Question Case fan upgrade question

Auchhior

Commendable
Oct 12, 2020
28
2
1,535
I'm looking at upgrading the fans in my build. Right now I'm running 3 Rosewill ROCF-13001s, two as intake and 1 as outtake(and 2nd outtake fan that came with the case.)

The ROCF-13001 has a 32 CFM, and I was looking either the Artic P12s or F12s. They both have CFM specs in the 50s and are pretty affordable at 31 dollars for a 5 pack. Is this a good move or would it not make that much of a difference? Also which one should I get for my use case? I was leaning towards the F12s.

I'd replace all three Rosewill fans, and use the extra Artic fan as a second one on my CPU cooler(Hyper 212 Evo).

Also, as a slightly related but different question, does anyone know what the rated CFM is for the exhaust fan that comes with the Courgar MX 330-G? I looked on their website and even asked them via the website, but haven't gotten an answer.
 

Auchhior

Commendable
Oct 12, 2020
28
2
1,535
Why upgrade?

What problems are happening?

If "not broke, then don't fix it"!
I bought the Rosewill fans on a tight budget because they were $13 for a pack of 4. I figured some air flow was better than none. My GPU and CPU both peak in the mid 80s for temperature when they are being pushed hard. I'd like to lower the temps, and better airflow in the case seems like the first step in that.
 

Ralston18

Titan
Moderator
Take a photograph or two of your case and fans. Mark up the photograph to show air flows into , through, and out of the case.

Post the photograph here using imgur (www.imgur.com).

Could be that a bit of rearranging things could improve cooling....

Especially if the fans are fighting each other or the overall air paths are convoluted. CFM becomes moot....

Someone may spot a problem and be able to offer other suggestions.
 

iPeekYou

Distinguished
Jul 7, 2014
392
77
18,790
I bought the Rosewill fans on a tight budget because they were $13 for a pack of 4. I figured some air flow was better than none. My GPU and CPU both peak in the mid 80s for temperature when they are being pushed hard. I'd like to lower the temps, and better airflow in the case seems like the first step in that.

Toyed around with fans a few months back, and I'll share my findings here. I hope it comes in handy. I'm discounting cases with solid front panels or otherwise poor intake airflow. In that scenario, the case needs replacing or modding first since it's the weakest link in the cooling department.
  • CFM for intake fans matters much less when you have at least 2 fans with good venting from the front. I have tested with cheap LED ring fans (the worst performers you can get really, rated at 30-40 CFM and non-existent static pressure), all the way to Corsair SP120 Performance and Scythe Wonder Snails at 2400 RPM. No change in temps with 2+ intake fans.
  • Keep in mind different manufacturers only rate CFM and pressure at ideal conditions. No use debating the specs if we can't see the PQ (pressure-airflow) curve. Even then, the curve of diminishing return is quite steep especially for case fans.
  • Good vents can be achieved with either 1) mesh paneling (or mere filters), and little obstruction behind it, or 2) solid front panel with unobstructed vents of about an inch wide. Alternatively, there's the side mounting like Lian Li does. Same principle applies, just to the side intake panel. Your case seems to have mesh panel, but I can't verify from pics that it's an unobstructed flow after the mesh.
  • 2 fans from the front is plenty; 3 fans can be better if you have powerful fans on a heatsink. You need all the CFM you can get to feed the heatsink fans, but it's nothing fatal. I see about 2 deg C of consistent gain from using 2400RPM SP fans on heatsink with good 70+ CFM fan at the top front. Short of that, or with an AIO, it matters little.
  • To expand a bit, you'd want to fill the front fan slots bottom first and working to the top. The third fan is nowhere as important as the first fan on the bottom.
  • Type of fans, generally SP-optimized for intakes and heatsink/rad, AF-optimized for exhaust. The less obstruction you have on intake, the more you can lean towards AF fans. I can confirm this on rear exhaust fan (-3 deg C on GPU, none on CPU and VRMs), on intake fans I don't see any consistent improvements.

TL;DR: get good front airflow, don't sweat the fan specs, 2 in 1 out is plenty good, and SP intake AF exhaust.
 
Last edited:

Auchhior

Commendable
Oct 12, 2020
28
2
1,535
Take a photograph or two of your case and fans. Mark up the photograph to show air flows into , through, and out of the case.

Post the photograph here using imgur (www.imgur.com).

Could be that a bit of rearranging things could improve cooling....

Especially if the fans are fighting each other or the overall air paths are convoluted. CFM becomes moot....

Someone may spot a problem and be able to offer other suggestions.
Here's a approximation of how its set up. I'm not at home so I can't take pictures of the PC itself. The CPU cooler is a CM 212 Evo with the stock fan on the right side. The front two in take and top exhaust are the Rosewill fans, the back exhaust is the fan that came with the case(Cougar MX 330-G)
dfSkXKt.png


Toyed around with fans a few months back, and I'll share my findings here. I hope it comes in handy. I'm discounting cases with solid front panels or otherwise poor intake airflow. In that scenario, the case needs replacing or modding first since it's the weakest link in the cooling department.
  • CFM for intake fans matters much less when you have at least 2 fans with good venting from the front. I have tested with cheap LED ring fans (the worst performers you can get really, rated at 30-40 CFM and non-existent static pressure), all the way to Corsair SP120 Performance and Scythe Wonder Snails at 2400 RPM. No change in temps with 2+ intake fans.
  • Keep in mind different manufacturers only rate CFM and pressure at ideal conditions. No use debating the specs if we can't see the PQ (performance-airflow) curve. Even then, the curve of diminishing return is quite steep especially for case fans.
  • Good vents can be achieved with either 1) mesh paneling (or mere filters), and little obstruction behind it, or 2) solid front panel with unobstructed vents of about an inch wide. Alternatively, there's the side mounting like Lian Li does. Same principle applies, just to the side intake panel. Your case seems to have mesh panel, but I can't verify from pics that it's an unobstructed flow after the mesh.
  • 2 fans from the front is plenty; 3 fans can be better if you have powerful fans on a heatsink. You need all the CFM you can get to feed the fans, but it's nothing fatal. I see about 2 deg C of consistent gain from using 2400RPM SP fans on heatsink with good 70+ CFM fan at the top front. Short of that, or with an AIO, it matters little.
  • To expand a bit, you'd want to fill the front fan slots bottom first and working to the top. The third fan is nowhere as important as the first fan on the bottom.
  • Type of fans, generally SP-optimized for intakes and heatsink/rad, AF-optimized for exhaust. The less obstruction you have on intake, the more you can lean towards AF fans. I can confirm this on rear exhaust fan (-3 deg C on GPU, none on CPU and VRMs), on intake fans I don't see any consistent improvements.

TL;DR: get good front airflow, don't sweat the fan specs, 2 in 1 out is plenty good, and SP intake AF exhaust.

That's really helpful! Thank You!
 

iPeekYou

Distinguished
Jul 7, 2014
392
77
18,790
Here's a approximation of how its set up. I'm not at home so I can't take pictures of the PC itself. The CPU cooler is a CM 212 Evo with the stock fan on the right side. The front two in take and top exhaust are the Rosewill fans, the back exhaust is the fan that came with the case(Cougar MX 330-G)
dfSkXKt.png




That's really helpful! Thank You!

Seems solid to me. If you ever wonder if you need more airflow from that point, just open the side panel and see if load temps drop >2°C consistently. Only in that scenario do you need to rethink airflow.

In good airflow scenarios, temps shouldn't change more than 2°C with side panel open, and might even get worse since we don't have a good, strong flow of air. My previous case exhibited this when I did just that, but that was with few mods and essentially 3x heatsink fans as intake.
 

Auchhior

Commendable
Oct 12, 2020
28
2
1,535
Seems solid to me. If you ever wonder if you need more airflow from that point, just open the side panel and see if load temps drop >2°C consistently. Only in that scenario do you need to rethink airflow.

In good airflow scenarios, temps shouldn't change more than 2°C with side panel open, and might even get worse since we don't have a good, strong flow of air. My previous case exhibited this when I did just that, but that was with few mods and essentially 3x heatsink fans as intake.
That makes sense. So should I just fire up a game I normally play, wait for it to reach the normal temps, then open the side panel to see if the temps change? Or check the temps with the panel on, then take it off and restart the game and see how high the temps go?
 

iPeekYou

Distinguished
Jul 7, 2014
392
77
18,790
That makes sense. So should I just fire up a game I normally play, wait for it to reach the normal temps, then open the side panel to see if the temps change? Or check the temps with the panel on, then take it off and restart the game and see how high the temps go?

Don't sweat the details, really. If your temps are good enough when doing your workload (games, rendering, etc), it doesn't really matter.

The ideal way, though, is using consistent loads from stress tests. That way, we have a known uniform workload and a more accurate temps measurement. Games fluctuate a lot in their usage; and temps will reflect that as well. For major cooling issues, games can serve as a practical load; but they're inconsistent if we're looking for minute delta in temps.

If you do want to test it, then fire up the stress test, monitor temps until it stops rising, then see what average and max temps are reported. For good measure, do the test 3-5 times and measure average. Or at least monitor the trend.
After the system cools back down, open the panel, then do the same thing. This is how I'd do it when I want to be thorough.

If you have a good idea of the temps, you can do it the "lazy" way. Run test with panel on, pop it open midway. If temps drop immediately and doesn't come back up, you've got your answer. I do this from time to time when I've run multiple tests already and don't feel like waiting long for more tests.

Softwares I use:
  • OCCT for CPU test. Small data set and AVX for maximum heat generated. Worst case scenario for heat, but also most consistent and revealing. Plus it's an all-in-one package of multiple stress tests.
  • Alternatives: Prime95, Intel Burn Test, AIDA64, Cinebench.
  • MSI Kombustor for GPU test. Not as hot as Furmark, but useful nonetheless. Stresses like a typical GPU-heavy game would.
  • Alternatives: Heaven Benchmark, 3DMark, Furmark, and OCCT's own.
  • HwInfo64 for monitoring. Generally accepted as precise, especially for Ryzens. Intel chips can use either this or HWMonitor, but HwInfo is held to a higher regard it seems.
  • Alternatives: none as well-accepted as HwInfo, but there's HWMonitor, CoreTemp, and even GPU-Z.
 
Last edited:
  • Like
Reactions: DRagor

Auchhior

Commendable
Oct 12, 2020
28
2
1,535
Don't sweat the details, really. If your temps are good enough when doing your workload (games, rendering, etc), it doesn't really matter.

The ideal way, though, is using consistent loads from stress tests. That way, we have a known uniform workload and a more accurate temps measurement. Games fluctuate a lot in their usage; and temps will reflect that as well. For major cooling issues, games can serve as a practical load; but they're inconsistent if we're looking for minute delta in temps.

If you do want to test it, then fire up the stress test, monitor temps until it stops rising, then see what average and max temps are reported. For good measure, do the test 3-5 times and measure average. Or at least monitor the trend.
After the system cools back down, open the panel, then do the same thing. This is how I'd do it when I want to be thorough.

If you have a good idea of the temps, you can do it the "lazy" way. Run test with panel on, pop it open midway. If temps drop immediately and doesn't come back up, you've got your answer. I do this from time to time when I've run multiple tests already and don't feel like waiting long for more tests.

Softwares I use:
  • OCCT for CPU test. Small data set and AVX for maximum heat generated. Worst case scenario for heat, but also most consistent and revealing. Plus it's an all-in-one package of multiple stress tests.
  • Alternatives: Prime95, Intel Burn Test, AIDA64, Cinebench.
  • MSI Kombustor for GPU test. Not as hot as Furmark, but useful nonetheless. Stresses like a typical GPU-heavy game would.
  • Alternatives: Heaven Benchmark, 3DMark, Furmark, and OCCT's own.
  • HwInfo64 for monitoring. Generally accepted as precise, especially for Ryzens. Intel chips can use either this or HWMonitor, but HwInfo is held to a higher regard it seems.
  • Alternatives: none as well-accepted as HwInfo, but there's HWMonitor, CoreTemp, and even GPU-Z.
Sounds good, thank you so much for your help!
 
  • Like
Reactions: iPeekYou