Cooling (cont’d), clocks, etc.

So far we’ve only detailed the layers of ACS³ that lie above the GPU, but EVGA has also made quite a few additions to the underside of the card for greater cooling performance as well.

Bottom layers of the EVGA cooling

EVGA starts the back of the card’s cooling off with a long, black aluminum plate. Sitting directly underneath the G70 GPU is a purple thermal pad. This pad is responsible for transferring heat from the underside of the GPU directly to the aluminum plate (via a heatsink which we’ll discuss later, which makes direct contact with the pad). In comparison, the plate on NVIDIA’s reference 7800 GTX design doesn’t make direct contact with the PCB, hurting its effectiveness.

By going with a long aluminum plate rather than the smaller plate used on NVIDIA’s reference 7800 GTX design, heat can be transferred over a greater surface area; essentially allowing the plate to “soak” up more heat, just like a larger sponge would soak up more water.



The only downside to this design is that it shares some of that area with the board’s memory modules, which can get pretty toasty under load themselves. With four memory modules and the GPU, you’ve got five hotspots on the bottom of the e-GeForce 7800 GTX KO ACS³, all which need to be cooled.

Note how large the EVGA plate on the bottom of the board is in comparison to NVIDIA reference board

NVIDIA gets around this on their reference design by using one large plate for the memory modules, and one smaller plate to help cool the GPU. This helps to separate the components from one another.

EVGA on the other hand helps mitigate this problem by including an additional layer of cooling on the underside of the board: RAMsinks for the memory modules, and a small heatsink directly underneath the GPU, sitting atop the thermal pad we mentioned earlier. Basically, these heatsinks take the brunt of the load for the large aluminum plate, with the RAMsinks responsible for the memory modules and the black aluminum heatsink at the center of the underside of the card responsible for the GPU. The large black plate then soaks up whatever these heatsinks can’t handle.

RAMsinks and heatsink under GPU provide additional cooling

Use of thermal pads

Thermal pads aren’t used solely to transfer heat from the underside of the GPU, EVGA also uses them for all of the board’s memory modules, including the modules on the top of the card. This is the same practice NVIDIA (and most board partners for that matter) uses for their reference GeForce 7800 GTX card. Thermal pads are also used frequently on ATI-based cards as well.

We mention this because some enthusiasts expected EVGA to use thermal adhesive, such as that made by Arctic Silver for their e-GeForce 7800 GTX KO ACS³ card. Thermal adhesive is preferred by enthusiasts because it’s a little more effective than thermal pads at transferring heat. EVGA chose to stick with thermal pads for two reasons, ease of assembly, and also the permanent nature of thermal adhesive: once it’s on there, the two components are permanently bonded together.