Fan Question

Associate
Joined
23 Aug 2010
Posts
237
Hi everyone.

Long time visitor, occasional poster.

I am trying to balance noise vs performance on my rig (9700k & 2080ti, fractal nano s) and due to my office being freezing in winter and a sauna as soon as spring hits this is quite hard.

I am currently have corsair ML fans (3x120 and 2x140) running, but they are quite loud when running at any decent speed for cooling.

Is it worth migrating my fans to noctua ones, or will there be no real difference in noise?

On paper they move a similar amount of air, with less noise, but I am interested in any real world experience people have.
 
I did a bad and bought the noctua chromax fans. Of course they still make a noise, but nowhere near as much as the corsair ones. Even when they rev up though, the sound profile is just better, far quieter overall
 
1/. What rpm are you running these fans at?
2/. What do you have them blowing through, ie radiators, heatsinks, etc?
3/. What sort of temperatures are you getting, compared to room temp, on the CPU, GPU and (if you have it) water?

A curve from ~800 through (now) to 1400 on the front fans (the 140s) the same on the 120s strapped to a top mounted rad. A rear fan (bequiet) that runs at a steady speed because they are quiet even if they don't move that much air.

All fans are linked to a commander pro and controlled through icue. Not the best solution maybe, but it does the job.

Temps are good, 2080ti idles at 27 to 35 depending on ambient temps and if rtx voice is running (that adds ~8c to any temps). Processor (9700k) idles at 35 to 40.

Under load cpu gets to between 60 and 70 depending on ambient, the gpu tops out at high 70s again on ambient.
So I'm not troubled by any of that, just want a decent fan noise if I can get one. The gpu runs a reasonable fan curve where I try not to get the fans to rev too much.

Trouble is the swings in ambient temp in my office is part of the problem, one I can't fix easily. It can be 6c in winter, and go north of 20c in spring/summer. That can trigger the gpu to hit 81c, which is the point where the fans go full steam ahead with killing my eardrums. So that's the next part to look at.

I am going to replace the paste on the gpu because I want to, and because there has to be something I can do to stop the fans getting too loud. I would love to go full water, but it would be hard to slip that amount of kit (and cash) past wifey...
 
That's a bit concerning... I'm sat in 26ºC but my GPU won't go above 65º under load. I have uncomfortably high idle temps, but the difference compared to load values is quite small.
How many fans are intaking vs how many exhausting?

Yep, hence me going to brave taking the card apart and redoing the thermal paste. I could RMA it I suppose, but tbh I would rather take care of it myself as it is probably quicker and easier.

How many fans are intaking vs how many exhausting

Now I have:

INTAKE
2x140mm Noctua (https://www.overclockers.co.uk/noctua-nf-a14-pwm-chromax-premium-grade-fan-140mm-fg-06l-nc.html)
1x120mm Corsair (running at fixed 800rpm to get some air direct on to the GPU. This will go after I redo the thermal paste

EXHAUST
2x120mm Noctua through a 240mm Corsair H100 AIO (https://www.overclockers.co.uk/noctua-nf-f12-pwm-chromax-premium-grade-fan-120mm-fg-06n-nc.html)
1x120mm Bequiet on the rear of the case

The 140mm intakes are curved to run faster than the exhausts
 
Last edited:
This is basically what you currently have:
DeavuUj.jpg

That's exactly what I have (good paint skills) except the GPU spreads across the whole area, so the bottom fan pushes air on to the last third(ish) of the card. The 2 140mm effective push air through the top half of the case and the bottom, if you assume the GPU effectively creates 2 zones.
There is no room for push/pull on a radiator and I moved the radiator to increase the airflow through the case over the GPU. Putting the rad in the front restricts airflow, so the bottom fan helps for sure as well. Does that make sense?

I like the idea of stopping the rear fan as things start to ramp up, and directing all air out the top though. I wonder how I can do an inverse curve...

I had seen that reddit post, it is very interesting for sure
 
Last edited:
It does make sense to some degree, but the bottom fan isn't really solving the problem - it's blowing a little air at part of the GPU, instead of having air pulled across all of the GPU.

I agree completely, and I do want to remove it soon. But I do think I need to repaste the card. If nothing else I would like to make sure it is done properly and the card is tightened down properly, something which (according to other forums can be a problem). It is not overclocked either.

All the other temps are ok (VRM, RAM etc. don't go above 55c on a GPU stress test run) so it does just seem that the GPU has an issue when being at load.

I have also tried removing the front filter on the case, and tbh there is no difference in temps
 
It's on the switch on the Nintendo online service if nothing else.

Noise from the fans is fine now, even if they are running at max. The noise profile of the noctua fans is more pleasing to my ear as well. Noise from the gpu fans, not so much when they go full tilt. So that is the next focus for me
 
Ok, the experiment for today is done. It was the harder option, that being to repaste the GPU.

I wanted to do this because I am stubborn basically, and if it didn't make a difference then I will sort out fans afterwards. I hate messing around with fans more than I hate taking apart graphics cards, and yes, I know the potential for error is much higher taking apart a GPU than moving fans.

So, the results...

With the fans still in the same positions in the fractal nano s (2x140mm in the front as intake, 1x120mm in the bottom as intake, 2x120mm in the roof with the AIO as exhaust), following a repaste the temps on a 3DMark stress test went down from 81c at peak to 76c at peak. This seems like a good result.

Tomorrow's test will be to move fans around and see if I can get that lower...

@Armageus thanks for all the ideas and help to me and others who have asked
 
And the results (for experiment 2) are in

Setup:

AIO in the front (thus 2x 120mm fans intake)
1x 120mm bottom fan intake
1x140mm top exhaust
1x120mm rear exhaust

CPU temps were good (as they were previously)
GPU temps were 2c higher at a peak of 78c

Ambient temp between the two tests was reasonably consistent (maybe a degree in it)

This means that (for me at least) the following is the winner for my particular setup (mainly because it makes more sense to me, and gave the lowest GPU temp which is what I care about atm)
2x140mm front intake
1x120mm bottom intake (pegged to 800rpm constant)
2x120mm top fans exhaust on the AIO
1x120mm rear fan which I may or may not run

Proof positive that each case and setup is different, and trial and error is the way forward.

That said, I do recommend the noctua fans. They sound good as well as performing, whichever case/orientation etc you have them in.

Be back later... I have some fans to move around... Again...
 
And also that what looks good in theory, doesn't necessarily correspond with practice :D

Yeah, sorry about that.
I was not going to try your way, but I did, honest.

I appreciate there was only a couple of degrees in it, but I have to go with what runs cooler for my GPU since I cannot do much about those fans without water as I said before.

Besides, it's nice that I was right at least this once.

With a decent fan curve I get plenty of air in now, quietly, and when running a CPU bench (like OCCT) I cannot get the CPU temp to go above 70c. So I am calling it a win, and I am not going to mess around with unscrewing fans again (well for a while anyway)
 
Back
Top Bottom