ever had heat cause a hdd crash?

[xbassman]

A couple of days ago my HTPC started crashing on me.
I figured dust build-up was the culprit, since I have been remodeling and there's been a lot of drywall dust floating around.

Well, I cleaned that out and noticed that my video card fan had crapped out (once again ).
It's an old Radeon 7500. Not really worth the cost of a new fan and I have brought it back about 3 times before. OK, new card on the way, but noticed something strange.

I normally install UD on my PC's. This program keeps the CPU @ 100% so it generates a lot of heat. With UD running, once the CPU heats up (keeping in mind dead vc fan is keeping temps higher than normal) the PC would crash. I can faintly hear the seek sound (except pretty quietly) that a hdd make when it is failing.

Now, if I don't run UD, it won't crash at all. I also double checked by installing prime95. It caused the same crash scenario. (No overclocking...will normally run prime95 indefinately)

I am going to make up a boot cd with testing software for Samsung drives so I can verify I haven't borked my drive. I have just never run into this before.

Anyway software temp readings are bogus:

idle- CPU-38C, mobo-19C (75F in the room )
load CPU-47C mobo-28C

 

[Tiamat]

either your system isnt stable or heat is the culprit or both.

Run memtest86 to make sure your ram is not at fault. If you get even 1 error in 3 consecutive runs of all 11 tests, your ram has some fault to it. This will take about 2 days straight to do 3 complete runs (not just the normal 7 tests).

When you get your new graphics card, see if it can go through video benchmarks like 3dmark without problem.

Check to see if your hdd has bad sectors (Norton disk doctor or something better)

underclock your processor and see if your program still causes your computer to crash.

Touch the harddrive when the system is loaded up and see if it gets very hot. if its hot to the touch, then it needs active cooling.

What power supply do you have. A bad power supply can cause things like this to happen.

 

[xbassman]

Ya know I didn't even think to load memtest....

I booted it with "Ultimate Boot CD" and the Samsung hdd utility didn't run, but it is currently running memtest86. ('bout 15 minutes so far...)

I'll see if I can figure out why the Samsung diagnostic didn't run. (probably chipset related (nforce2)) Meanwhile while running memtest I heard the occassional tap that sounds like it is coming from my hdd. It didn't cause any errors though.

Anyway my PS is an Antec True430. It is mounted upside-down in a desktop case with my hdd mounted over it's intake fan. (I have 2 other places I can mount my hdd, one of which would probably get a little fresh air from my Panaflo 80L1A intake fan)

Upon its initial build it was tested and passed Prime95, memtest86 and 3dMark without errors.

I still think the extra heat has caused hdd errors that chkdsk didn't fix. (hopefully not physical ones)

 

[xbassman]

update:

memtest passed (1 hour) good enough for now.
I got the Samsung utility running from the boot cd now (passed all preliminary test doing surface scan now)

Actually this boot cd is a pretty good free download. It has a ton of utilities. Pretty easy also. Just downloaded the .iso and used Nero to burn it.

If you want to check it out google "Ultimate Boot CD".

 

[Tiamat]

Perhaps the issue is with software. Have you installed anything new between the time it was flawless and now?

 

[xbassman]

Originally posted by: Tiamat
Perhaps the issue is with software. Have you installed anything new between the time it was flawless and now?

No, but when it started crashing, I noticed my hdd running in pio mode, so I removed the controller driver through device manager (WinXP) and let it redetect and reinstall. That driver could possible be corrupted from a hard crash......

It passed the Samsung drive tests including surface scan.
This software reported drive temp @ 27C.

Then I found on the cd a prime test.
It runs from a dos prompt, but appears to be the same tests that prime95 torture test runs.

I started it up and let it go. I was surprised when i checked it again (about 30 minutes later) and it was still chugging away. When I checked the next time I saw a blank screen. I hit ^C and found that the test was still running and then aborted per my interrupt.
The PC goes to reboot and hangs on detecting ide devices....

hmmmm..... chipset too hot?
anyway I have until Monday before my new video card arrives.
I hope I don't regret ordering a passively cooled card.

 

[Shadowknight]

55C and higher equals HDD problems. I know from experience. Fortunately, my drive seems to be fine now, but when it was overheating, the mouse pointer would hang before moving on the screen, switching programs or using the taskbar would hand, and Windows Software registry was nuked (though that wasn't the first time Windows crapped out on me anyway).

I thought the drive was screwed, but after my Raptor spontaneously died, I slapped my Samsung back in, reinstalled Windows, and it's been running great since Saturday.

 

[xbassman]

Well, I think I am going to leave this box off until I get my video card.
No tivo this weekend I guess...

Too bad speedfan doesn't support my Biostar mobo so I could gain a little better fan control.
Right now they are running on the "fan only" connectors of my power supply.
Ya know I don't think I have ever noticed them spinning up any faster.

 

[Zim]

Originally posted by: ShadowknightI thought the drive was screwed, but after my Raptor spontaneously died...

It died because it was made by Western Digital, no other reason.

If you install a program such as hddhealth you can watch your drives slowly deteriorate over time. At higher temps the deterioration is considerably faster.

 

[Shadowknight]

The Raptor drive still powered up, it's just it no longer could be detected in BIOS.

I wasn't specific enough, but it was the Samsung that overheated. The Raptor was kept quite cool, I guess the data connection was just screwed.

 

[Zap]

Originally posted by: xbassman
Right now they are running on the "fan only" connectors of my power supply.
Ya know I don't think I have ever noticed them spinning up any faster.

I've noticed that on my Antec Truepower units also. Those PSUs seem really resistant to heat, and won't spin up the fans (it's own or the ones hooked up to "fan only") for nothing. All three of the ones I had did that.

I've encountered two HDDs that had minor issues from heat. One was a WD 5400RPM drive in an external USB box. It was hooked up to USB 1.1 and had a multi-GB transfer going on that took hours. Well, transfer stopped because HDD stopped working. Took it out of the USB box and it was too hot to handle. After it cooled down, worked fine again. Another HDD I encountered was in a client's computer. This was YEARS ago. I think it was a 386 system in a desktop case. The only cooling fan was in the PSU, and it died. Well, they had us check out the system because after it was on for a while it would make a horrendous squealing noise. Think of a slipping rubber fan belt in an old car, and imagine it right-in-front-of-your-face-noisy. Guess the drive bearings didn't like the heat. After the fan was replaced, the noise went away. Even with the noise it worked fine.

 

[GnomeCop]

Ive had my WD drives inside silencing enclosures and under extreme heat they never crashed

 

[albumleaf]

as hdd temp rises above 30C MTBF decreases exponentially.