Pro/Forums - View Single Post

BillA · 01-01-2001, 03:20 PM

The following is from a post at overclockers.com.au
This investigation was done with respect to cold plates but is somewhat applicable to HSs as in almost all applications the HS is significantly larger than the die.
Take a look at the Thermacore site for a discussion of heat spreading resistance (and their vaporware Athlon and Pentium HSs).

I finally found a heat spreading resistance calculator at http://www.mhtl.uwaterloo.ca/tools.html# and the results were enlightening.

Assuming a heat source of 11*11mm (120mm sq Athlon) applied to an isotropic cold plate of 48*48mm, I calculated the spreading resistance for the thicknesses 1/8, 1/4, 3/8, and 1/2in.s (3.18, 6.35. 9.53, and 12.7mm) in aluminum, copper, silver, and (my favorite) Cusil (72%Ag & 28%Cu).

The results:
Matl . T.C. . 0.125 0.250 0.375 0.500 in. thick

Al . . 220* . 0.259 0.170 0.149 0.142

Cu . . 388**. 0.147 0.096 0.084 0.080

Ag . . 418 .. 0.136 0.089 0.078 0.075

Cusil . 515 .. 0.111 0.073 0.064 0.061

*This is the thermal conductivity of PURE Al, so the actual spreading resistance values are probably 20% higher than shown.
**for Alloy 11000

What can be understood from the numbers:
1) For all materials (and a given aspect ratio) there is a thickness beyond which the spreading resistance decreases only slightly.
2) Alumimum is a poor choice for small hot dies.
3) A 1/4 in. copper cold plate has 25% LESS spreading resistance than a 1/8 in. silver plate !
(and a 3/8 in. Cu will still beat a 1/4 in. Ag !!)
4) The Cusil values are very good, but the price may preclude all but absolute fanatics.
5) The commonly provided 1/4 in. copper is probably the most cost effective cold plate, though I'd go for 3/8 in. just to beat out silver.
(sorry GOD)

A note on the "Film Coefficient - h (W/m sqK): the value used was "2" as found in the literature. Also ran with 200 and 20000 and while the calculated values changed, the ratios did not.

Someone good with graphs, I'm not, might plot the curves and post it for the group.

be cool

[This message has been edited by BillA (edited 01-01-2001).]

01-01-2001, 03:20 PM	#34
BillA CoolingWorks Tech Guy Formerly "Unregistered" Join Date: Dec 2000 Location: Posts: 2,371.493,106 Posts: 4,440	The following is from a post at overclockers.com.au This investigation was done with respect to cold plates but is somewhat applicable to HSs as in almost all applications the HS is significantly larger than the die. Take a look at the Thermacore site for a discussion of heat spreading resistance (and their vaporware Athlon and Pentium HSs). I finally found a heat spreading resistance calculator at http://www.mhtl.uwaterloo.ca/tools.html# and the results were enlightening. Assuming a heat source of 1111mm (120mm sq Athlon) applied to an isotropic cold plate of 4848mm, I calculated the spreading resistance for the thicknesses 1/8, 1/4, 3/8, and 1/2in.s (3.18, 6.35. 9.53, and 12.7mm) in aluminum, copper, silver, and (my favorite) Cusil (72%Ag & 28%Cu). The results: Matl . T.C. . 0.125 0.250 0.375 0.500 in. thick Al . . 220* . 0.259 0.170 0.149 0.142 Cu . . 388*. 0.147 0.096 0.084 0.080 Ag . . 418 .. 0.136 0.089 0.078 0.075 Cusil . 515 .. 0.111 0.073 0.064 0.061 This is the thermal conductivity of PURE Al, so the actual spreading resistance values are probably 20% higher than shown. **for Alloy 11000 What can be understood from the numbers: 1) For all materials (and a given aspect ratio) there is a thickness beyond which the spreading resistance decreases only slightly. 2) Alumimum is a poor choice for small hot dies. 3) A 1/4 in. copper cold plate has 25% LESS spreading resistance than a 1/8 in. silver plate ! (and a 3/8 in. Cu will still beat a 1/4 in. Ag !!) 4) The Cusil values are very good, but the price may preclude all but absolute fanatics. 5) The commonly provided 1/4 in. copper is probably the most cost effective cold plate, though I'd go for 3/8 in. just to beat out silver. (sorry GOD) A note on the "Film Coefficient - h (W/m sqK): the value used was "2" as found in the literature. Also ran with 200 and 20000 and while the calculated values changed, the ratios did not. Someone good with graphs, I'm not, might plot the curves and post it for the group. be cool [This message has been edited by BillA (edited 01-01-2001).]