Not sure if this has been discussed but what about a cooling mod for what lookslike a 3rd processor unit underneath the graphics card?
When you completely remove the Graphics card, on the bottom there is another thermal pad that attaches to fan that is used for the Graphics Card... has anyone attached a copper plate to this component (i don't know what it is)?
I figure might as well go for it since it's very similar to the GPU & CPU. There is a dye/core of the processor to transfer heat and copper is certainly a much better conductor than the thermal pad. Which leads me to the question posed about the VRAM receiving more heat from the GPU because of improved conductivity via the copper plates. In theory, this is true, providing copper plates gives a better path for the heat to travel back to the VRAM and could heat up the VRAM more than before, but it seems to be less significant compared to the improved conductivity provided by the copper. Which is better, to have the thermal pads in place and the GPU and VRAM have a poor conductor for transferring heat away from them to the heatsink/fan combo... or to have solid copper plates to transfer the heat away from the GPU & VRAM to the heatsink / fan? Granted, the fan is weak and arguably may not provide much heat removal, but at least the the copper plates give it more heat to work with.
And by heat diffusion and assuming that the heat does transfer back to the VRAM and heat up the VRAM, i'm fairly sure that when the GPU & VRAM as a combo reach a stabilizing running temperature (let's say gaming for several hours and the heatsink / GPU / VRAM are maxed out and are relying on heat to dissipate via the fan mostly hopefully), the VRAM and GPU will be much lower in temperature with the copper plates than with the thermal pads. So my conclusion is copper plates for the VRAM is better than the Thermal Pads. Any and all copper plates are better than the thermal pads. even if you were to install only copper on the VRAM, it still would be better even if your theory of heat transferring from the GPU to the VRAM is correct because my assumption is the more heat sinks you provide, the better... which leads me to my next discussion...
Malone and Evo were discussing making the plate bigger so it touches the heat pipe
http://notebookforums.com/showthread...&page=29&pp=15
which is what i want to do. Actually, I want to make the copper plates as large as possible because the bigger heat sink you provide, the more heat you can remove from the GPU / VRAM / CPU / and whatever that third chip is under the graphics card. However, there are resistors, capacitors, jumpers, etc...(i have no idea) that are on top of the GPU where the copper plate is placed. Ideally, I would make the copper plate as big as possible and connect it directly to the heatsink but i'm worried it may short these components on top of the GPU...
ANybody know what these tiny buggers are on top of the GPU and if metal running across them will short them? I'm guessing they won't short since Malone and Evo have not mentioned any problems. Malone... Evo ??
back to my more heat sink the better theory, for the CPU i'm gonna try to cut a single piece of copper that runs along the heat sink and try to contact it with fins similar to the original heat sink. I may even throw a few layers extra along that curved heat sink to beef it up even more cuz The more heat sink (copper preferably) the better ...
If anyone has the fabrication skills, the solution to this problem would surely be to extend the heat pipe so that it can also touch the heat spreader from the VRAM and thus give it an outlet of lower potential energy also. Come to think of it, would it not be possible to cut the VRAM spreader with a small finger that does in fact contact the heat pipe directly??? I will leave you Physicists to work out the possibilities and consequences here 
(I wanted to see the banana man...heh)
, and I’m still working on my V.RAM cooling idea.
