Since hot air rises, it’s always been recommended to set your top case fans to EXHAUST - y’know, to remove all the warm air from your CPU and GPU. But I recently tested this, and this case fan configuration can actually harm your computer’s thermals.

Which sounds counter-intuitive, but as I show off in this video, it actually sort of makes sense when you see the airflow analysis results (I used smoke matches - which are really fun to mess around with, by the way!).

I also done some stress tests with Prime95 and Furmark to record the exact thermal differences for four different fan configurations:

  • Fan Config #1 - both top fans are set to exhaust
  • Fan Config #2 - one top fan is intake, the other is exhaust
  • Fan Config #3 - both top fans are set to intake (not a good idea)
  • Fan Config #4 - zero top fans (I taped up the top fan grill)

(Link to direct YouTube video)

If you prefer text over video, please read on for the guide/transcript version of this video.

Video Transcript And Guide

Hey everyone! Can you tell me what’s wrong in this clip? So this is my ATX gaming computer, and it’s got space for eight fans: three at the very front, two on the PSU shroud, two at the top, and then one at the rear. The idea behind this is obviously that warm air rises, and so as we all know, case fans - especially at the top and the rear - should, you know, exhaust all of that warm air out.

However, as you’ve probably seen from that clip, that’s not actually happening because one of the exhaust fans is sort of stealing the air from the CPU cooler, which obviously could be affecting thermals. Actually, in this video I’m going to prove that it definitely does affect thermals.

I’m not the only one saying you shouldn’t have two top exhaust fans. Noctua recently published a guide saying the exact same thing - saying you should actually have one of your top exhaust fans flipped around as an intake.

Testing Methodology

That brings us on to the testing methodology, which is just a fancy way of saying I run some stress tests and then written down the results, to be honest. But essentially what I did in my Lian Li Lancool 2 case is I changed the configuration of the fans at the very top of the case.

So I left, you know, three intakes at the front, two on the PSU shroud, and one at the rear because they make sense - you know, I’ve done some testing before, but they make sense to keep them like that. But the debate is about the actual top exhaust fans, or the top fans in general: whether they should be exhaust or intake, or whether you should have them at all.

With all of those fan configurations, I done airflow testing with smoke matches, which are really, really fun to play with. I did have to tweak the colors so you could see the smoke a little bit, so that’s why my arm looks a bit radioactive - so I’m sorry about that. I’m not sort of some mutant or something; it’s just to tweak the colors. But I did some smoke testing so you could see the actual airflow.

Then I also run Prime95 and also FurMark. So obviously one stresses the CPU and one stresses the GPU. I know Prime95 is a slightly older test compared to like AIDA and tests like that, but actually I still like it because it, you know, warms up the CPU a lot and stresses it quite a lot. So that is the test methodology.

Test 1: Two Exhaust Fans

The first test I actually done was to have two exhaust fans. Usually in this particular case, what I actually have is an all-in-one cooler from Corsair with the radiator at the very top, but I’ve got rid of that temporarily for this test. So I’ve gone back to a Noctua air cooler, and that then freed up the actual two top slots where I then put Corsair 140mm fans. It’s them that I actually, you know, rotated around in different configurations to bring all of the test data to basically prove that two top exhaust fans can actually harm your computer’s performance.

Without further ado, let’s actually get on with the test results!

Smoke matches are always fun to play with. If you vape, you can also use that, although it may coat your computer components in a sticky residue, so you probably don’t want to do that. But anyway, this is the bottom intake fan, which is mainly going to cool the graphics card. The middle one then - and I know it looks like the GPU is causing issues, but this is fairly normal, you know. I’m just trying to visualize the air, and this air is going to cool both the GPU and the CPU. But this top intake fan at the front of the case is obviously doing nothing - it’s just… well, really the top exhaust fan at the very top of the case is stealing all the air.

Next, let’s look at the actual stress test results. We got a baseline for what’s actually going on. As I mentioned earlier, I ran both Prime95 and also FurMark. The CPU in this system is the Ryzen 7600X, so a six-core, 12-thread CPU. Actually, I run Prime95 for enough time for it to actually heat up, and, you know, then I left things cool down before running the next test, which was obviously FurMark. My graphics card here is the RX 6700 XT, so again I run that for a certain amount of time and I track the results, which you can see here.

Now obviously in isolation these don’t mean much - which I’m going to be doing comparisons in just a second - but the CPU is in green: 60° under Prime95, dropped out to 47.4° with FurMark. Then you can see the GPU one as well. So they are our baseline results.

Test 2: One Intake, One Exhaust

Now that we have our baseline results, by doing what most of the internet actually tells us to do, what I did was actually take one of the exhaust fans - the one closer to the front of the case, you know, between the front of the case and the CPU cooler (and that’s an important point I’ll come back to later) - but that particular one I actually flipped around and set to be an intake. That then matches the diagram on the Noctua website.

So let’s take a look at the airflow and the temperature data for that.

So now we can see the airflow for the bottom intake fan at the front. Actually, I’m not going to show the exact airflow the entire time ‘cause it’s the same as the first test, so we’ll just jump ahead to that top fan. As you can see here, the air is not being sucked straight out. It’s a bit hard to see exactly where the air is going, but clearly it’s not going straight out of the top of the case, which is a good thing. You know, there’s more air inside the case. This is a front intake, and so it’s cool air from, you know, the room actually coming in, and it’s going to be cooling the CPU, which is brilliant.

These are the actual results - you don’t have to remember what they are before ‘cause I’m going to show a comparison - but just wanted to show the absolute results there. Now we’ll move on to the comparison.

This is just the CPU now. The CPU was better with fan configuration two, so I’m going to the intake. So in the yellowy color there, you can see that the CPU was cooler in both tests, and it makes sense that it was, you know, more cooler in FurMark than with Prime95, because Prime95 will stress the CPU a lot more, so those temps will be higher.

When we look at the GPU temps and compare the two, you can also see a clear difference. But again, more with Prime95 - FurMark again is going to stress the GPU more - but even then, you know, in this case you can see that with the intake fan at the top, the graphics card responds better to that with both Prime95 and FurMark.

So that’s a clear win with all four of the temperature points actually looking better, and I think the airflow analysis looks better as well. I mean, it’s harder to see, but actually, you know, you don’t see some of the air coming straight out of the case. And then obviously the CPU temp results actually prove that that is a better configuration. So that’s great!

Test 3: Two Intake Fans

But now I wanted to play double or nothing and go to two intake fans and actually see whether that would help or not.

So this is the airflow for test number three with both intakes. Actually, you can see the air is not going directly out the case. If we bring out fan config number two with the intake, it’s hard to actually tell a difference. But I did actually feel - with the main footage with the two intakes - I did feel that the air was kind of hanging around the case a little bit more. It’s hard to put my finger on it exactly, but it did just feel like the airflow wasn’t as good.

So again, these are the absolute results, which again don’t mean too much by themselves, but I just wanted to show them here. Then we’ll go to the comparison.

Oh no! Something bad happened under the CPU results with FurMark, and I’ll talk about that later on. I did rerun the result multiple times, but actually the CPU was just really unhappy with FurMark, and I think it’s to do with the fact that all the hot air from the GPU wasn’t being cleared whatsoever.

But actually, you can see even Prime95 wasn’t as good as having an intake fan there. Then we come to the graphics card, and yeah, no real difference actually. So it was mainly really the CPU that seemed to be struggling when we had two intakes - you know, no real change on the GPU.

Okay, so that’s actually kind of a strange result. The GPU data was basically the same, as well was the CPU data on Prime95, which I sort of expected because Prime95 is going to heat up the CPU a lot, and so the case fan configuration won’t, you know, change things as much. But obviously what we’ve seen was a very weird result with the CPU temperatures when running FurMark.

Actually, you know, at first I thought it was a mistake and I rerun the test multiple times, and I kept getting the same result. I think what was probably happening is that all of the, you know, temperatures from the graphics card - all that hot air was coming out - but then instead of usually that hot air being exhausted out either by the top or the rear, what was probably happening is those two top fans which were set to intake were bringing all that cool air in. But essentially, I think maybe it was trying to like trap the hot air from the GPU or something like that.

If you’ve got your own ideas about what might be happening, then please let me know down in the comments, because yeah, it’s a bit of a weird one. But really, you know, don’t set both of your top fans to intake - I think that’s the conclusion.

Test 4: No Top Fans

But next I wanted to actually try not having any top fans at all. So I actually kept the fans there just ‘cause I was a bit lazy, but what I did instead is: firstly, I went into Fan Control - so my software on my Windows computer - and I actually set the fans to zero so they weren’t running. Then I actually taped up that top completely so that no air was actually coming through it.

So now we can look at the test results to actually see what happens in this configuration. In this case, I’ve shown the results in real time - I’ve gone back to 100% speed and I won’t put any comparisons on the screen to confuse anyone. But actually, yeah, you can see the airflow here. It looks certainly healthier than when we had two intakes - everything, you know, the airflow seems to be a little bit better. Obviously, you know, this is a concerning fan configuration.

So next we look at the results. Again, these are the absolute ones, which don’t mean much by themselves, so let’s look at the head-to-head comparisons.

So our new bluey sort of color is the one we just done. Actually, you can see that, you know, for Prime95 on the CPU, it’s not a bad result at all. But for FurMark, you know, it’s one of the worst results - not the absolute worst results, but it’s worse than the other two fan configs. Then for the graphics card, it is the worst result overall. You know, the graphics card really does seem to be struggling with just having no top fans whatsoever. It’s the worst result in Prime95 and also FurMark.

Conclusions

So kind of a mixed bag of results, really. I mean, in general, you know, I run these tests multiple times, and it was clear that actually, you know, doing what Noctua said is the best thing overall - which is not completely surprising since Noctua are the GOATs for things like this.

But actually, I think this does make sense, because if you have, you know, a top fan between your front of your case and also your CPU cooler, obviously, you know, if that top fan is set to exhaust, then, you know, it just makes sense that actually any cool air coming in from the front of your case is going to be exhausted straight out before it can hit your CPU cooler.

Obviously, if you’ve got a particular case where, you know, one of the top fans - maybe you’ve only got one top fan - if that’s sort of right above your CPU cooler, or maybe a bit to the right of it or the left of it depending on which way you’re looking at it, then actually in that case that should be set to exhaust. I think that’s kind of clear, and certainly my data shown that too. It was very bad when that was set to intake, ‘cause yeah, that just caused things to go a little bit haywire in both the airflow and the temperature data.

But overall, you know, I found this kind of interesting. And, you know, previously I have set my top fans to exhaust just by default because that’s what the interwebs always says, but now I know not to do that. So going forward, you know, I’m not doing that in my case anymore, so that’s really good to know actually.

But if you’ve got any further questions or thoughts about anything I’ve covered, please do let me know in the comments - especially if you’ve got thoughts about, you know, the double intake, you know, thing driving up the CPU temps during FurMark, ‘cause that was really interesting.

But if you enjoyed this video, please click the thumbs up button and please subscribe to see more videos like this. Thanks for watching!