The biggest problem with coil whine is reproducing and measuring it. It not only depends on the card, but also on the PSU. The biggest factor is the FPS rate. The higher, the more coil whine, also depends on the actual FPS rate to find resonance.
Any ideas how to test this?
I've tested those RX 6900 XT on:
- 3 motherboards (X570 Gigabyte Gaming X | B550i Gigabyte Aorus Pro AX | B550 MSI MAG Tomahawk)
- 3 different CPUs (2x 5600X | 1x 3400G)
- 3 different PSUs (Fractal Design ION+ 860W | be quiet! Straight Power 11 1000W | Corsair HX850W)
- 2 different houses (230V) with and without UPS protection (APC Back-UPS Pro 1200 S | BR1200SI)
- 2 different monitors (BenQ MOBIUZ EX3415R UWQHD 144Hz | Samsung Odyssey G5-G55T UWQHD 165Hz)
SAME-EFFIN-COIL WHINE.
Then, at last, I've started tweaking the max GPU clock speed and I've started to see a pattern (thanks to the glorious GPU-Z): the lower the Watts needed, the lower the coil whine.
I've been using "The Forest" as a test-game because it easily triggered the noise from the cards:
- Downvolting the GPU did not affect coil whine.
- Lowering the Max power to -10% from the AMD Software did not affect coil whine.
- Then, when I started lowering the max boost clock, the noise got reduced. When I've set it to the reference 2250 MHz, the noise got low (but still present). The point was not the GPU clock, but the power needed. The more I was closer to the 200W, the more silent got the coil whine.
Hummm...
Check for coil whine with unlimited FPS, 60, 120 and 144 FPS.
60 - 144 - 240 - unlimited FPS on a 240Hz monitor should be better.
60 is for "i don't care" gamers.
144 is for "so my eyes can see past 60FPS" gamers.
240 is for "pro" gamers.
∞ is for "I can see anything" gamers.
But still, coil whine should be tested against resolution too.
I guess the best catch should be the Samsung Odissey G9-G95T.
With the resolution of 5120x1440p (~7.4MP) is almost dense as a 4K (~8.3MP).
It runs @240Hz with FreeSync and G-Sync.
EDIT:
or the new Odyssey Neo G8 32-inch 4K 240Hz
but then he would have to test with different PSUs and maybe even with 230 and 120 volt...
Good call about the 120V.
Guess the test should implement the best quality PSU available like the Seasonic Prime Ultra Titanium 850 W used in the review.
Then, IF there's coil whine, test the card with other two high-quality PSUs, but just running one test. If the coil whine is the same, there's no need to repeat all the iterations.
Yes, it's a hell of a job. I know.
At least add the info in the noise part of the reviews. "In this particular test setup, the card (did't/) suffered from coil whine".
That should help.