Silver's % advantage vs. copper
 

This is sort of a follow on to the thread about alloys.

A response given to questions on that thread showed me that only prure copper or silver works really well, the alloy wouldn't make it.

So how much of a advantage would a silver block offer vs. the same design made of high grade pure copper? Would this silver block offer more than a 2-4C improvement in a well designed water cooling system where only the metal used in the CPU block was differant?

Re: Silver's % advantage vs. copper
 

In a word, no. The difference in conductivity is on the order of 7% for the pure metals. Consider that the delta-T through the block metal is at most a few degrees C and the advantage is roughly 0.07 * "a few degrees C".

My charts list conductivity of pure copper and pure silver at 401 and 429, respectively. I'd put the units on it, but they are irrelevant so long as you're using the same for both.

.07C ? ?

How, if the differance between the 2 metals is 7% can this be?

If you have a CPU with a 80C potential, how can you end with a differance so tiny?

A good high flow and high pressure system using a copper block of good design can bring this down to 35C for sake of having a number to start from.
Now as silver is 7% better, how is it that it makes such a small differance? It would seem that you should have a differance of at least 2C. I don't get this one. Can you show me the why/and or how of this?

OK. On a good watercooling system you probably have on the order of 20-30°C between the true die temperature and the ambient air that's cooling the radiator. Of this 20-30°C, some is in the die/block interface, some is in the metal of the block, some is the delta-T between the surface of the block and the liquid, some is between the liquid and radiator tubing, some is between the radiator tubing and fins, and some is between the fins and air. If you take 7% of the portion that exists across the block, you're talking a very, very small number. Consider that aluminum and copper vary to the tune of ~230 to 400, or nearly a doubling of conductivity. Yet the difference between good copper and aluminum blocks tends to be on the order of 5°C.

And no, I didn't mean the difference would be 0.07°C. I said 0.07 * a few degrees C. Say maybe a final result of less than 1°C.

What? With good copper you're looking at maybe 10C difference, maybe a good deal less with higher flow systems. Then some of that is in the CPU/metal interface and still more is in the core itself (heat still must be conducted through the outside of the core).

Just look at Cu vs. Al. ~2 times the conductivity, but a good Al block will be with a few C of everything else.

Ok, let's see if I've got this now.

I was looking at the block and it's interface with the water on one side and the thermal paste and CPU on the other.

But the proper way is the whole cooling range of interactions or steps until the heat is disapated into the air off the rad. fins. Correct? So the advantage of silver over copper is diluted through the numerous stages of getting rid of the heat. Is this right?

Thanks for trying to help me guys.

Yep. You want to focus on the major points of heat transfer resistance, like the radiator, if you want to improve performance. Heat moves very efficiently across a piece of copper. There's no gain in improving that.

Alchemy

Make the whole rig out of silver :D, hmmm?, silver tubing, silver barbs, silver jubillee clips.....

/me waits eagerly for the first fullerene waterblock...
(it's supposed to conduct heat as well, if not better than diamond !)
Would be a bitch to drill though.

Don't be silly. You can line an unlimited number of radiators until you run out of space. If you have the will its not hard to bring the coolant to within a fraction of ambient. People have done it. Furthermore even a simple heatercore with a modest 120mm fan will cool the water air interface to a lower delta T then the block-water interface.

Now try and get a copper block within 5C of coolant. You can't without TECs (or an innaccurate probe).

Hi,
I may be missing something here but with my ambient chiller i normally see water temps of 19~20 and cpu-waterblock temps of 20~22 (only 2 degree rise) and cpu temps (die) of 22~24.
wj

But you're still using a pelt, just in a different fasion, but get basically the same results.

So, if silver`s performance advantage over copper is so low; The whole deal of C110 copper versus normal one is absolute nonsence!