Thermal paste
 

Artic silver has a thermal conductivity of 9 w/m k... solid silver has 429... iron 80....


But it will improve performance, is this because there otherwise would be a gap between core and block?

Yes.

Talk about bottle neck...

I experimented little with thermal transfer and preassure. I had a cold copper pipe and a one warm, when i had them thouching eachother with no pressure there where virtualy no heat transfer at all. When pressure where given there where significantly more heat transfer..

This experiment gave me the understanding of that even hard materials nead to be pressed together in order to exchange temperature properties effectively.

Now, how much pressure can a core cope with? (amd xp2000)

Officially, 25lbs (11kgs)

Unofficially, I've probably applied up to 25kgs of force on one.

If it's the organic packaging (green/brown) you're reasonably safe putting a lot of pressure onto an AXP core as long as that pressure is "flat" (ie. not pushing on an edge or corner).

With the older ceramic packaged Athlon T'Bird, I wouldn't put much more than 15kgs of pressure on one as the packaging can snap.

You have to remember that the thermal paste is ONLY there to fill in any microscopic voids where the two metals are NOT touching.

Some people put way too much on, wich reduces thermal conductivity between CPU & block/HS. You just want a very thin layer so that the TP fills in any voids that would otherwise be filled with air.

More pressure will give better temps. But that same pressure WITH TP will give even better temps.

I found AS2 used to migrate within two months anyway :shrug: , so I'm a little more 'liberal' now than I used to be :) . I hav'nt had a chance to try AS3 yet, I brought a tube but it 'migrated' off the shelf it was on into my wastepaper basket!(along with my nylon block mounting stuff and springs):rolleyes: (I found out two weeks later!!):cry: . but they say the migrating's fixed with AS3. anyone had a look~see?...
PS. would 24lbs equate to 6lbs on each corner?, or does it not work like that?...

I've got systems that have run in excess of one year without issue using ASII. I take care to blow the dust out now and then, but nothing more. I have noticed odd patterns on both the chip packaging and heat sink base outside of the die region that could be described as "ghosting". These appear to be the result of some volatile component of the ASII, but haven't changed performance by my reackoning.

As to you "PS", the answer is yes if the 24 pounds is centered within the "box" defined by your corners.

Yep, I meant as in the four mounting(AMD) holes, which are centred in their way I s'pose.

I don't know if the performance was adversly affected, but there was just some kinda oily residue left on most of the core area. all(well, most) of the 'silver' had moved to the edge or beyond.

I might have put it on to thick (even by my 'thin' standard), I could imagine this exagerating the problem now I think about it. to much volume of oily stuff allowing the silver to migrate???...

Can you get torque drivers that go down to that lbs with a good 'resolution'(degree of acuract). or is that overkill?. when that fine a measurement the drag on the thread or spring snagging would throw it out would'nt it?. or is 6lbs quite hefty?...

There is no universal way to accurately determine clamping force based on torque applied to a fastener. There are only rules of thumb that are no where near the accuracy of what you would like.

For mounting it is far more accurate and practical to use springs. With these you can measure (or calculate) force versus deflection. The force really isn't that critical unless:

1) You use thermal tape (final thickness varies with force)
2) You use too much goop (similar to #1, but not as severe)
3) You have excess tubing hanging off the block or tubing with tight bends close to the block (tries to "pry" the block)
4) In air cooling if you have a "tall/heavy" fan arrangement (oversize fan or tall funnel, which will also try to pry the block off the CPU)

More important for four-holers is equal pressure at each corner. Compress each spring an equal amount and you should do fine. As for final force, compress them in your hand and get a general feel. Ideally, the springs are part of a good kit and have already been selected to give the right force at a prescribed level of compression.

some words on the subject here

if a couple of degrees either way are 'good enough', don't sweat it

if you're trying to optimize, oohh - what a chore indeed

Optimizing here...

I had to lookup CF, but got linked back to OC...:D

This is going to take me longer than designing the block! By the time I've got these issues resolved, I'll be into an Opteron 4400+...

I lost my mounting kit and spent 3days hunting parts for another, the M4 50mm(brass,longest I could get) bolts were the hardest to find!. I could'nt find any springs short enough so had to cut two in half :( , I flatened the nasty point that's left as best I could without throwing the shape of the spring right out, but I have no idea of the tension of the springs!. Is it determined by the gauge of the spring metal mainly?. what kind of gauge (in MM) are the springs used?. could someone show me a close up pic with a rule next to them for comparison?.