Thermal dissipation when OC'ing CPU

[soccerballtux]

Limiting factor on my current e2180 OC seems to be heat (well, I've hit the voltage wall anyways, I'm at 1.5075 in BIOS (1.48 in CPU-Z)).

Now the die on the C2Q is like this:
http://www.cnet.co.uk/i/c/blg/...esktops/core2quad2.JPG

The die on my C2D e2180 is half of that size.

I've touched the heat pipes on my TR-Ultra 120 when my processor is as hot as it gets (absolute max is 74C, from Orthos Large), and they're only a little warm to the touch. So clearly I'm getting some thermal dissipation limits in the cooler I think. Or bad contact with the CPU. Ya I've applied the paste right, and lapped my cpu + cooler to no avail.

SO I guess I wanted to make sure, I wouldn't notice a huge temp spike (and my Ultra-120 wouldn't be overloaded) with an OC'd Quad, reasoning being while there may be double the heat coming out, there is double the surface area, so the rate of transfer of heat down the pipes will be double.

Simple thermal dissipation I guess, just want to check the logic behind it.

 

[VirtualLarry]

My E2140s at 3.2ghz, 1.425v bios 1.36v cpu-z, run up to 9C from tjmax. I dont know what freq your 2180 is clocked at, but 74C sounds like an amazingly good temp.

 

[RussianSensation]

Originally posted by: VirtualLarry
My E2140s at 3.2ghz, 1.425v bios 1.36v cpu-z, run up to 9C from tjmax. I dont know what freq your 2180 is clocked at, but 74C sounds like an amazingly good temp.

I don't know about that. My Q6600 @ 3.4ghz w/ Tuniq gets 63-65*C on the hottest core in CoreTemp and I consider those temperatures OK at best, nothing amazing. So 74*C on any dual core is unacceptable on a TRU. I would expect a good cooler to cool a 3.2ghz E2140 at 53-54*C max. My old E6400 @ 3.4ghz never exceeded 52*C at load with an inferior Big Typhoon. So i am not buying the argument that because there is 1 die vs. 2 in the quad, the pipes are less effective, resulting in 70*C+ load temps. You have the heatspreader anyway that helps to dissipate the heat from those cores.

I say you reseat your heatsink one more time, or reapply your thermal paste. Also, those voltages are probably the culprit for this (i.e. let's say if your system is fully stable at 3.15ghz with just 1.40V, your temps might fall 6-8*C).

 

[bryanW1995]

that has nothing to do with his voltages, it's contact issues. my e6750 I ran at over 1.50 v in bios for 9 mos at around 3.5ghz, and it never got above 60c.

 

[VirtualLarry]

Originally posted by: bryanW1995
that has nothing to do with his voltages, it's contact issues. my e6750 I ran at over 1.50 v in bios for 9 mos at around 3.5ghz, and it never got above 60c.

Ok, I guess I'm wrong then. I thought that his high voltage (1.5v) was causing his high temps. On my chips, I can't even imagine how high the temps would get if I pumped that much voltage into them. I'm using CoolerMaster HyperTX2 heatsinks btw, which are fairly mediocre.

 

[sonnygdude]

It could be a combination of both. Assuming similar current draw, higher voltage = higher power dissipation. For heat transfer through conduction (Q=kA/L(delta T)) and convection (Q=hA(delta T)) you can see the temperature difference from recovery temp to source of power dissipation is linear with repsect to power. In other words, if you raise your power by 10% expect the temperature difference between ambient and your source to go up by 10%.

Inefficient thermal connection between the HSF and CPU adds contact resistance, which can also force delta T up.