intake and exhust fans, how many?

humanevil · 29 Apr 2018 at 22:58

atm i have 3 intake (3 front) 3 exhust (1 rear 2 top) and i was wondering would i get better cpu temps if i made the two top fans intake as well ? for a total of 5 intakes as i have my aio rad at the top.

spec:
i7 6700k @4.6
16 gb ddr4 ram
evga ftw 1080
colermaster cosmo 700p case
asus maximus 8 ranger mobo
cooler master 280 aio

Ross Thomson · 30 Apr 2018 at 08:43

I'd move your rad to the front as an intake

Bugbait · 30 Apr 2018 at 09:15

Your case has ventilation holes/grills at the rear so you can try positive pressure, so switch top 2 to intake. Always better to have cool, external air (intake) flowing through any radiators. Positive pressure tends to reduce dust inside the case as well.

doyll · 30 Apr 2018 at 13:30

Why is it people often say they want cool air flowing through radiators, but almost never say anything about the heated air from radiator as intake is what is supplying the GPU and other components not being cooled by the CLC on CPU. \Using a CLC radiator as intake when it's only cooling CpU defies logic. :mad:

I would setup case with 2x front and 1x bottom 140mm intakes decent pressure rated fans (1.3mm H2O @ 1300rpm minimum), remove all PCIe back slot covers (more rear venting around GPU for better front to back flow and less upward flow of GPU exhaust), top radiator exhaust. Have front intakes speed control based on CPU heat and bottom intake based on GPU temp. This is to have case fans cycle with component GPU cooler fans and CPU radiator fans and supply cool airflow as needed. Experiment with back fan exhaust or not running and see which gives lowest CPU temps.

Bugbait · 30 Apr 2018 at 13:38

doyll said:
Why is it people often say they want cool air flowing through radiators, but almost never say anything about the heated air from radiator as intake is what is supplying the GPU and other components not being cooled by the CLC on CPU. \Using a CLC radiator as intake when it's only cooling CpU defies logic.

Good point

. In my haste I assumed the GPU was being water cooled as well. Oops! In my mind I find AIO CPU and air cooled performance GPU a bit bizarre.

Why would you prefer static pressure over flow for the fans though?

doyll · 30 Apr 2018 at 13:57

Bugbait said:
Good point . In my haste I assumed the GPU was being water cooled as well. Oops! In my mind I find AIO CPU and air cooled performance GPU a bit bizarre.

Why would you prefer static pressure over flow for the fans though?

LOL. You were not the only one suggesting radiator as intake. Yeah, in a custom loop the air does not warm up a lot because coolant never gets very warm, but in CLCs it runs warmer .. probably because their pumps are cc 40-60 L/h while custom loop pumps are a couple hundred L/h and higher. Also CLC radiators are not near as good and not near as much area as typical custom loop using much bigger or multiple radiators. almost always both GPU and CPU in the loop.

Fan spec are not much good. CFM is taken on fan running with 0 resistance .. often in a chamber system where the exhaust side chamber is slightly lower than intake side so the air 5cm in back of fan is same pressure as 5cm in front of fan .. a condition impossible to have in normal use. Static pressure is amount of pressure fan makes pushing air into a sealed chamber .. again not a way we use fans. Our airflow is needs are somewhere in-between airflow and static pressure specificatons. A fan rated 100cfm and 0.5mm H2O behind a grill and filter creating 0.5mm resistance will flow almost no air even at full speed because the resistance is the same as it's rating. But a fan rated 50cfm and 2.0mm H2O will flow 40cfm at full speed and about 20cfm at half speed. As long as fan are speed controlled and will idle at 500-600rpm they cannot have too high a maximum rpm, cfm or mm H2O.

Hopefully you can understand what I'm trying to say. If not, please ask questions.

Bugbait · 30 Apr 2018 at 14:42

Thanks for the detailed response, makes sense. Now I'm wondering if I should use the Noctua F12 fans instead of the P12 fans for the intake as well. Hmm.

doyll · 30 Apr 2018 at 14:57

Do you already have NF-P12 fans? They are good with 1.68mm H2O @ 1300rpm. NF-F12 is only a little better with 1.83mm H2O @ 1200rpm and noise levels are very similar.

What case do you have and what fan placement?

Bugbait · 30 Apr 2018 at 15:03

Case: Corsair 900D. I have 4 NF-F12's on the top radiator (360) and another 4 on the bottom radiator (480) on one side, all are in an intake, push configuration. Then 2 NF-P12's on the front for intake across the drive bays. The front filter is restrictive with the cover on though. All fans are intakes with the single 140mm ass exhaust on the rear. Custom loop, CPU and VRM monoblock, full cover GPU block and dual D5's.

Current temperatures at idle:

Water: 22.47C
Motherboard Front: 23.91C
Motherboard Back: 26.53C
HDD cage (front): 21.29C
Bottom: 21.61C

These are all from external temperatures probes running into an Aquaero 6.

doyll · 30 Apr 2018 at 15:37

Your temps are good, but just to paly devil's advocate, how is all the intake airflow potential of your 9 intake fans trying to pull air into case supposed to get out? 900D has what .. 16x vents? Even if the 4th vent in top is exhaust it's just circling air back into radiator intake next to it. I didn't know 900D had bottom vents .. do you mean in one side and out the other side of bottom chamber?

Instead of explaining airflow I'll give you a link to a guide about case airflow. It's mostly for air cooling but applicable to water cooling too.
https://forums.overclockers.co.uk/t...-i-put-my-temp-sensor.18564223/#post-26159770

EsaT · 30 Apr 2018 at 15:56

Bugbait said:
Case: Corsair 900D. I have 4 NF-F12's on the top radiator (360) and another 4 on the bottom radiator (480) on one side, all are in an intake, push configuration. Then 2 NF-P12's on the front for intake across the drive bays.

That sounds exhaust limited with fans likely basically "fighting against each others" in pushing air into case.

doyll said:
Do you already have NF-P12 fans? They are good with 1.68mm H2O @ 1300rpm. NF-F12 is only a little better with 1.83mm H2O @ 1200rpm and noise levels are very similar.

Which again isn't very useful, because fans aren't pushing air into closed box...
P-Q curve would be only true useful data.
http://www.arx-group.com/pq.html

And for home use of course noise level per airflow would be also needed.
While advertised dBs are often about as honest as tax returns of Trump&Co.

Bugbait · 30 Apr 2018 at 16:07

Most of the rear of the case, including the blanking panels are perforated. There's room for 2 PSU's at the bottom and the blanking plate on that is also perforated. If I do the tissue test, it bends outward at the back to varying degrees so air is being pushed out the back at all points.

The 900D has 4x120 on top, 4x2x120 on each side at the bottom, 1x140 at the rear and 3x120 at the front. The only places I don't have fans is on the bottom right (4x120) and I had to remove a 120mm from the front to fit the 480 at the bottom. The bottom right side is used to house the dual pumps, PSU and some cables. Filters are installed at all locations except for the rear 140mm. So there's 10x120mm intake fans and a single 140mm exhaust. Although, as mentioned above, the rear, from top to bottom is perforated. The second picture from the left is a good picture of the rear: https://www.overclockers.co.uk/corsair-obsidian-900d-super-tower-case-cc-9011022-ww-ca-120-cs.html

The bottom air would likely escape out the back or middle. The side directly opposite is pretty cluttered and has an air filter fitted. Up and back would be less resistance.

EDIT: Typo.

doyll · 30 Apr 2018 at 16:07

In my defense my post you quoted was before we know case or case fan placement.

NF-P12 and NF-F12 perform similarly at same rpm in all respects, including sound level .. and Bugbait's temps are all very good being only a few degrees assuming room ambient is normal 20-23c.

doyll · 30 Apr 2018 at 16:11

Bugbait said:
Most of the rear of the case, including the blanking panels are perforated. There's room for 2 PSU's at the bottom and the blanking plate on that is also perforated. If I do the tissue test, it bends outward at the back to varying degrees so air it being pushed out the back at all points.

The 900D has 4x120 on top, 4x2x120 on each side at the bottom, 1x140 at the rear and 3x120 at the front. The only places I don't have fans is on the bottom right (4x120) and I had to remove a 120mm from the front to fit the 480 at the bottom. The bottom right side is used to house the dual pumps, PSU and some cables. Filters are installed at all locations except for the rear 140mm. So there's 10x120mm intake fans and a single 140mm exhaust. Although, as mentioned above, the rear, from top to bottom is perforated. The second picture from the left is a good picture of the rear: https://www.overclockers.co.uk/corsair-obsidian-900d-super-tower-case-cc-9011022-ww-ca-120-cs.html

The bottom air would likely escape out the back or middle. The side directly opposite is pretty cluttered and has an air filter fitted. Up and back would be less resistance.

Indeed, you have good temps. I was just projecting my OCD.

Out of curiousity, what are our thoughts about aquaero 6?

Bugbait · 30 Apr 2018 at 16:21

EsaT said:
That sounds exhaust limited with fans likely basically "fighting against each others" in pushing air into case.

I'd agree if the back wasn't full of holes. The 140mm exhaust fan is positioned in the usual place next to the CPU slot. With a monoblock the space is very open and clear. Below the GPU the space is also relatively clear since a full cover block is a lot thinner than an equivalent air blower. The GPU is the only card in any slots so there's a nice cavity there which is in line with the front intake fans. All the blanking plates are well perforated as well.

EDIT: I could improve the front airflow by removing the cosmetic front panel but I like the look of it. Removing the HDD would help as well but that's not practical given I'm using 3/6 bays and the upper, empty one is used as a mounting point for my reservoir.

doyll said:
Indeed, you have good temps. I was just projecting my OCD.

Out of curiousity, what are our thoughts about aquaero 6?

Sorry, don't mean to come across as defensive, I'm here to learn and improve my set up as well as others

. I just don't agree with the "closed box" scenario from @EsaT given the open back of the case.

If the cost isn't an issue the Aquaero 6 is fantastic. Not perfect but a dream for people who like to tweak. I have physical temp monitors as well as software ones (reads from HWInfo and others). Custom and fully independent temperature curves for fan and pump control, LED lighting tied to case and water temperature for visual indication. I wouldn't build a custom water loop without one again

.

Software is only used to configure the device, all monitoring is internal so it will work even if the OS crashes or freezes. You can set it to shut down the system (needs additional plugs) or fire off an audible alarm which is also on board.

Bugbait · 30 Apr 2018 at 16:26

When gaming for hours the temperatures increase proportionally. So the MB Front, HDD, and Bottom always stay lower than the water temperate with the MB Back being higher. The MB Back probe sits behind the CPU socket and there are loads of cables behind the tray and very limited air flow so that's expected.

doyll · 30 Apr 2018 at 18:03

Thanks for the feed-back about aquaero. Yeah, only complaints I have against aquaero 5 is not enough PWM channels and voltage control in aquaero 6 is much better/no heat buildup at low voltage like on 5. I don't know of any other fan controller that is even a tiny shadow in comparison to what aquaero are.

Bugbait · 30 Apr 2018 at 18:07

doyll said:
Thanks for the feed-back about aquaero. Yeah, only complaints I have against aquaero 5 is not enough PWM channels and voltage control in aquaero 6 is much better/no heat buildup at low voltage like on 5. I don't know of any other fan controller that is even a tiny shadow in comparison to what aquaero are.

I moved from the AQ5 to 6 for more PWM channels. Now have radiator fans on one channel, intake on another and each D5 has a channel. Wanted D5's on own channels for failure monitoring, otherwise could easily daisy chain them.

doyll · 30 Apr 2018 at 18:13

Bugbait said:
I moved from the AQ5 to 6 for more PWM channels. Now have radiator fans on one channel, intake on another and each D5 has a channel. Wanted D5's on own channels for failure monitoring, otherwise could easily daisy chain them.

Most all of my systems are air cooled .. except one with be quiet! Silent Loop 280 on CPU.

Couldn't you use any rpm input to monitor 2nd D5 and have both on same PWM signal? Not needed now but if you ever wanted to control something else on it's own PWM controlled curve.

Ross Thomson · 1 May 2018 at 08:50

Bugbait said:
Good point . In my haste I assumed the GPU was being water cooled as well. Oops! In my mind I find AIO CPU and air cooled performance GPU a bit bizarre.

Why would you prefer static pressure over flow for the fans though?

made no difference in my case before I watercooled the GPU. The card still has a 90mm fan on it cooling the VRM's the the kraken G12. no issues with temp, they are 10C cooler than when it was on air. The air that comes out of the rad for my CPU isn't hot.