Ran a CPU without a working fan for ~20 hrs

[Praetor]

I had to replace a faulty HDD yesterday on my file server/HTPC backend. Apparently, I knocked one of the power leads to the cpu fan out of place and it wedging the fan in place.

It's been up and runing for a few minutes over 20 hours. I just happened to be checking the dmesg log and saw a number of

CPU1: Temperature above threshold
CPU0: Temperature above threshold
CPU0: Running in modulated clock mode
CPU1: Running in modulated clock mode

messages. I thought maybe it had something to do with updating the kernel and a new setting I had missed, but again on a whim, I opened up the case and noticed the fan not spinning.

Should I shut the machine off now? Or leave it running? This is a 2.6 p4 on an Intel 865PERL board.

 

[ScrapSilicon]

shut the machine off now ...replace fan <edit> reconnect fan's pwr lead

 

[Praetor]

Originally posted by: ScrapSilicon
shut the machine off now ...replace fan

to prevent the chance that the motor spinning the fan could have burned itself out trying to rotate? .... It's working right now, for what it's worth.

 

[MagnusTheBrewer]

Reconnect the fan. If it ran this long long without going up in smoke it's probably ok and you just took some time off it's lifespan. I do have a question though, are there no thermal warning or shutdown points set? Why not?

 

[Praetor]

Originally posted by: MagnusTheBrewer
Reconnect the fan. If it ran this long long without going up in smoke it's probably ok and you just took some time off it's lifespan. I do have a question though, are there no thermal warning or shutdown points set? Why not?

I had to replace the primary drive yesterday due to continued crc write errors. I restored an older backup onto the new drive and just compiled a new kernel and started the basic services to get me through to today.

Thus, no cron jobs setup to check these things and shoot an e-mail to warn me.

Looking through the timestamps, the messages were posted over a 50 minute period while I was upgrading mysql and the databases this morning remotely. I don't have any other physical audible warnings around the case, and the only visual clue I would have is on the case-display, which I hadn't bothered to look at earlier.

So. Should I start crying now and look for a new CPU or is the general consensus that the chip should be fine?

 

[Bozo Galora]

um, why would you even consider leaving it running with no fan?

your cpu is fine, the messages show it was throttled down INTERNALLY - modulated clock mode (prob several levels) due to heat

a CPU with no fan OR HEATSINK will violently fry itself in about 5 seconds

 

[Ruptga]

Originally posted by: Bozo Galora
um, why would you even consider leaving it running with no fan?

your cpu is fine, the messages show it was throttled down INTERNALLY - modulated clock mode (prob several levels) due to heat

a CPU with no fan OR HEATSINK will violently fry itself in about 5 seconds

Violently? Is there a video of this somewhere, it sounds interesting.

Anyway, I agree, it's probably fine since it's still working.

 

[Bozo Galora]

http://video.google.com/videoplay?docid=5393904704265757054

notice if a chip has throttling - nada
all depends on brand, chip type and if throttling is software or hardware, enabled or disabled
notice game playing on system in background

http://video.google.com/videoplay?docid=3336933539640249505

 

[tbike06]

Originally posted by: Bozo Galora
http://video.google.com/videoplay?docid=5393904704265757054

thats fake.

 

[Bozo Galora]

yeah, picked wrong one - its been a long time

 

[cryptonomicon]

ROFL, they put the heatsink back on all the Intel CPU's and they let the AMD ones BURN..

hahahaha

 

[Praetor]

Originally posted by: Bozo Galora
um, why would you even consider leaving it running with no fan?

your cpu is fine, the messages show it was throttled down INTERNALLY - modulated clock mode (prob several levels) due to heat

a CPU with no fan OR HEATSINK will violently fry itself in about 5 seconds

(emphasis added by me)

I wasn't planning on leaving it running. When I first posted, I had already noticed the wire blocking the fan from spinning and had moved and tucked it back into place. The only time the fan wasn't running was between the time I first closed the case ~20 hours previous and when I inspected the logs.

I'm not a hardware guru, I've heard stories of people permanently ruining their processors doing stupid ****** like this. Thankfully, I have more than adequate ventilation and cooling for this case and I suspect that may have helped out as well. I needed confirmation on whether or not the cpu was going to survive my blunder.

Thanks for the kind words and advice guys!

 

[lxskllr]

The chip should be fine. Just keep running as usual. It protected itself the way it's designed to. My old Dell (P4 2.4gz)didn't even have a cpu fan. It was shrouded and had a fan that sucked air past the heatsink, and out the back. It got pretty warm on occasion, but it's still powering my daughters machine.

 

[bob4432]

i am assuming it is a intel cpu as they have/had better throttling down due to heat. but you may want to check in the bios to see if there is an option that says if the cpu fan is lower than XXXX rpm, it will beep or not boot at all.

since it worked for 20hrs, i would say you are ok, but just in case you may want to run a couple of stress test like prime95 to make sure some silicon didn't get f*cked up

 

[ch33zw1z]

Originally posted by: cryptonomicon
ROFL, they put the heatsink back on all the Intel CPU's and they let the AMD ones BURN..

hahahaha

I'm guessing it's because intels cost way to ****** much!

Praetor: if it's still working, should be fine for now. may not last as long as it would have.