDs were around 40C for a while, and I changed from a SFF case to a P180 primarily to get their temps down.

 

[Piyono]

Ok, update...

The drives are now hover at about 40°C in their new (temporary), open-air configuration.

Now that I've been sitting here for a few hours the Samsung has started its song and dance, only it seems that I was imagining the "whirrrr" part of "click-whirrrr". I The drive is clicking, but it does not appear to be spinning down.
Gonna HD Tach it to see how the transfer rates are holding up.

Piyono

 

[Bozo]

Sounds like the drive is dying a slow painful death. Hopefully you got it backed up.

Bozo :joker:

 

[Piyono]

Just picked up a new WD3200KS. I'll dump the contents from the samsung tonight and prepare an RMA when that's done.

Two interesting points to ponder:
1. I bought this WD3200KS retail at CanadaComputers; I don't know how these stores do it, but they matched my channel suppliers' prices. Same price and less paperwork? I'll take it.

2. Samsung doesn't have an advance drive replacement RMA program, which really bites.
... I was just going to contrast that with Maxtor's advanced return policy, when it struck me that Seagate might have overridden that. I checked. They did. No more advance drive replacement, although they're "looking into it".

Piyono

 

[ddrueding]

Nope. The only people who do advanced replacement are the people that have lots of practice doing replacements: WD.

Having never had a Samsung die; I am not familiar with their replacement process. WD, however, does have the best RMA process out there (except that they ship you WD drives).

 

[Piyono]

Ddrueding, in the 10 years that I've been at this game I've replace hard drives from every major player. Actually, every player but Samsung because this is but the second drive of theirs that I've ever purchased (the first one lives happily in a USB enclosure on my desk, backing up my photos). Now that Samsung is pulling out of the Canadian HDD market I have a feeling I won't be installing too many more.

Anyway, here's an update:
I tested the defective Samsung using their HUTIL diagnostic tool. It sped through the test, clicking and stuttering all the way. HUTIL's diagnosis? No problems detected here, buddy! Your drive is A-OK! A clean bill of health!

Juuuusst great.

Now I envision the bench technician at Samsung plugging my drive into his test system, firing up the same tunnel-visioned utility, and walking off to get a coffee instead of monitoring the progress. He comes back to find flying colors all over the results page, marks the drive as 'OK' and has it shipped back to me (C.O.D.).

Can anyone think of a utility that might be able to identify the defect? I'm going to tape a note to the top of the drive with instructions on how to reveal the problem, but I'd like something to back up my claims.

Piyono

 

[ddrueding]

The pessimistic suggestion would be to assure that their util finds something before returning the drive.

 

[Handruin]

If you've said this happen regularly during I/O...how about a tool that tests seek time or sustained transfer rates? If the drive is spinning down/clicking...the seek time or transfer rates should be terrible. That might show a significant problem that you could complain about. It sucks to have to play such a game, but it might be the only way to show first hand a problem.

 

[Buck]

I tested the defective Samsung using their HUTIL diagnostic tool. It sped through the test, clicking and stuttering all the way. HUTIL's diagnosis? No problems detected here, buddy! Your drive is A-OK! A clean bill of health!

Juuuusst great.

Now I envision the bench technician at Samsung plugging my drive into his test system, firing up the same tunnel-visioned utility, and walking off to get a coffee instead of monitoring the progress. He comes back to find flying colors all over the results page, marks the drive as 'OK' and has it shipped back to me (C.O.D.).

Can anyone think of a utility that might be able to identify the defect? I'm going to tape a note to the top of the drive with instructions on how to reveal the problem, but I'd like something to back up my claims.

Piyono

I just cannot see the drive passing through the recertifcation process at Samsung. When the drive arrives, it will be physically inspected (to make sure it hasn't been tampered). Then it will be put in a test chamber with many other drives and tested by proprietary software much is more complex than HUTIL. If it fails (which is what I would expect) it will go through another test sequence that will take several hours in an attempt to correct the issue if at all possible. If it does and it can pass the test, the final process will involve an acoustic test. If any internal parts need to be replaced (e.g. heads), the drive will be scrapped.

 

[Bozo]

Piyono "Can anyone think of a utility that might be able to identify the defect? I'm going to tape a note to the top of the drive with instructions on how to reveal the problem, but I'd like something to back up my claims."

44 Magnum?

Bozo :joker:

 

[Mercutio]

Spinrite?

 

[Piyono]

OK... latest developments are indicating that the issue might just be with the SIL 3512 controller on my motherboard... I had the brilliant idea to try the drive on my dad's computer which sports a brand-new GeForce 430-based motherboard. The Samsung clicked once or twice but I could not get it to behave like it did on my computer, despite loading it down with multiple simultaneous scans and audio files playing back. Now I'm back on my computer playing back music from the new WD drive and it is exhibiting symptoms similar to the Samsung (as I described above), although it's not clicking.

Deleted the SiI3512 drivers and rebooted Windows but that didn't help. I've been using SATA drives on this controller for a while, now without problems. I can't think of anything I did recently to invite this kind of behaviour.

Will investigate further tomorrow.

Piyono

 

[Sol]

Maybe you should test the drive using a live Linux cd, that should tell you if it's drivers or if the firmware or hardware is at fault more easily than mucking around trying to figure that out in windows. From there I would envisage that reinstalling windows, flashing the firmware or replacing/RMAing the board should be the solution...

 

[Piyono]

Sol, good idea...
I'll boot using the Ubuntu live CD and see what I can find out that way. Not that I have any idea what utils to use to test the controller...

No chance of RMAing the board as it's well past its prime.

Piyono

 

[Piyono]

OK, I'm in Ubuntu now.
While booting off the live CD I got a string of "Buffer I/O error on device hdc" prefixed and suffixed by long strings of numbers.

I'm scrolling through the logs but to be honest I don't really know what I"m looking for.
Any tips?

Piyono

 

[Sol]

I guess your really just looking for any errors and listening for the clicking... If you can ensure that your controller should work in the version of Ubuntu your using and that the cd passes it's self test (assuming the version of the live cd you have has the option of testing itself) then any problems/noises experienced would tend to rule out a windows software problem.

 

[Piyono]

OK, so I've been slacking off on this issue a bit, but I wanted to clear up the issue of the skipping audio. I remembered shortly after mentioning it above that this is actually an old and noted problem with the Sil3512 controller and the M-Audio Delta 1010 audio interface (which I use). Any audio being played from a hard drive on this controller through this particular sound card will glitch. I knew this from troubleshooting the issue once upon a time when I recorded vocals for a guy who had to run out of town, only to learn that there was clicking and popping all through the take. (We managed to re-record and thereby salvage the session.)

Anyway.


Piyono