PDA

View Full Version : How cold until PIII Core might crack ?


zoopa_man
07-18-01, 08:36 AM
I just added a 80w peltier to my PIII 733 and I love the thing to death. With my DD waterblock and cube radiator I am geting temps below zero !! My temp probe that I installed touching the die reads anywhere form -10C ilde to 3C at full load. The bios reads -2C when I boot up. My software monitoring program for my gigabyte motherboard just freaks out and reads 250C and panics, so I had to shut that off. But anyways...

I wanted to know just how cold my processor can get before I have to worry about the risk of the die cracking? I'v heard of this happening but don't know at what point the temp becomes a real factor. Could someone help me out ?

*edit* Is the spell check not working today? It just hangs up, the spell check screen pops up but then just sits there and says Loading... but never starts up.

Kingslayer
07-18-01, 10:25 AM
I don't have a clue at what temp it cracks. But I think it's from expansion and contraction due to the temp change.

Anyways, that cold is pretty good coming from only an 80W pelt. Even if it's on a PIII.

zoopa_man
07-18-01, 10:33 AM
So If the cracking comes form the core expanding and contracting which makes sense, then should I leave my water pump and peltier running all the time, or is it ok to shut it off when I shut down my system ?
I have noticed that as soon as I shut the system down and turn off the peltier that my cpu probe temps start going up rapidly to the point where they reach the water temp in the block, which is right arround 40C - 50C This sounds bad to me, so should I just leave the Peltier and water pump running all the time or is it safe to shut the sytem off and let it heat up like that ?

Pinky
07-18-01, 11:09 AM
I don't think that room temps will cause the chip to expand that much, but over time the constand 20-30 C change from turning the Peltier on and off could cause small fractures in any material, and I would assume the same chemistry laws would apply here.

Maybe try turning down the cooling, have it run at 15C and 20-25C under full load, that way you can turn the Peltier off and still keep it relatively cool.

I don't know that there's much benefit with your setup having the chip run at subzero.

NeoMoses
07-18-01, 03:29 PM
post # 4 brings up an interesting point. why run temps that low? I would be very interested in seeing benchmarks at like 5 degree C increments, from around 45 C to 0 C. That's data that I've been looking for for a while. I think many would be interested, and would make a really kick-@$$ article. Hope you do it.

zoopa_man
07-18-01, 03:33 PM
I just might do that. If I can find a large LARGE enough rheostat, I could hook it up to the pelt to give me full contorl over cpu temps. Then run some benchmarks. That's a great idea. I bet O/C.com would love to post something like that for me.

Now where To find a rheostat for this monster. Here's the specs on it.

80W At 12V it draws 6A So what size rheostat ?

outhouse
07-18-01, 08:01 PM
Well as far as how cold, well they have done test and found out that at 100 below 0 that a CPU will run twice its rated speed i'm not exactly sure its 100 below but its very close to that it was definatly under 50 degrees below 0 but i'm pretty sure the article said 100. So I would think that at your temps you run absolutly no risk of getting your CPU too cold, just think about the guys runing liquid nitrogen and the guy who used dry ice to cool his CPU. As far as your temps going up and down these CPU's are made to handle changes in temps and they do an amazing job at this to boot, so again with your set up I would not worry I have yet to hear about a peltier getting a CPU so cold it cracked unless it failed or unless uneven pressure or somehow you had a instant change in temps and what i mean by that is if you took something 50 below 0 and touched it to a 35C CPU.

goodluck I wish i had your problem :)

JigPu
07-18-01, 10:39 PM
I wouldn't leave the cooling system on all the time running at full blast. I would turn it down though. If you leave the system running at full blast all the time, I would think that because your CPU isn't putting out any heat anymore, it would cool down greatly, and would probably have the same effect as turning off the cooling with the computer. I don't do any super cooling so I dunno, but it makes sense to me! As for the rating, it would have to handle at least 80W to turn it off completely.

JigPu

outhouse
07-18-01, 11:48 PM
zoopa_man (Jul 18, 2001 10:33 a.m.):
So If the cracking comes form the core expanding and contracting which makes sense, then should I leave my water pump and peltier running all the time, or is it ok to shut it off when I shut down my system ?
I have noticed that as soon as I shut the system down and turn off the peltier that my cpu probe temps start going up rapidly to the point where they reach the water temp in the block, which is right arround 40C - 50C This sounds bad to me, so should I just leave the Peltier and water pump running all the time or is it safe to shut the sytem off and let it heat up like that ?

If your worried about it leave it running for a minute after you shutdown and then there will not be as much of a tempature change.

zoopa_man
07-19-01, 08:43 AM
Thanks for all your thoughts and sugestions. Wouldn't it be nice if everyone was worried about -10C temps LOL I'v never used a peltier and had only a slight idea of what to expect. So I set my standards low thinking I would get temps arround 3 to 10 C I was suprized to see my temps hit -10C Peltiers are real cool devices. Pun intended ??? Now if I could find a electronics shop with a big rheostat or trimmer I'd be all set.