Got a stumper for ya Mikewarrior. Its about shims.

[WarCon]

First off I am not a proponent of shims for cooling. I think shims should be make of a non conductive substance that is slightly compressable to allow best contact with core.

That said, I was playing earlier today with a silver heat spreader (turned out to be an alloy, which killed the thermal properties) on my water block. I had lapped the silver commerative coin to almost mirror-like. I decided to see if since I am using a spring-loaded hold down on the 4 motherboard holes, if I could do without my shim.

Well the silver heat spreader sucked and I took it off. I re-cleaned everything and ASIIed my core like your supposed to and mounted my water block. When I turned my system on it wasn't stable at a previously stable point. My first thoughts were that my waterblock was mounted wrong so I loosened everything and leveled it and held it flat as I tightened the four wingnuts in a best as I could do star pattern to get even pressure. Same thing, so I used my Insert key to boot at default speed and my temps were on average 5C higher (by my mobo KT7A with diode touching) for the same speeds. I was freaking. I took it off to see if I crushed my core. Only tiny little chips off corners. So I mounted it again. Same result. I was pretty frustrated thinking those tiny little chips had degraded my chips cooling that bad. So I got my shim out again and decided I was just going to run it as best I could til it died, which I expected to be soon. After I got it all together, I booted it and I was pleasantly surprised by getting the same temps I had before and the same stability.

Mikewarrior, what the heck is up? I am sure it was as flat on the core as I could get it. Essentially did it 3 times, twice with different ASII applications. And I could see from the AS that it left on the waterblock that it was making good contact. Man I am confused.......HELP

 

[Mikewarrior2]

It would seem that it wasn't flat before, or, in the process of loosening the waterblock, it came "uneven" or not enough pressure was applied to the core, or any combination of things...

When the shim was installed, you were able to tighten all without fear of pushing too far, so you probably got a really good contact between core/shim...

P.S. It is kinda hard to notice something sitting just a bit crooked on a socket-board..


Mike

 

[WarCon]

Tell me about it. But wouldn't the four sprung hold down tend to center the pressure and make it flat? I did have a really good imprint on my waterblock. What I am really wondering if all I am doing is cooling the ceramic through the shim and making my readings lower. (Since my thermistor is touching the ceramic on the back.)

What really confuses me is that it did seem to effect stability too.

I really hope that core designs become more aware of the problem of core damage and need for a tight fit. Intels way I believe is better, but still should be better. I like the idea of the chip being protected under a heat spreader made of cusil or silver. Tell me they can't afford to put a 1/4 ounce or 1/2 ounce slug of silver over the core to make sure that first heat junction has more draw and less destructablity. And man I can't wait til they put a die thermistor in AMD chips.

 

[Mikewarrior2]

I wouldn't worry about the shim/waterblock..

also, the good impression of grease on the waterblock could have been there, but then it could have shifted...

Intel is going with aluminum heat spreaders on future p3s/p4s, hoepfully amd will do something similiar to protect the o so fragile cores...


Mike

 

[WarCon]

If they don't then I would have to ask myself how they can couple such wonderful advances with such a weak implementation. People just shouldn't be able to break a core, putting on or taking off the cooling device, period. Number of times shouldn't be a factor.

Thanks for the quick response.