Potential Cause of Coil/Electrical Whine in GPUs Under Load

They make noise, too.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
nintendoeats
Posts: 59
Joined: Sun Feb 22, 2015 11:57 am

Potential Cause of Coil/Electrical Whine in GPUs Under Load

Post by nintendoeats » Sun Nov 12, 2017 11:46 am

For a while now (months) I have been dealing with increasing noise from my GTX 1060 while under load. As frequency increased, so would the pitch of the noise. I run F@H on the GPU at all times, so this was a real problem. However, I believed there was nothing I could do about it. I have just discovered that this was not so and would like to record my experience on the internet in the hopes that it will help somebody.

My 1060 Windforce OC was equipped with an Arcitc Accelero Xtreme IV, which takes it from incredibly loud to essentially silent at over 2.1 Ghz. However, the weight of the cooler created a great deal of droop, so I suspended the cooler at the unmounted end using a pair of zip-ties run through a handy hole in the cooler. This worked fine for my old 7970 (which also had an Accelero on it) for years, and for quite some time worked with the Windforce.

For reasons that are not very interesting I replaced the Windforce with an ROG Strix 1060 in the hopes that the buzzing would go away. After quickly finding the performance and noise from the cooler on the Strix unacceptable I put the Accelero on it and returned to the zip-tie system. To my dismay the noise was even worse than with the Windforce. It was so loud that I could hear it from the other room with the door closed.

I developed and tested many hypothesis over the course of a week and have finally resolved the issue. Despite the fact that both cards were working perfectly the stress placed on on either the GPU or the PCI slot by the zip-tie system was creating exceptional amounts of noise. Fortunately the backplate on the Strix gives it a great deal of strength and it doesn't sag nearly as badly as the Windforce. On its own the GPU (or possibly the socket) emits only a tiny amount of electrical noise which I could call essentially normal and which is effectively muffled by the case.

The Accelero Xtreme IV comes with a bracket for holding GPUs straight. unfortunately it is only compatible with standard width PCBs. This almost exclusively limits it to reference boards.

I hope that this sheds new light on potential sources of noise from GPUs.

Notwist
Posts: 18
Joined: Mon Feb 27, 2017 3:05 pm

Re: Potential Cause of Coil/Electrical Whine in GPUs Under L

Post by Notwist » Thu Nov 16, 2017 2:08 pm

I'm a little confused. Coil whine occurs even in cards that don't have custom coolers on them, and MOST people don't use zip ties to secure their card, so I'm not sure how this helps us figure out the "cause of coil whine" under load. =\

nintendoeats
Posts: 59
Joined: Sun Feb 22, 2015 11:57 am

Re: Potential Cause of Coil/Electrical Whine in GPUs Under L

Post by nintendoeats » Thu Nov 16, 2017 2:11 pm

Notwist wrote:I'm a little confused. Coil whine occurs even in cards that don't have custom coolers on them, and MOST people don't use zip ties to secure their card, so I'm not sure how this helps us figure out the "cause of coil whine" under load. =\
That is not my thesis. What I found is that torsional stress on the PCB and socket can be a cause of electrical noise from a GPU. That is all.

Wild Penguin
Posts: 83
Joined: Mon Jul 01, 2013 3:30 pm

Re: Potential Cause of Coil/Electrical Whine in GPUs Under L

Post by Wild Penguin » Mon Dec 18, 2017 11:43 pm

If you can hear it to the other room, you must have been experiencing some very exceptional electric noise!

Perhaps caused by poor contacts at the slot (caused by the mechanical stress).

That being said it is unlikely this will help users experiencing regular coil whine. Still, there could be users who have their graphics card sagging and in that case your advice can still be worthwhile!

Post Reply