Thermal Paste working its way out due to thermal expansion rates

[Harry Muscle]

I've read that according to AMD they don't recommend using thermal pastes
because they work themselves out from in between the heatsink and the CPU
due to different thermal expansion rates of the two. Unfortunately they
don't say how long it takes for this to occur. I don't do any major
overclocking (66 Mhz overclock on a Athlon XP 2500+), and am currently using
the original HSF and thermal pad that came with my CPU. However, during
full load (Prime95) my CPU Diode hits 68C after two minutes. I'm thinking
of maybe upgrading to a better heatsink (like the Vantec AeroFlow or the
Thermaltake Silent Boost ... don't like noise) but I'm worried about this
phenomenon of thermal paste working it's way out. I don't want to find my
CPU overheating one day, cause that will be the day when I don't have MBM5
running and nothing will catch the problem. I also don't want to find
myself having to reapply thermal paste every few months. I'd like to not
open this maching up for the next two years. Anybody got any idea of the
time lines that we're looking at for the thermal paste to work its way out?

Thank you for any help,
Harry

 

[Anthony R]

You have nothing to worry about. The paste will do it's job just fine. I
had an athlon 1ghz Tbird with plain cheap ass thermal paste for almost 2
years and it never had a problem.

Infact, when I went from the thermal pad to plain thermal paste on my Athlon
2600 + (333mhz fsb) I shaved off about 4 degrees.