Test Bed

Testing Methodology

We used the Omega HH506R thermometer, which supports two thermocouples input and RS-232 logging functions, using two Type-K thermocouples.

Testing was done over a time span of 10 minutes per test by running Aquamark 3 demo mode in a loop. Ambient and GPU temperatures were recorded simultaneously using a secondary PC via RS-232.

The thermocouple was taped right next to the GPU. The second thermocouple measured the room temperature outside the case. The final results in the table below show the differences between the core and ambient temperature measurements.

RAM heatsinks were not installed during these tests.

Test Results

According to the graph, the performance VF700-AlCu at 12V was about 7°C higher than the ZM80D-HP (at 12V) on average. While running at 5V, the ZM80D-HP was 8°C cooler than the VF700-AlCu. Doesn't look promising, right?

Well, I have to admit that our testing methodology is not fool proof. It's virtually impossible to measure the temperature of the GPU without external variables affecting the results. Such external variables include airflow around the thermal probe, thermal probe movement during the installation and removal of the card. But still, the performance difference of 7-8°C is still too much for measurement error. So, the VF700-AlCu is definitely not better than ZM80D-HP.

I wouldn't recommend running at 5V unless silence is what you are looking for. Then again, even running at 5V will not make the VF700-AlCu completely silent.

I did a crazy test by not plugging in the fan connector. The temperature went up to an unbearable 114°C, about 40°C higher than with the fan at 5V! So, it is not possible to use the VF700-AlCu as a passive cooler.

If you want passive cooling, ZM80D-HP is what you should be looking at. Well, this is to be expected, since the VF700-AlCu has a lot less surface area than ZM80D-HP.