Does it make a waterblock better?–Joe
I sent the silver plate back to Rick Milano and he returned it after lapping it flat. I retested on a Duron 800 @ 1000 MHz, 1.91 volts, using the silver and copper plates between a BE Cooling copper waterblock to see how much of a difference the silver plate would make. I ran both for one hour with Prime 95, then recorded temps:
CPU Back Temp
|Silver on Duron @ 1000/ABIT KT7 (64 watts)|
|Copper on Duron @ 1000/ABIT KT7 (64 watts)|
Not much of a difference – the Deltas are not statistically significant, although the CPU back temps are. This test is inconclusive, so I ratcheted up the test using a static test block. I maxed out the power supply and tested the exact same setup at 138 watts:
Die Temp Temp
|Silver (138 watts)|
|Copper (138 watts)|
Now we can see some differences, although in the scheme of things, it’s relatively minor: 1.6 C at the waterblock’s base.
Silver is a better thermal conductor than copper; as the table below shows, it’s about 5% better. Consistent with this difference, test results show a minimal gain. Considering the relative cost differential between the two materials, I don’t think the gain is cost justified, although some may consider it performance justified.
SUMMARY 8/27/01: A warped silver plate performs as well as copper – how much better will a flat one do?
There has been lots of interest lately in silver as a better material for heatsinks due to its higher thermal conductivity. Below is how much better silver is as a conductor than copper:
- Silver (429 W/mK)
- Pure Copper (401 W/mK)
Source: Cooling – An In-Depth Look
About 5% better than copper. Is this difference enough to justify 100 times higher cost (silver: $4.22/ounce; copper: $.042/ounce)?
Rick Milano contacted me about some products he is planning to market, among them is a 3/16″ thick solid silver plate. He asked me to test it out to see how it compares to a similar copper plate. He sent me one to test out.
I decided to test it by mounting each plate to a copper water block. I then mounted this assembly onto a T-Bird running 1397 @ 1.98 volts (99 watts). I recorded temps of the waterblock (thermocouple mounted above the CPU core), CPU back and air at the radiator’s intake.
CPU Back Temp
|Silver on T-Bird @ 1397/Iwill KK266+ (99 watts)|
|Copper on T-Bird @ 1397/Iwill KK266+ (99 watts)|
C/W = Delta / CPU Watts
Interpreting C/W: For every watt the CPU radiates, the heatsink will cool the core by the (C/W x watts) plus ambient temp. For example, at an ambient temp of 25 C, a C/W of 0.25 with a CPU radiating 50 watts means that the CPU temp will be 50 x 0.25 = 12.5 C over ambient temp, or 37.5 C. The lower the C/W, the better.
Looking at this, you might conclude there is no difference. However, what bothered me is the different CPU back temps – over 11 C higher for Silver than copper. Something is not right.
I then took a closer look at the Silver block and found that it was not perfectly flat; it was warped enough that it did not contact the waterblock as fully as the copper block – just enough to throw off the results.
The jury is still out – this is not an apples-to-apples test. Even warped, silver pulled even with copper, suggesting that better performance may be possible. However, whether 5% (or more?) better is noticeable and worth 100 times the cost is debatable.
Rick is sending me a lapped silver plate to test, so stay tuned.