Chipset temperatures

markjia · Post by **markjia** » Sun Jul 25, 2004 12:09 am

I have a 2.8 P4 on an ASUS P4P800, and I was wondering what safe temperatures are for the chipset.

Here are my temperatures:

Idle:
CPU: 43C
HD: 45C
Chipset: 48C

w/ CPU burn:
CPU: 64C
HD: 45C
Chipset: 54C

I think that the last temperature is from the chipset...am not certain, but that's my guess (could it be the case?).

Anyways, is this normal? safe? I also noticed that the PSU (NX3000) speeds up when I put it under these loads.

trodas · Post by **trodas** » Sun Jul 25, 2004 1:52 am

Yes, safe. 58C is not that high. My SB is same hot as your NB and no problem so far. Your PSU speed-up, because it's fan are temperature controled. More temp - more RPM - more noise, of course too

shathal · Post by **shathal** » Sun Jul 25, 2004 3:44 am

Trodas - what the heck do you do to you SB to get it that high?

What sort of traffic are you shoving down there?

Post by **Ralf Hutter** » Sun Jul 25, 2004 5:59 am

1) Intel rates the max temp of it's NB chipset as 99°C. I've linked to their pdf spec sheet several times, if you search the forums you can find it.

2) The P4P800 seems to give somewhat warmer than normal NB temp readings. Rumor has it that the sensor (which actually isn't on the NB itself) is in a place on the board with poor airflow.

______________________________
Bottom line: I wouldn't worry about your NB temps unless you can show that it's making your system unstable.

EvilNick · Post by **EvilNick** » Sun Jul 25, 2004 7:43 am

I'm measuring my NB temps with a Cooldrive4 probe stuck under the heatsink and the tip in contact with the raised NB core.

Mine are: Idle-41C, Load-46C
However, I'm using a Thermalright NB-1 and lemme just say that the 45mm Delta is far from quiet.

wumpus · Post by **wumpus** » Sun Jul 25, 2004 10:57 am

I was wondering this myself recently, and I found the same spec sheet that Ralf quoted for Intel's NB temps: max operating 99c. So I stopped worrying about my NB load temps of ~65c with the 486-class heatsink.

What worries me more, is the voltage regulation circuitry temps. Without any directed airflow, on newer/faster systems, these parts get to over 100c (!) which kinda... scares me. From HardOCP:

We used two ABIT boards that we had available. The AI7 and the IC7-MAX3. Both boards passed stock CPU speeds stress tests at 3.2GHz (Prescott CPU). We were working on stressing these boards before we got our 2.8E so we had days of stress testing done when the 2.8E arrived. All of the motherboards we have used with Prescotts run hotter than we are used to. It is not uncommon to see components on the motherboard at 250F/120C degrees.

The IC7-MAX3 did do something that we have never seen before. If you have noticed, we have used square blocks of plastic foam to elevate our boards for years now while we are testing them. The IC7-MAX3 got so hot that it actually melted the plastic block into a liquid form.

What you are seeing above is the plastic stringing and cooling as the board was pried loose from the block. Amazingly enough though, cutting off the motherboard circulation of cool air did little to hamper its stability. It ran for more than two days in very adverse conditions.

My system is an Athlon FX-53, but the temps on the VRM circuitry are very similar to the ones quoted for the Prescott systems.. as measured by my laser temp gun, over 100c under load.

ghowarth · Post by **ghowarth** » Sun Jul 25, 2004 4:12 pm

My ASUS VIA chipset results in BSODs if it gets over 40C

smifou · Post by **smifou** » Sun Jul 25, 2004 10:58 pm

i ranned my kt400 at 60-68 C no probz

markjia · Post by **markjia** » Mon Jul 26, 2004 10:15 pm

Thanks everyone! I guess I won't worry about my temps.

dago · Post by **dago** » Tue Jul 27, 2004 4:57 am

Well, if you want to consider the ads google delivered on top of your topic, ...

Mellen High Temp Furnaces : Small and Large High Temp Furnaces 200C to 1850C Standard and Custom

Laboratory Furnaces : Electric Fired High Temp to 1800 C Elevator Hearth for operator safety

TvU Laboratory Furnaces : Low cost, efficient vacuum, inert gas, transparent, hi-temp, more -

High temperature furnaces : Furnace Concepts provides off-the- shelf furnaces ovens and parts.

sthayashi · Post by **sthayashi** » Tue Jul 27, 2004 7:27 am

Um... does anyone else think that Markjia's processor is running a bit hot and is probably doing some speed limitng? Especially since it's been reported that the P4P800 underreports temps?

Tibors · Post by **Tibors** » Tue Jul 27, 2004 8:52 am

No, I don't think so.

I have a P4 3.0E om a P4P800-VM. It runs 64°C too. It runs Folding@Home 24/7 with the same average PPW as when I kept it running at 59°C.

wumpus · Post by **wumpus** » Tue Jul 27, 2004 9:00 pm

P4's won't throttle until about 74c. Preshotts won't throttle until even higher, over 80c. There's an article about this on digit-life:

http://www.google.com/url?sa=U&start=1& ... ng/&e=7620

wumpus · Post by **wumpus** » Tue Jul 27, 2004 9:00 pm

wow, freaky, google is now inserting their URL into the results?!?!?! Weird!!

http://www.digit-life.com/articles2/p4-throttling/

sthayashi · Post by **sthayashi** » Tue Jul 27, 2004 9:11 pm

Well, remember that Markjia has the Northwood, it reaches 64C w/ CPUBurn on a board that's known to underreport. That throws a red flag for me, but if it's not throttling then there's no need for concern. Sorry to raise a stir.

markjia · Post by **markjia** » Tue Jul 27, 2004 10:53 pm

Is throttling controlled by some internal sensor in the CPU? or does it rely on the mb?

Jan Kivar · Post by **Jan Kivar** » Wed Jul 28, 2004 6:37 am

markjia wrote:Is throttling controlled by some internal sensor in the CPU? or does it rely on the mb?

To the best of my knowledge the processor enables PROCHOT# signal, which makes the board drop the clock speed by predetermined amount. There is a separate temp sensor in the PIV core for this, and it's not accessible by user.

Interestingly, this (activating the Thermal Control Circuit, or TCC for short) could also enabled by hand (read the datasheet). I DO think that a hacked BIOS would be required for such an operation, though.

Once the processor gets TOO hot (135°C, IIRC) THERMTRIP# signal is enabled in the processor, which requires the board to cut all power to the processor within 500 ms.

Cheers,

Jan