In <001401bf65b9$9e7fe440$80bdfea9@alnglenn>, on 01/23/00
   at 10:50 AM, Allen Glenn <[log in to unmask]> said:

>>
>> Fact: Silver grease is 10 (4.8 Vs 55 Btu/hr) times more conductive thermally
>> than zinc. IMO, a factor of ten is more than "slightly more". This greater
>> conductivity results in real world temperature drops of 10°+F at the CPU.
>> This amount may not seem significant at first, but let me walk you down the
>> path of real world.

>My room temp is 70F and the CPU temp is 76F using zinc compound.
>If I use silver compound how much of a drop in CPU temp should I
>see?

Well, the math is pretty straight forward:

Power flow(watts) = (thermal-conductivity)*Area/thickness Thermal
conductivity (TC) is defined as:
Watts/((temp-difference= Kelvins)(thickness= meters))

swapping everthing around; Algebra was so much fun:)

delta-T = Power*thickness/(area*TC)

Assuming a 486 sized package of about 2" square, and a grease or
pad thickness of .001 inch:

delta-T = .001*P/((2*2)*TC)

with conversion factors to get deg F:

delta-T(F) = 0.0177*P/TC

Assume a 50W cpu (Alpha maybe)

delta-T = .0177*50/.7 = 1.26 degrees F with zinc oxide delta-T =
.0177*50/8  = 0.11 degrees F with silver

Note that a lower power cpu (P-III) and bigger package (P-III)
will both lower the temperature. There is not much you can do
about the thickness. The idea is to fill in the micro bumps with
goo instead of air.

Personally, i'd use a pad, or non-silicon based ZO paste.

jan lambert



-----------------------------------------------------------
[log in to unmask]
-----------------------------------------------------------

               PCBUILD mailing list is brought to you by:
                            The NOSPIN Group
                  http://nospin.com - http://nospin.org