Cooling, Noise, and Power

Beyond the normal synthetic and in game benchmarks, most of which really only depend on clock speeds. There are a few tests where the different card manufacturers can really make their cards stand out. The first of those is power usage, so to test that I ran the Aero 1060 through Valley Benchmark and noted the highest wattage out entire system pulled using a Kill a Watt. The Aero GTX 1060 came in between two RX 470’s but was surprisingly high compared to the other GTX 1060’s by about 30 watts. Given the higher clock speed, I expected the EVGA to come in higher, but even at 238 watts, this is still more than low enough for any SFX power supply.

graph32

Next, I tested noise levels at both 50% and 100% fan speed but sadly our decibel meter died. I went with a new meter and changed a few things about our test procedure as documented in the test procedures section, but for now, the Aero GTX 1060 is the only card in our new results. With the new meter and procedures, the old results aren’t comparable. I will say that in my testing the fan on the Aero came in about the mid-range for noise, but I did also experience some coil whine, something I rarely hear these days.

graph33

The last set of tests were for cooling performance. I tested the card by warming it up looping Valley Benchmark. I did this twice, once with the stock fan settings and again a second time with the fans at full speed. This way we can see what you can expect out of the box and what the cooler is capable of if needed. In this case, the ITX sized card ran in line with the other GTX 1060’s at stock fan settings though I should remind you that the EVGA card did see higher numbers with their newer firmware that gave the card the ability to turn the fan off at lower loads. The Aero did have some extra headroom in it as well, so when I turned the fan up it did drop down 11 degrees, putting it below the long 1060 Founders Edition and the Zotac card.

graph34

graph35

 

Log in to comment

We have 1590 guests and no members online

supportus