Replacing IHS w/ copper...

Many of you have read my other post of how I removed my IHS from my A64 and went for straight core to heastink contact.

http://www.hardforum.com/showthread.php?t=839108

Well, before I did that I was cooking up another kooky plan of replacing the IHS with a piece of solid copper possessing the same dimensions (roughly). Well, I finally got a piece I ordered.

So this picture comparison is for those who may be interested in replacing your IHS, since some hsf clamping methods need that fraction of an inch to obtain proper pressure on the processor.

This was my original plan for my A64, but since I just went the other route (and got great results) I don't feel the need to go with my original plan of IHS substitution.

Anyways, here you go (sorry about the last pic, I cant figure out how to get REAL good pics of REALLY close objects (yes I played with manual focus)).

Side-by-side : roughly the same length and width


What would have been the new copper IHS sitting on top the old IHS


Thickness comparison

Copper piece is flat, some excess material from the cutting was on the edge of the copper, easily removed after picture was taken)

With a little sanding and a good lapping, the copper IHS-replacement could easily be worked done to the IHS thickness (and have a nice mirror shine and better heat transfer properties).

The only issue I really saw with doing this is the original A64 IHS is recessed on the bottom by a VERY small amount (like 1/2 a millimeter), so obviously you'd want to make the copper one a little thinner than the outer edges of the IHS (or similar thickness of the recessed portion of the stock IHS).

It sounds like a step backwards from a direct die-to-heatsink contact. Copper still has a lot of thermal resistance.

I know, thats why I was going to leave it the way it is now.

I was merely posting this for those interested in replacing their IHS instead of removing it altogether (since some HSF setups would have contact problems with a chip that did NOT have the IHS present (or atleast something to replace the IHS)).

care to tell us how much changing the ihs from the stock one to the lapped copper one improves/hurts perforance (and cause you alredy have the data, throw in the numbers from direct die )

Tenchi4U said:

I know, thats why I was going to leave it the way it is now.

I was merely posting this for those interested in replacing their IHS instead of removing it altogether (since some HSF setups would have contact problems with a chip that did NOT have the IHS present (or atleast something to replace the IHS)).

Click to expand...

Everything you say makes sense. Right up to a point…

Aluminum sucks as a heat spreader; I think AS even points this out in their instructions on the application of AS-5 on AMD 64 and Intel P-4 CPU’s. They suggest using only a small drop of compound in the center of the CPU spreader because that is where most of the heat will pass up to the HSF. Makes perfect sense.

Ponder this; copper has much better characteristics for spreading heat over a larger surface. Think copper-bottomed cookware as an example.

Is it possible that your copper shim would in fact spread more heat from the core to a larger area of the HSF assembly, thus increasing it’s efficiently? More heat spread to more fins would (should) equal better cooling.

This would of course require proper finishing on all flat surfaces.

Just some thoughts

but if your heatsink is already copper, direct die would definitly be better anyhow. though copper should show a good improvement over the stock alum. one (2-4ºc would be my guess, since he got a 6º drop on full load without it)

(cf)Eclipse said:

but if your heatsink is already copper, direct die would definitly be better anyhow. though copper should show a good improvement over the stock alum. one (2-4ºc would be my guess, since he got a 6º drop on full load without it)

Click to expand...

Yup, you could be right

Thermal dynamics is not my field. Just thought it might be fun to experiment with though, an interesting project so to speak.

In a lot of my cookware analogy they use different layers of copper for more “spread factor” and they use different density copper as part of the core in better cookware. (No, I’m not a cookware expert either, just to clear that up )

Just thinking out loud

Has been an interesting project, I wish I had another A64 to test this on and NO I'm not going to test it on my current setup, that just seems like a step backward for me (and I'm too lazy to take everything apart and run hours of Prime95 based tests).

Main points, so nobody will fight over what we have gleaned off my work already:

1. Direct heatsink to die/core contact is the most efficient means of heats transfer (since the heat will NOT have to pass through an intermediary (instead of the core -> thermal material -> IHS -> thermal material -> HSF; the IHS-free method is simply core -> thermal material (AS5) -> heatsink).

2. I merely posted this for those who may be interested in swapping the IHS, since some HSF setups will not allow for the proper pressure. I've read others who have removed their IHS have had problems of NOT enough pressure being supplied due to the lack if IHS retention method takes the chip AND IHS height into account (most utilize the clamp method of retention, NOT the screw/spring method like mine).




10min lap job - yes, that is the same piece of copper from the first thread,

Isn't the stock IHS nickel plated copper like Intel CPUs?

Isn't the stock IHS nickel plated copper like Intel CPUs?

Click to expand...

To be honest, I'm not entirely sure, I've heard both. I wonder is AMD would tell me if I emailed them.

I still wish I had another a64 I could test, just to see if my copper IHS or the stodk one is more efficient.

I'm still happier with the route I chose though (total IHS elimination).

Interesting article about this sort of thing (search "IHS")

http://www.intel.com/technology/itj/q32000/pdf/thermal.pdf

Tenchi4U said:

Has been an interesting project, I wish I had another A64 to test this on and NO I'm not going to test it on my current setup, that just seems like a step backward for me (and I'm too lazy to take everything apart and run hours of Prime95 based tests).

Main points, so nobody will fight over what we have gleaned off my work already:

1. Direct heatsink to die/core contact is the most efficient means of heats transfer (since the heat will NOT have to pass through an intermediary (instead of the core -> thermal material -> IHS -> thermal material -> HSF; the IHS-free method is simply core -> thermal material (AS5) -> heatsink).

2. I merely posted this for those who may be interested in swapping the IHS, since some HSF setups will not allow for the proper pressure. I've read others who have removed their IHS have had problems of NOT enough pressure being supplied due to the lack if IHS retention method takes the chip AND IHS height into account (most utilize the clamp method of retention, NOT the screw/spring method like mine).

10min lap job - yes, that is the same piece of copper from the first thread,

Click to expand...

Heh, you said "lap job"

no....I said "10min lap job" - wonder how much that would cost over at the gentleman's club

BillR said:

Heh, you said "lap job"

Click to expand...

Shoot. You're not too far from Philly. If you want to learn about lapping, you should go to Daydreams. Talk about thermal transfer... damn.

Daydreams, where YOU become the integrated heat spreader...oh yeah...