If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below. |
|
|
|
Thread Tools | Display Modes |
|
#1
|
|||
|
|||
Copper Shims - Any good
Is it worth getting a copper shim for my AMD XP when I fit a new
cooler? What exactly do they do? TIA Alex |
#2
|
|||
|
|||
I wouldn't bother. They're meant to "prevent" the core from being crushed
due to too much pressure from the heatsink. I've never experienced this on any of the systems I've built. "Alex" wrote in message ... Is it worth getting a copper shim for my AMD XP when I fit a new cooler? What exactly do they do? TIA Alex |
#3
|
|||
|
|||
"Alex" wrote in message
... Is it worth getting a copper shim for my AMD XP when I fit a new cooler? What exactly do they do? TIA Alex I wouldn't bother. They're meant to "prevent" the core from being crushed due to too much pressure from the heatsink. I've never experienced this on any of the systems I've built. Yeah your really only in any danger of crushing the core if you have a screw on heatsink. |
#4
|
|||
|
|||
I wouldn't bother. They're meant to "prevent" the core from being crushed
due to too much pressure from the heatsink. You are neglecting the potential additional heat transfer of cpu heat from the substrate to the heatsink with a copper shim. If you were careful in coating both sides of the shim with compound, I wouldn't be at all surprised to see say 1-2C temperature difference from a shim-less install. Small of course, but so what? Small differences make tinkering with such simple mods fun. rms |
#5
|
|||
|
|||
rms wrote:
I wouldn't bother. They're meant to "prevent" the core from being crushed due to too much pressure from the heatsink. You are neglecting the potential additional heat transfer of cpu heat from the substrate to the heatsink with a copper shim. Actually, it's the reverse. The substrate has a much higher thermal resistance compared to copper (or aluminium), so at any particular point on the substrate, the HSF surface above it has a higher temperature. So you'd actually be dumping heat back into the substrate, though in any case there isn't much conduction/convection away from the substrate, so there would be a negligible performance difference. [...] -- Michael Brown www.emboss.co.nz : OOS/RSI software and more Add michael@ to emboss.co.nz - My inbox is always open |
#6
|
|||
|
|||
you'd
actually be dumping heat back into the substrate Interesting theory, but I'm not sure I buy it, especially in the case of a water-cooled system like mine. I can touch the perimeter of my waterblock under full load, and it is essentially the water temperature. In other words there is a significant temperature gradient between the outer and inner areas of the waterblock surface, and I would imagine this gradient is much steeper than for the average air-cooled sink, given the poor heat removing capacity of air vs that of water. The purpose of the copper shim is to increase contact from the substrate to the outer area of the waterblock surface, which in this scenario is quite likely to be cooler than the substrate, thus removing additional heat, not adding it. Also, in the ideal situation, with a more or less perfect heatsink that instantly removes all heat without spreading it out over the heatsink surface, it should be clear that more surface area equals more heat removal. So the better the heatsink, the more a cpu copper shim should help. That's my argument anyway. No, I haven't tried a copper shim yet, tho I have bought one and have been meaning to try it out, along with doing a simple waterblock mod and using larger fans on my radiator. rms |
#7
|
|||
|
|||
rms wrote:
you'd actually be dumping heat back into the substrate Interesting theory, but I'm not sure I buy it, especially in the case of a water-cooled system like mine. Ah, yes, I should have been a little more specific In most air-cooled systems, the horozontal gradient is relatively low on the bottom of the HSF, with a steep gradient right near the edge (where there is significant air circulation and whirlpooling). This is primarily due to the dead area of air circulation above the HSF base. The main temperature gradient is vertical in these cases. I saw some thermal imaging camera pics of this somewhere (LostCircuits I thought) but I don't have any links on me at the moment. Waterblocks and (some) non-standard HSF designs get around this. For a well-designed waterblock, you can effectively treat the inside surface as being at water temperature. This means that the horozontal gradient is very steep (as is the vertical gradient as you've only got a few mm of copper between the heat source and the inner surface), and hence the shim would actually help cooling as long as it didn't reduce the pressure, etc between the block and the die. I still don't think it would be particularily great, as the thermal resistance of the substrate is relatively high, but it would probably drop your substrate temps (ie: socket temps) and make you feel a bit better It'd be interesting to see what it would do for die temps (though repeated installation and removal of the shim would be necessary to get some idea of the standard error of the measurements created from WB installation variations). [...] Also, in the ideal situation, with a more or less perfect heatsink that instantly removes all heat without spreading it out over the heatsink surface, it should be clear that more surface area equals more heat removal. So the better the heatsink, the more a cpu copper shim should help. That's my argument anyway. Agreed [...] -- Michael Brown www.emboss.co.nz : OOS/RSI software and more Add michael@ to emboss.co.nz - My inbox is always open |
#8
|
|||
|
|||
Michael Brown wrote: rms wrote: I wouldn't bother. They're meant to "prevent" the core from being crushed due to too much pressure from the heatsink. You are neglecting the potential additional heat transfer of cpu heat from the substrate to the heatsink with a copper shim. Actually, it's the reverse. The substrate has a much higher thermal resistance compared to copper (or aluminium), so at any particular point on the substrate, the HSF surface above it has a higher temperature. So you'd actually be dumping heat back into the substrate, though in any case there isn't much conduction/convection away from the substrate, so there would be a negligible performance difference. Actually, a shim does NOTHING for thermal transfer. Think about it this way: when was the last time you felt the actual ceramic base on a chip? It stays comparatively cool. Ceramic is a thermal insulator, which means that it will keep heat from going from the silicon of the processor to the rest of the packaging material. That means that the heat has nowhere to go except to move into the heatsink. Also you should be careful about putting any type of thermal interface material on a shim that is conductive for the simple fact that XPs have bridges on the ceramic that could get shorted by the TIM. ---------- To reply, replace 'deadfishies' with 'bigfoot.' We live our life in our own way, never ever listen to what they say, the kind of faith that never fades away, we are the True Believers! |
#9
|
|||
|
|||
Christopher Rawlison wrote:
Michael Brown wrote: rms wrote: I wouldn't bother. They're meant to "prevent" the core from being crushed due to too much pressure from the heatsink. You are neglecting the potential additional heat transfer of cpu heat from the substrate to the heatsink with a copper shim. Actually, it's the reverse. The substrate has a much higher thermal resistance compared to copper (or aluminium), so at any particular point on the substrate, the HSF surface above it has a higher temperature. So you'd actually be dumping heat back into the substrate, though in any case there isn't much conduction/convection away from the substrate, so there would be a negligible performance difference. Actually, a shim does NOTHING for thermal transfer. Think about it this way: when was the last time you felt the actual ceramic base on a chip? It stays comparatively cool. First of all, it's not ceramic. Ceramic went out with the Thunderbirds and early Palominos. The current XP's are on an "organic" substrate, whatever that means. The substrate on my XP2500's does get noticably warm after it's been on for a while. Secondly, the substrate has a high thermal resistance, so could be called a thermal insulator (a thermal insulator is anything with a high thermal resistance, and a conductor is one with a low resistance, high and low being rather subjective). This does not mean that it won't heat up. All it means is that heat will not flow through the material quickly, in the same sense as current flowing. In the electrical sense, you can have a device with a high resistance (thermal resistance) but also have it at a high potential voltage (temperature). As long as there is a low heat loss through conduction and convection, the substrate will heat up. This is why socket sensors report a somewhat accurate temperature. However, this high thermal resistance means the heat-carrying capability of the substrate is quite low, which is why there is (or should be) a negligible temperature increase/decrease compared to not using a shim. [...] Also you should be careful about putting any type of thermal interface material on a shim that is conductive for the simple fact that XPs have bridges on the ceramic that could get shorted by the TIM. That's what a drill press is for Seriously though, modifying a shim is not recommended as it's very easy to warp it, which leads to bad die-hsf contact. I beleive there's pre-machined copper shims for the new XPs that have holes around the bridges, so as long as you're careful with the goop (and ideally use silicon goop) there shouldn't be any problems. -- Michael Brown www.emboss.co.nz : OOS/RSI software and more Add michael@ to emboss.co.nz - My inbox is always open |
#10
|
|||
|
|||
rms wrote:
I wouldn't bother. They're meant to "prevent" the core from being crushed due to too much pressure from the heatsink. You are neglecting the potential additional heat transfer of cpu heat from the substrate to the heatsink with a copper shim. Thermal resistance of the substrate is such that the heat transfer is negligible anyway. If you were careful in coating both sides of the shim with compound, I wouldn't be at all surprised to see say 1-2C temperature difference from a shim-less Difference in the temperature of what? install. Small of course, but so what? Small differences make tinkering with such simple mods fun. What about the reduction of air flow to the side of the core, and to the substrate? Would that not have a negative effect on temps? Ben -- A7N8X FAQ: www.ben.pope.name/a7n8x_faq.html Questions by email will likely be ignored, please use the newsgroups. I'm not just a number. To many, I'm known as a String... |
|
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Anyone know a good deal on a good mp3 player? | travel | General | 1 | November 30th 04 10:51 PM |
Which Wireless Option Is Good For Streaming Video? | Jay Chan | General | 0 | October 21st 04 04:32 PM |
Are cpu shims worth using? | Jim | General | 4 | April 11th 04 04:10 PM |
Help me pick out a good mobo to use with XP 2800+ | Ben in TN | Overclocking AMD Processors | 3 | February 7th 04 11:25 AM |
Good 2500+ ? | Thomas Andersson | Overclocking AMD Processors | 3 | July 21st 03 04:10 PM |