Silver Thermal Grease Test

Add Your Comments

Following is a report of a test on Circuit Works Part No. 7100,
Conductive Silver Grease.

OBJECTIVE

The objective of the test was to establish the products’ suitability for use as a heat transfer interface material between PC microprocessors and heatsinks, with a singular consideration for the products’ ability to resist flow and not contaminate nearby electronic equipment and so cause short circuits.

CONCLUSIONS

THE PRODUCT DID NOT FLOW OUT OF THE TEST JOINT under conditions more severe than will be encountered in normal or processor overclocked operation. It will be an excellent heat transfer interface material because of its superior thermal conductivity; 11.5X higher than a currently popular SiZnO grease. This higher conductivity translates into a 7 to 11C reduction in temperature drop across the interface and therefore to the processor internals as well. A drop of the silver grease deliberately placed in an open area did flow 5/8 in. after 80 minutes however, which indicates a need for some care in application and perhaps a post assembly cleanup of any joint overlap areas.

DESCRIPTION OF TEST

A model was constructed to simulate an actual interface. A 1.25 x 1.25 x.051 in. stainless plate was used to simulate a Cyrix MX CPU case top. A 3 x 5 x .063 in. aluminum plate was used to simulate a heatsink. The stainless plate was polished on 600 grade paper but it was deliberately left not perfectly flat to simulate real life situations and to involve some thicker areas of silver grease. The silver grease was applied and the joint made using the repetitive stippled effect observation technique.

The finished joint was then very lightly clamped and the whole assembly fixed in a vise with the joint in a vertical plane. A small drop and a thin smear of silver grease were placed on the aluminum plate adjacent to the stainless plate to evaluate any tendency of the material to free flow. A heat gun clamped 3 inches from the rear of the 3×5 aluminum plate was the heat source. Test data indicate that the interface was at temperatures well in excess of processor manufacturers’ case top limit of 70C.

OBSERVATIONS AND RESULTS

2:36 PM: Power applied to heat gun.
2:45 PM: 74C measured on the stainless plate with low cost uninsulated millivolt meter type temperature probe. So the temperature in the interface must be higher.
2:50 PM: Waited for temperature to level out at 81.4C. Carefully inspected the lower edge of joint with 6X loupe and observed no leakage of silver grease. No movement of drop and smear past markings.
2:51 PM: Temp of hot air source 1/2 way between heat gun and aluminum plate beyond the instrument’s limit of 150C.
2:57 PM: Loupe inspection does not indicate any joint leakage. Probed the drop to test for drying and might have caused 1/8 in. flow. Material still fluid. Smear has not moved.
3:36 PM: Stainless plate temperature 85.6C. Loupe inspection does not indicate any joint leakage. Drop and smear unmoved.
4:51 PM: Stainless plate temperature 85.8C. Loupe inspection does not indicate any joint leakage. Drop and smear unmoved. Smear looks dry but not, it responded to resmearing.
4:57 PM: Still no joint leakage. Drop finally flowed 5/8 inch. Smear unmoved. Continue test to observe the drop.
6:25 PM: No joint leakage. No drop movement. No smear movement.
7:00 PM: No joint leakage. No drop movement. No smear movement. 107.8C at original drop location. End test, power off.

DISCUSSION

Bengt Johansson was the first to see products’ potential for reducing temperatures inside microprocessors. He contacted the Cyrix-list and subsequent co-operative information gathering established that an 11.5 fold improvement in thermal conductivity was possible which would be worth at least 7C to a Cyrix MX PR200, more if overclocked. These are not insignificant improvements. So a supplier was located, the product ordered and this test conducted to determine the practical safe use of the product.

Silicone/silver grease application will duplicate that of the familiar SiZnO grease since both greases have similar consistencies. But because a test drop of silver grease did flow 5/8 in. it would be wise to swab any overlapping areas with a Q-Tip or similar after final assembly. This would be a normal precaution for SiZnO grease as well. But also note that there was absolutely no flow from an exposed smear nor any leakage from the interface joint, so the risk of short circuits should be negligible.

PRODUCT DESCRIPTION

[General Description from package]

7100 grease combines “pure silver” with advanced silicone lubricants. Conductive lubricant, maximum electrical conductivity, excellent thermal conductivity, protects against moisture, corrosion, and oxidation. Compatible with metal, rubber, and plastic. Ideal for applications requiring lubrication and maximum electrical and/or thermal conductivity, and remains stable over a wide temperature range. Applications include static discharge, grounding, HEATSINKS, and “soft” electronic connections.

[Technical Details from package].

Material——————Silver Grease
Consistency—————Smooth Paste
Temperature Range———(-57 to 252C)
Electrical Conductivity—(less than 0.01 ohm-cm)
Thermal Conductivity——50.0 [BTU/Hr/SqFt/In/F]
Thermal Conductivity k—-4.167 [BTU*Ft]/[Hr*SqFt*F] (conversion)
Unworked Penetration——210
Worked Penetration——–250
Dropping Point————255C
Chemical Resistance——-Excellent

APPENDIX

Calculation of temperature drop improvement.
***Aavid SiZnO joint resistance test, .274C/[sqin.*W] This translates to ~ .2C/W for 1.5 sq in Cyrix case top. But Aavid uses flat smooth polished blocks for test consistency. Real world case tops have raised lettering, perimeter ridges, and central cavities. So the “most optimistic” thermal resistance for real world case tops is felt to be .4C/W.

Silver grease k = 4.167 (Chemtronics/Circuit Works)
SiZnO grease k = .363 (Chemtronics product/tech support)
Improvement ratio = 11.5
Fractional improvement = [1 – 1/11.5] = .913
Un-overclocked power = 20W
Temp improvement = .913*.4*20 = 7.3C, 11C at a 30W overclocking.

***See “Thermal Performance Of Interface Material etc” aavid.com

SUPPLIERS

Action Electronics (action-electronics.com) – 800-563-9405
Mouser Electronics Part # 5168-7100 – 800-346-6873
Chemtronics (chemtronics.com)
PRICE ~ $14 per 1/4 oz. (est. about $1.50 or less per application)
PERFORMANCE/PRICE RATIO Silver/ZnO = 11.5/[$14/$2] = 1.64

John Carcich

And something about John:

“My background is engineering, testing large steam turbines, (and writing technical reports thereof). I am 73 years old, retired and living on the Outer Banks of NC to enjoy windsurfing at the windiest spot on the East Coast. Steam turbines/thermal experience explains interest in thermal aspects of computers, cooling, especially when
overclocking…which we push to the limit.

Currently driving a K6-2 450 to 504/112fsb 2.6V running Samsung GH memory synchronously. But overall I am little more than a novice re operating systems. “Used” computers since 1953, but you don’t learn anything “about” them from the keyboard. Decided to learn about PCs only a little over a year ago. Put together 3 systems…ie know how to read manuals and not color blind re cables 🙂

Thanks to Bengt-Arne Johansson of Sweden as he was the individual who called my attention to the product and its potential use as CPU thermal grease because of its excellent thermal conductivity. Its primary use/app is “electrical” conductivity.”


Leave a Reply

Your email address will not be published. Required fields are marked *