P5W DH form or function-NB cooling

Associate
Joined
5 Aug 2006
Posts
230
Location
in a well
Others have done this before: Removed pretty tinsel eyecandy from NorthBridge of the P5W DH and slapped a 40mm fan on it.

Well I did this today. I removed all the factory installed heatsinks and TIM from NB. cleaned up, removed the tinfoil golden prettyness and added a cheapo 40mm Akasa and fresh TIM.

Result 15°C reduction in NB temps from 50°C to 35°C at idle. Not only that CPU temps have dropped by 2°C as well. I guess that the NB was acting as a little radiator, hot air rises as we know, or should. The NB is below the processor when the board is mounted in a tower case. So I guess with the NB at 50°C the CPU cooler had to deal with heat rising from the NB as well as cooling CPU. It's the only explanation I can think of for the reduction in CPU temp. Anyhow, imho replacing the TIM on the chipset and adding a fan to the NB is a good thing to do with this mobo. :cool:

I will see if I can get a stable 3.33Ghz tomorrow.

What a cruel twist it is to allow the selection of a 550Mhz fsb, yet limiting the Voltage to the MCH to 1.65v :confused:
 
kitten said:
What a cruel twist it is to allow the selection of a 550Mhz fsb, yet limiting the Voltage to the MCH to 1.65v :confused:
If you think that's bad, how do you feel knowing theres a new revision of the board out already that allows 1.85v MCH?
 
Which cooler did you use, the Akasa AK-210 Blue LED Chipset Cooler?

What about cooling the voltage regulators or whatever they are, is that needed?
 
Jimi said:
Which cooler did you use, the Akasa AK-210 Blue LED Chipset Cooler?

What about cooling the voltage regulators or whatever they are, is that needed?

No blue LED's. Akasa DFS40102M. Plain Black.

I noticed that there is a thermal pad on the base of the HS that sits on the voltage regulators (about 2mm thick). Judging by the impression or lack of it, made by the regulators on this pad I decided NOT to replace. If I removed it I would need a 2mm layer of TIM to even make contact with the regulators. I don't have any thermal pads lying around.

Note that temps have now stablised at 40°C idle on NB. But CPU is still 2 deg cooler
 
Durzel said:
If you think that's bad, how do you feel knowing theres a new revision of the board out already that allows 1.85v MCH?

Mixed.. to get more out of CPU I think I need water cooling anyway. I get a stable 3.28Ghz 410Mhz FSB on air and get random crashes @ 3.36 420FSB both with 1:1 memory ratio. CPU temps are over 60°C under load @ 3.36Ghz.

If I would have waited 2 weeks I would have likely received the newer revision board.

Still I'm not complaining I have an E6400 system thats faster than E6800 @ stock speed.
 
The general consensus is that the bit of the heatpipe that covers the voltage regulators is just there to act as more surface area for the actual NB part of the heatsink. If you replace the NB heatsink you probably don't need to do anything about the vregs, they don't get hot at all.

The MOSFETs on the other hand get pretty damn hot.
 
Jimi said:
Oh, I got mine last week and it's one of the ones with only 1.65MCH.

I suppose I wouldn't have done then :) I first read of the new revision boards about 2 weeks ago. I guess retailers will use up any old stock first.
 
Back
Top Bottom