Saturday, November 12, 2005

Athlon XP: Inexplicable Shutdowns

Original Post (HDD Dying): 10 November 2005, 11:00pm
I know which one it is. I have two hard drives in the "audio workstation" box. It's the Fujitsu. Every one of these that I have ever used has died. Maybe that's why they stopped making IDE drives? You think??

This Athlon XP box dual-boots between Win98 and Linux. Windows 98 serves a dual role. The boys can play games on it, draw, write stories, etc. AND I can mess around with the free version of ProTools (which apparently only runs under Win98, not NT or beyond for some unknown reason) and Kristal on it. The Linux partition is for the serious audio apps.

I started to notice it yesterday. One moment this box was running along in Win98... walk by it later and it would be completely shut off. At first I thought someone might simply be shutting it down more frequently then usual, but no, this was happening far too often. I run Folding@Home on it, so I try to keep it running non-stop anyway (even though the boys always like to shut down when they are done using the box).

I was starting to get suspicious, so tonight I kept my eye on it while we were watching TV. I thought it might be CPU thermal shutdown (I don't overclock and I have the automatic shutdown level set to the lowest temperature setting in the BIOS). I decided I would test my theory, and found a free system monitor/configuration app for Windows, SpeedFan 4.25. I installed it and set it to logging. A decent little application if I might say so, but it did not get much of a chance to run. I sat there watching the sensor readings fluctuate when all of a sudden I heard it, "ka-chunk" from one of the HDDs and poof! The box was off...

Well, now I know. I just hope the drive lasts long enough for me to clone the partition onto it's replacement. I'll just have to make sure no one turns it on until I can get to it. Man! I do not want to reload the dang thing from scratch again! It's such a waste of time. I can't complain though, I knew this day was coming when I put that drive in the box...

Update: 12 November 2005, 12:42am
Upon further investigation this evening I find that my initial impression on the cause of the shutdowns was correct after all: CPU temperature exceeding the thermal shutdown threshold set in the BIOS.

I gave the CPU heat sink/fan a thorough dusting with some canned air. The processor has been running at 100% (courtesy of Folding@Home) for quite a while, and right now it is sitting right around 57ºC. This is acceptable as the (Athlon XP 2000+) has a maximum die temperature of 90ºC. I currently have the BIOS shutdown threshold set at 75ºC. I will let the system run all night, and check the SpeedFan logs in the morning. I doubt it will go much above 60ºC based on what I am seeing now.

Looks like my Fujitsu drive will live to see another day. Still, a snapshot of the current partition image for backup purposes is definitely in order. That drive will fail sooner rather then later, I can count on it...

Update: 13 November 2005, 9:31pm
Thank goodness for automatic thermal shutdowns & warnings. I did some more testing and had the shutdown disabled but I had the warning set at 75ºC. The beeping started and I checked the temperature 80ºC! Shutdown and fast. Long story short(er), it was the CPU heatsink fan. I removed it, took the back cover sticker off, verified it was a ball bearing fan, cleaned it with DeoxIT, replaced the sticker, reinstalled the fan, and have been running it at 100% load ever since (again, courtesy of Folding@Home).

Fan Speed Avg. 5700RPM
CPU Die Temp. Avg. 47ºC

Oh so much better then the numbers I saw previously. If the fan holds up after this cleaning it should be good to go. If not, thermal shutdown is on and ready to kick in. Case closed...


No comments: