Highest safe temperature limits for the i7 & PCH

PCH ~ 111C, CPU ~ over 80C will be unstable; before damage >90C ~ I don't temp(T) fate (pun).

Hi Jaquith,
Thanks for the info. That's great.
I didn't realize the PCH could take such a high temperature! That's hot! The P55A-UD7 motherboard does report the PCH temperature on the MIT BIOS page.

Do you know if there are any fan modules that can be used above the PCH to keep the temp down on the GB Motherboards? Or if they are ever necessary.

The board does come with that optional vertical fin which covers up PCI-E #1 slot, but the review reports show it only drops the temp about 2 degrees centigrade.

{Like you I had an issue w/GA MOBO for new office pseudo workstations; the MOBO's I am most familiar with and personally build on are EVGA} meaning there are all sorts of water-blocks for this purpose; Koolance. The UD7's are typically cooled well.

Assuming I was running a P55 to any limit that would require additional cooling for the PCH chipset then "air" cooling wouldn't enter my mind instead "everything" would be water cooled.

If I may ask, what OC are you targeting?

Hi Jaquith,
I am not targeting any specific high end overclocking. I am just finishing a nice higher end system with a Silverstone FT-01 aluminum case with 230mm fans, a GA-P55A-UD7 MB, an 875K unlocked CPU, 4 sticks of fast Dominator RAM, & an XXX 5850 video card. Just trying to learn the limits of these components and to understand how to manage the system.

The PCH on this board is ready for water cooling, so they did engineer it nicely. I see just loafing, the PCH does report 6 degrees higher temp than the CPU, but using a fluke IR digital thermometer, the reported temp is actually on the heat pipe near the water cooling port, rather than the top of the PCH which is about the same temp as the CPU.

I may be looking at it from more an engineering point of view, with that old adage "if you can't measure it, you can't manage it"! And of course, the more you understand your components & rig, the more likely you are of not becomming part of a problem.

The problem is two identical PCs can have ±5C of stable tolerance, either via bin or random assembly which can be frustrating. This isn't typically an issue for most builds until you push the limits (BSOD or post failure) when stressed vs idle.

Meaning, there are no ±1C empirical answers - there are too many components - there are no absolute tolerances - there are spec'd limits (below "true" limits but these are pushed as soon as you OC. I've built 1 rig that OC 4.3 with specific settings, but the next fails.

Thanks for all the great info everyone.

Yes, Bill, that graph is just what I have been searching for. Will copy it and the accompanying table to put with my i7 data.

The temperature profiles have been readily available for older CPU's, just couldn't locate the i7 spec sheets.

Best answer selected by John_VanKirk.

In case you want the whole set, the other pair:
Intel® Core™ i7-800 and i5-700 Processor Series Datasheet - Volume 1
and
Intel® Core™ i7-800 and i5-700 Processor Series Datasheet - Volume 2

I don't want anyone to think I pulled a number out of the "air"; the temps were from Intel and not someone's Google'ing.