• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

20 degrees difference between two Celeron D CPU's?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Mastiff

Member
Joined
Mar 7, 2003
Location
Norway
I have bought new setups for my kids, with Elitegroup P4M800-M7 mobos, 512 meg Twinmos memory and Celeron D. The shop only had one Celeron 3.2 left, so I got one of that and one 3.0. Both use the boxed cooler, a stupidly placed component makes it impossible to use any cooler with a H-shaped mounting bracket. Without load the 3.0 is around 45 degrees in a cool room, while the 3.2 is around 67 degrees in a 10 degrees hotter room. Is this to be expected, or is there something very wrong here? I have Arctic Silver 5 applied identically on both systems.
 
yes because there is increased room temp and your pc has hotter air being pushed around inside it. these are idle temps are they?
 
Yep. And man, you're fast! I'll take the PC into another room and see what happens.
 
Nope, that didn't help much. Went down to 63 degrees in a rom that's 20 degrees (68 F). Sounds far too high for me. The scary thing is that the aluminium fins on the boxed doesn't even feel warm to touch. Shouldn't it be when the CPU's that hot? I have only been running this for a couple of hours after the last Arctic Silver change. Is that perhaps the reason? The other one has probably run around 30 hours or so.

Edit: I forgot to mention that I had problems with the memory on that same mobo. It didn't work so I took to sticks from another computer, and they worked as they should. And the original memory from that mobo worked on the computers where I took the other two sticks! Can that be related? Also it seems to me like that boxed cooler is bending the mobo a lot.

Edit 2: OK, now I'm confused! I installed SpeedFan 4.27, which should support this chipset, and that shows only 47 degrees on the CPU! The above mentioned numbers are from the BIOS (latest BIOS for the mobo). What is this? :bang head
 
Last edited:
But if it wasn't, wouldn't the system fail on Prime? I have been running Stress Prime for a while now, and SpeedFan says it's running at 66 degrees, but it's not failing.
 
If the memory was generating a lot of heat it could be contributing to the issue. Man those stock HS's must suck. My dual 3.2 xeons are running at 35 degrees C.

Try the 3.2 CPU in the other board.
 
Also there's something wery weird about this stuff. I just used the reset button and went straight into the BIOS setup. After 15-20 minutes of Stress Prime, that made the temp in Speed Fan go a lot higher, up to 67 degrees, the temp in the BIOS is still only 64 degrees. The process took less than 10 seconds, and I don't believe that the temp falls that fast. Can this be a BIOS/monitor problem?

Edit: Ebola, this was on no load what so ever. The memory wasn't even lukewarm to touch.
 
Not to insult ur intelligence -

But you did remove the stock TIM B4 applying the AS5, right?

Figured it couldn't hurt to bring it up.

Strat
 
Yes, of course. I'm not insulted I know that total newbies probably wouldn't. But I don't go to extremes either, I don't lap my heatsinks or CPU's. ;) Anyway I need to look at the appliance method. I used the old fashioned credit card method, but should I have done the "one grain of rice in the middle and press down" method for a CPU with a heat spreader? Is it possible that the layer of AS is too thick?
 
The artic silver could be too thick. that caused a problem with my comp but reaplying it correctly sorted it out. Try out the method that is recommended. The "dab in the middle then press down". Hopefully doing that will solve your problems.
 
I will. It's a pain in the ***, though. This board is so soft it'll bend when I fasten the CPU cooler, and I have to do it when I hold it, so I can support it from the back instead of doing it with the board screwed in the case, like the manual for the boxed cooler says. Idiotic fastening system! Intel should learn to use H-brackets, so all mobo manufacturers had to leave room for them.
 
Use the "rice grain" size dollop method...it really works.

It's amazing how little is required.

Also, if you can snug the HSF down, & rotate it 10 - 20*, back & forth, once or twice, it's a good idea too (not sure if you can do that w/the stock HSF ***'y).

I also "buff" some AS5 into both mating surfaces, first, before applying the "dollop". It's definitely not a layer, but more like just enough to lightly "cloud" the surfaces. Not even enough to leave much of a film. Not sure if anyone else does this.

But, yeah, a thick layer of AS5 can actually act as a thermal insulator, of sorts. Can't say that's ur problem for sure, tho.

GL,

Strat
 
Temp in the BIOS after an hour or so in a 20 degrees room and the grain of rice method is 62-63 degrees. So it did help a tiny bit, but didn't really help the main problem. I'll have to wait until tomorrow to try the other CPU, it takes some time to rip it out of the computer that I have it in now. In the meantime I'll let this one stay on the BIOS temp monitor page overnight to see if that helps the temp a bit when the AS settles. Don't believe it will make that much of a change, though.
 
OK, the computer has been there over night. It has gone down a few degrees in the BIOS, from 64 to 62. But now to the really, really confusing stuff: On the other identical mobo with a CD 3.0 I see the same temp in SpeedFan 4.27 as in the BIOS, around 45 degrees. On the problem mobo with the 3.2 I see 49 degrees. If I change from diode to thermistor, they both go up to around 75 degrees, but I think diode is the correct setting. That's what it finds automatically. Now here comes the real fun part:

If I set the BIOS to allow the CPU fan to go slower when the CPU temp is below 75 degrees and let SpeedFan be at Diode, the fan slows down. If I set the sensor in SpeedFan to thermistor, the fan speeds up again. And this is not because of SpeedFan, because I haven't set SpeedFan to automatically vary the fan speed. I think this is a case where the BIOS is reading the temp totally wrong, and where SpeedFan is correct, that this is running at around or just below 50 degrees. Does anybody agree?

Oh, the VCore in SpeedFan is shown at 1,33 on the hottest computer and 1,36 at the coolest computer.
 
Back