Does thermal paste need to "break in" or "set"?..........CPU temps kind of high

[iamme]

I just got a new HSF for my Athlon 2100+, and my temps are a little high. Just wondering if I need to let it set. If so, how much lower can I expect?

The HSF was rated for Athlon 2200+ XP's. When I overclocked my CPU to 2.0GHz (2400), the idle temp was 50 degree C

 

[Creig]

Some thermal paste manufacturers suggest a "setting" period of a few days. Such as Arctic Silver. But if you're overclocking your processor, you could run into temps in the 50s at idle. There are so many different factors that it might be hard to isolate a single cause. Case ventilation/temps, CFM of fan, heatsink design, contact with CPU die and even the CPU itself might run hotter than others of the same speed.

 

[KuRnUp]

ok, yea, that may be a little high, but 2200+'s i dont believe to be very good overclockers. im not too sure about that. anyways, to the main issue: thermal paste. just put it on, and place the hs on after you do it (some manufacs do infact give a setting period, just not all). over time, it will set in, and over some time, you will notice the temp going down slowly. it does need to set to increase the preformance, but it will set as you use it over time.

hope that helps.
~KuRnUp

 

[txxxx]

Originally posted by: KuRnUp
ok, yea, that may be a little high, but 2200+'s i dont believe to be very good overclockers. im not too sure about that. anyways, to the main issue: thermal paste. just put it on, and place the hs on after you do it (some manufacs do infact give a setting period, just not all). over time, it will set in, and over some time, you will notice the temp going down slowly. it does need to set to increase the preformance, but it will set as you use it over time.

hope that helps.
~KuRnUp

He's running a 2100 , so T-Bred B most likely. Depends more on the mainboard itself and how its reading the temperature too. My MSI nForce2 reads the CPU die (53c average) and CPU socket average (42c average). I have the same processor running at 1.775v , 2.17ghz (just standard 166fsb crankup, fast enough ). This is on a silent Zalman flower cooler (CNPS-6000).

Edit : Those are temps on load, temps without load are typically 4-5c lower.

 

[McCarthy]

Yes, some need break-in time for benchmark tests....you know, where one paste is declared outstanding and another to be useless because of a 1 degree difference.

The break-in usually results in a 0 to 3 degree difference with .5 to 1 being more common.