Well, I read all the forums for opinions on the merits of different thermal compounds and this discussion highlighted the difference in silver content in two common silver based products. AS5 seems to be the market leader and has an enthusiastic following. Dynex is another product that is easy to get at your local BB store, so many of us have tried that because it was there.
I heard some rather dramatic claims on the superiority of AS5 so I thought it would be a good product for long term, but before I got my first tube of AS5, I used some Dynex to mount my new OCZ Vendetta II CPU cooler (a great product, but that's another story). I decided to do some testing myself with both products to see if there is anything to the claims.
I ran the cooler first with Dynex as the thermal interface, then about a week later, I broke it all down and started over with AS5. I spent about an hour cleaning the CPU heat spreader and cooler interface before applying AS5.
I chose to apply both the Dynex the AS5 as an evenly applied thin layer on the heat spreader as opposed to just a dab. The dab technique is appropriate when the cooler and heat spreader surfaces are both perfectly flat because the thermal compound will spread fairly evenly when the two are compressed together. This is not the case with the OCZ and similar coolers where the heat pipes are inset into the cooler base leaving small crevices between the pipes and base.
I was careful throughout the process to keep the test environment as consistent and stable as possible. Just to make sure, I let each application run for a week to ensure they went through plenty of heat cycles. However, I never saw any significant difference in temp from that measured within the first 10 minutes of operation. Following is a summary of the setup:
CPU: AMD Phenom 9850 BE at stock speed (2.5)
Cooler: OCZ Vendetta II
Case: Standard ATX Tower; 2 exhaust fans; case open
Ambient Temp: 69F
Temp Measurement: Speedfan and Everest Ultimate Edition
Loading: Everest Ultimate System Stability Test (running all 4 cores at 100%)
Idle temperatures were measured with Speedfan after ensuring no program activity for 30 minutes.
Load temperature was measured with Speedfan and validated with Everest to ensure no variation. The System stability test was configured to exercise the CPU, FPU and Cache only since I was only interested in the heat generated under the cooler.
The duration of the test was 10 minutes for both test cases. After the first two minutes at load, the load temp reached maximum and roughly flatlined after that point. Here are the results comparing the stock setup to both setups with the OCZ cooler and silver-based compounds:
Stock
Idle: 38C Load: 50C
OCZ w/Dynex
Idle: 26C Load: 38C
OCZ w/AS5
Idle: 25C Load: 37C
As mentioned previously, these were open case results, closed-case temps ran approximately 3C higher for both setups.
Conclusion: While this certainly isn't a scientific laboratory test, it was enough to satisfy my curiousity at the relative efficiency of two silver-based thermal compounds. As you can see from the measured temps, there was not a dramatic difference between the two, but the AS5 did provide roughly 1C cooler temps under the same conditions.
One unexpected result was in the stability of the temperature at load with the AS5. Once max load was reached, there was less variation up and down during the AS5 test.
So in the final analysis, I have no reservations at all recommending the AS5 compound when cooling efficiency is critical. At the same time, I wouldn't discourage using the Dynex product if it were more readily available or was to be used for multiple applications. Both provided acceptable results in my test setup and during later overclocked operation.
Other thoughts: The Dynex compound was easier to apply because it was lighter and seemed to have a finer texture. The AS5 was much "stickier" and harder to spread, but once it was in place, it tended to stay put. Neither compound was in any danger of running away, but once you started manipulating the cooler on the heat spreader, the Dynex compound moved around more than the AS5.
I'm sure that 10 different people could get 10 different results, but hopefully someone might find this comparison test enlightening.
Regards, Beaz
Edited by Beaz - 1/22/09 at 5:00pm
| CPU | Motherboard | Graphics | RAM |
|---|
| Phenom 9850 BE | ECS GF8200A | 9800 GT | 8 GB |
| Hard Drive | OS |
|---|
| 1TB SATA-II | Vista Ultimate 64bit |
| View all |
hide details
| CPU | Motherboard | Graphics | RAM |
|---|
| Phenom 9850 BE | ECS GF8200A | 9800 GT | 8 GB |
| Hard Drive | OS |
|---|
| 1TB SATA-II | Vista Ultimate 64bit |
| View all |
hide details