Air flow balance: should intake and exhaust flow be equal?

Control: management of fans, temp/rpm monitoring via soft/hardware

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
danielG
Posts: 68
Joined: Thu Jul 17, 2008 3:40 pm
Location: Leiria, Portugal

Air flow balance: should intake and exhaust flow be equal?

Post by danielG » Fri Jul 18, 2008 4:18 pm

I've bought an Antec P182 case. I've also mounted a middle fan in the case. All fans output about 40 CFM.

As I have two 40 CFM exhaust fans, should I to use a 80 CFM intake fan to balance the flow? 80=40+40? Since I have a 40 CFM fan for intake, doesn't it block and slow down the airflow of the two 40 CFM exhaust fans?

I'm confused. Unfortunately, all my knowledge of fluids is about water (hydraulics).

cmthomson
Posts: 1266
Joined: Sun Oct 09, 2005 8:35 am
Location: Pleasanton, CA

Post by cmthomson » Fri Jul 18, 2008 5:44 pm

Well of course the air inflow must equal the outflow, regardless of rate.

But that's not really the issue. The P180/182 suffers from a basic design flaw: the inlet cross section is less than half that of the outlet cross section. If you want to have a very quiet P180/182 system, you need to either use only one outlet fan (and block the other outlet), or create a second inlet (typically with a Kama Bay or equivalent).

What you don't want to do in a quiet system is have both inlet and outlet fans! This adds lots of noise and no extra air flow.

zoatebix
Posts: 99
Joined: Thu Jun 08, 2006 1:57 pm
Location: Virginia
Contact:

Post by zoatebix » Fri Jul 18, 2008 6:32 pm

Running fans close together in series is a problem, though it's generally fine with the fans on opposite ends of a case, I thought.

Exel
Posts: 90
Joined: Tue Jun 27, 2006 3:25 am
Location: Finland

Post by Exel » Sat Jul 19, 2008 2:44 am

I'm not sure if the intake fan is a hindrance really, but it might be unnecessary. To move the same volume of air as you would with the intake fan, you could run the two exhaust fans at higher rpm for the same noise level that the intake fan would produce.

Of course, if you are using the intake fan to generate extra airflow locally, be it for HDDs or GPUs, it's another matter entirely.

But I too would like to know the answer to the original question, since I've recently installed an intake fan for the latter purpose - am I in doing so in fact reducing the overall case ventilation?

Luminair
Posts: 223
Joined: Wed Mar 21, 2007 10:45 am

Post by Luminair » Sat Jul 19, 2008 5:22 am

> should intake and exhaust flow be equal?

It depends on what you want. If the flow is not equal between active inputs and outputs, and the rest of the case is totally sealed, then some of the fans are working harder for no intended purpose.

If the case is NOT totally sealed, then you have to ask yourself if the flow difference that is going through those unsealed parts is what you want. For instance, a P180 with more output than input and the default perforated slot covers results in air getting sucked in through those slot covers. Should the intake and exhaust flow be equal? Well, do you want flow through those covers? Do you want flow through the vent at the bottom, or the other open holes in the case?

danielG
Posts: 68
Joined: Thu Jul 17, 2008 3:40 pm
Location: Leiria, Portugal

Post by danielG » Mon Jul 21, 2008 4:19 pm

@cmthomson: mounting a second intake in the 5.25'' drive bays is a good idea. Hadn't thought of that.

Unfortunately, without a middle fan my motherboard overheats and the network controller stops working. The P182 seems to acumulate hot air in the graphics card area. My motherboard has crap heatsinks too, but that's beside the point.

Should I really remove the middle fan and bump up the rear fan speed?

I was thinking of changing the Tricools for 800 rpm Scythes (as advised by Luminair to another P182 user), but I'm not so sure that's enough. I'm currently running the rear Tricool at 1600 rpm because the ambient temperature in my room is too high and the motherboard overheats when playing games. I hate summertime. :(


---------------------------------------------------------------------------------

@Luminair: My P182 is still stock. Isn't intaking fresh air from the card slots bad? Won't it disrupt the air flow inside the case?


----------------------------------------------------------------------------------

Water flow is something I understand, but water cannot be compressed. Air, however, is compressible.

I've been thinking about this a bit. Wouldn't it be better to have more intake than exhaust?

If all the fans are at the exhaust, then the fans are decompressing the inside of the computer case. If the case is decompressed, then there are fewer particles of air inside the case than if it were at room pressure. It has low air mass.

Air cooling depends on the molecules touching the heatsink, absorbing the excess energy and then throwing them out so that other lower energy molecules can absorb the heat too.

I think running the PC case with positive pressure be better, since there would be more cold molecules per unit of volume traveling through the case, therefore absorbing more heat. Also, if water pumps are more efficient at compression than suction, shouldn't case fans be better at pushing air in than at extracting it?

In a nutshell: more air mass=more cooling power?


----------------------------------------------------------------------------------

@zoatebix: how close together are you talking about? I was wondering if mounting Noctua P12 fans on both ends on the P182's top HDD cage would be a good way of building a relatively silent two stage compressor for the intake.


-----------------------------------------------------------------------------------

I apologize is this post is a bit confusing.

Luminair
Posts: 223
Joined: Wed Mar 21, 2007 10:45 am

Post by Luminair » Mon Jul 21, 2008 4:35 pm

> if water pumps are more efficient at compression than suction, shouldn't case fans be better at pushing air in than at extracting it?

Maybe, but only if the water pump uses the same kind of propeller as the fan? :)


> In a nutshell: more air mass=more cooling power?

Probably, but immeasurably so, I bet.


> @Luminair: My P182 is still stock.

This is still pretty accurate, though my climate might be tens of degrees cooler than yours: viewtopic.php?p=404018#404018
Last edited by Luminair on Mon Jul 21, 2008 4:48 pm, edited 1 time in total.

Vicotnik
*Lifetime Patron*
Posts: 1831
Joined: Thu Feb 13, 2003 6:53 am
Location: Sweden

Post by Vicotnik » Mon Jul 21, 2008 4:39 pm

I usually go for 100% exhaust, no active intakes at all. Then I tape up unwanted intakes to control the intake.

Here's a drawing of the principle of my airflow.

jaganath
Posts: 5085
Joined: Tue Sep 20, 2005 6:55 am
Location: UK

Post by jaganath » Mon Jul 21, 2008 4:45 pm

first of all, there have been a million threads discussing the relative merits of positive vs negative pressure, please do a search. Bluefront in particular is the champion of positive pressure, but for dust control rather than the reasons you give here.
I think running the PC case with positive pressure be better, since there would be more cold molecules per unit of volume traveling through the case, therefore absorbing more heat.
whether positive or negative pressure, you are still moving the same amount of air through the case, so theoretically same cooling performance. although most of the heat-producing parts of a PC are near the rear exhaust fans (ATX specification designed it that way) and generally the further a part is away from a fan the less cooling it gets, so if all the fans are on the intake cooling will probably be worse.
if water pumps are more efficient at compression than suction, shouldn't case fans be better at pushing air in than at extracting it?
in general, axial fans are quite poor at working against backpressure (ie compression). centrifugal and cross-flow fans are much better in this respect (but also much noisier).

danielG
Posts: 68
Joined: Thu Jul 17, 2008 3:40 pm
Location: Leiria, Portugal

Post by danielG » Tue Jul 29, 2008 8:00 am

Thank you all for answering my questions. I've learned a lot.


I finally got round to experimenting a bit and this is what I found:

1) only Tricool top+rear fans on low (1200 rpm) -- insuficient cooling. CPU and MB temps rise slowly at idle. Cannot handle full load.
2) Tricool top+rear fans on low + 1200 rpm Coolermaster middle fan -- temps stable at idle, barely enough at full load. A bit louder than 1), but the middle fan isn't great.
3) Tricool top+rear fans on medium (~1600 rpm) + 1200 rpm Coolermaster middle fan -- much better than 2), stable temps at full load. Loud.
4) only Tricool top+rear fans on medium (~1600 rpm) -- best cooling performance and lowest temps of the 4 configs. Loud, but less than 3).

Using the 'unscientific' method of putting my hand in front of the front intake, I concluded that 4) draws more air than 3). So using a slow intake fan with fast exhaust fans actually restricts air intake.

However, 2) draws more air in than 1). I'm guessing this is because air speed drops off with distance, just like jaganath said.

dhanson865
Posts: 2198
Joined: Thu Feb 10, 2005 11:20 am
Location: TN, USA

Post by dhanson865 » Tue Jul 29, 2008 11:37 am

You mention 1200 RPM for the tricools but http://www.silentpcreview.com/article69 ... ml#tricool shows them doing under 1000 RPM on low.

Sounds like you could use some better fans.

Yate Loon 120mm is about 1300 RPM and has a good noise signature and is cheap.

Scythe Slipstream 1200 RPM moves more air and has a better noise signature. Don't let the RPM numbers fool you on this one. It is a different kind of fan so it moves more CFM per RPM. It is higher in price than the Yate Loon but is reported to be one of the best sounding fans. Don't make the mistake of buying the 1600 RPM slipstream, it really is loud. The 1200 is the right slipstream based on your description of using tricools on medium.

Scythe SFLEX 1200 RPM or 1600RPM is more expensive yet but lasts longer especially in high temp environments. The 1200 RPM won't move as much air as the slipstream 1200 but should sound better than your tricools. I doubt you want the 1600 RPM version but unlike the 1600 RPM slipstream the SFlex 1600 can be slowed down to a reasonable speed by any number of fan control methods.

http://www.silentpcreview.com/article255-page8.html shows blocking off some of the area around the PSU. Recommended if you aren't putting tons of hard drives in that lower chamber and your PSU has a fan.

If it were me I'd try without a fan in the top position (blocking off that top vent) and see how that goes with your existing fans. If you have 5 fans you have plenty of configurations to try but only the top vent fan is worth blocking off. http://www.silentpcreview.com/article255-page6.html shows you some that have been tried.

Then look at the price of Yate Loon vs Slipstream shipped to your part of the world and go with one of those to replace a few of your fans. And retest after you have more fans in hand.

The P18x cases are not fire and forget. The CPU, CPU heatsink, PSU, video card, and hard drives all effect your choice of quantity and placement of fans.

seemingly.random
Posts: 176
Joined: Mon Oct 16, 2006 11:51 am
Location: Southeast, USA

Post by seemingly.random » Tue Jul 29, 2008 6:22 pm

This article reveals that the antec fans don't spin at 1200,1600,2000 rpm (low,medium,high) as advertised but at 870,1430,1930.

The 1080 rpm nexus might be just what you need and are quieter. They're available with open corners now. The 1200 rpm scythe slipstream might work also. I refer to the Fans & Controls reviews often as it appears others also do - there are over a million views combined.

In the end, no amount of reading or advice is going to do it for you completely. You're just going to have try a few combinations like you've already started to do.

HAHA
Posts: 20
Joined: Tue Apr 26, 2005 4:13 am

Post by HAHA » Tue Jul 29, 2008 11:55 pm

It is probably not necessary to have more than one case fan in P182.
I have a workstation here built around P182 containing: Intel E6850 (3 GHz) on a passive heatsink, NVidia 8600GT (passive), one Raptor 74 GB and two Samsung 500 GB harddisks. Motherboard is Asus P5K-E WiFi. The PSU is a Corsair 520 W, which is very silent.

I ripped out all three Tricool fans because they were far too noisy and couldn't be controlled. There is now just one Nexus 120 mm fan cooling the entire machine and it runs on about 800 rpm when the machine is in 2D duty. However, it took some modifications to guide the air properly.
What I did was the following:

I sealed all holes in the chassi (including the top exhaust) with tape, except for the rear exhaust and the front intakes. I also cut off the rear exhaust grill to reduce drag and noise. I sealed the upper and lower chamber from each other. I sealed all rear intakes around the PSU so that it sucks air from the front only. I put the Raptor in the lower chamber and it is now sufficiently cooled by the PSU airflow.

The GPU is passively cooled and it sucks in air through its slot grill by negative pressure. The upper harddisks and the passive CPU are also cooled generously by the horizontal airflow. The machine is now very quiet and cool.

Yesterday, I built another machine with almost the same formula but in an Antec Solo chassi. (I wanted to try the suspended harddisk mounting.)
It seems to work almost as well in this chassi. All it takes is some tape, careful measurements and time.
The secret is really the taping, which makes it possible to guide the air exactly where you want it. You get rid of a lot of leaks and losses and don't have to use as much fan capacity to keep the pressure and flow.

alecmg
Posts: 204
Joined: Thu Mar 13, 2008 5:56 am
Location: Estonia

Post by alecmg » Wed Jul 30, 2008 12:40 am

Arctic Silentium I have operates only with exhaust fans. Intakes are passive and placed strategically near hot components. So far it works wonders and is really quiet. You can feel the cold air flow near vents.
So negative pressure works, but the Silentium has pretty unusual air flow inside, so for you P182 junkies it may not be optimal :)

dhanson865
Posts: 2198
Joined: Thu Feb 10, 2005 11:20 am
Location: TN, USA

Post by dhanson865 » Wed Jul 30, 2008 5:11 am

HAHA wrote:It is probably not necessary to have more than one case fan in P182.
You can't assume. The original poster is in Portugal. We don't know if he has air conditioning. We don't know his ambient temps. We don't know what hardware is inside the case.

While 1 case fan might be the best case scenario there definitely will be scenarios where more fans are necessary.

In a smaller case 3 case fans + 1 fan on the heatsink were used for a PC going to Thailand. http://www.silentpcreview.com/article293-page6.html and the GPU and PSU had their own fans as well.

There is no magic in the P182 that would make it any cooler in an unconditioned environment that any other case.

danielG
Posts: 68
Joined: Thu Jul 17, 2008 3:40 pm
Location: Leiria, Portugal

Post by danielG » Wed Jul 30, 2008 10:12 am

It does get really hot here some days and I don't have air conditioning.

However, I just found out that my Asus M3A-78 EMH is a bad motherboard. I suspected, but wasn't sure. It has a poorly written BIOS and the southbridge heatsink is too small, so it overheats and causes problems. Therefore, the first step to silence my PC is replacing my current motherboard with something better. :( Unfortunately, that is something I can't afford for the time being.

I've come to realize that Tricools are not good, but it's all I have right now. Buying the P182 broke the bank and I have to wait a week or two before buying new fans. I mean to replace the Tricools with Scythe fans. I am considering 800 rpm Slipstreams, since the 1200 rpm are too loud. None of the local shops sell Nexus, but those 1000 rpm fans look really nice.

I've disconnected the top fan and covered ithe opening with a book. With the Tricool on low, the PC is much quieter. Can't tell if this is because the top panel is no longer vibrating or because the top fan throws noise up. Probably both. Both MB and CPU are 2ºC hotter at idle, though. Only now I've realized why one has to seal the top vent if the top fan is off. With the top open, the rear fan draws air from the top instead of from inside the case!

Noticed something strange. With the rear Tricool on medium, the fan is louder with the top sealed than it is with the top open. A bassier, rushing sound. Odd, no?

Is it easier the keep the temps under control with a high end motherboard, with all those heatsinks and heatpipes? Or does it just give off more heat?

seemingly.random
Posts: 176
Joined: Mon Oct 16, 2006 11:51 am
Location: Southeast, USA

Post by seemingly.random » Wed Jul 30, 2008 11:53 am

danielG wrote:It does get really hot here some days and I don't have air conditioning.
Amazing how much difference there is between summer and winter. Last winter, I built a pc with a low end amd/igp in a case that didn't come with fans and I wanted to see how hot it really was before choosing. It ran acceptably for a couple of months - until outside temps got warmer - before I had to install some fans.
However, I just found out that my Asus M3A-78 EMH is a bad motherboard. I suspected, but wasn't sure. It has a poorly written BIOS and the southbridge heatsink is too small, so it overheats and causes problems. Therefore, the first step to silence my PC is replacing my current motherboard with something better. :( Unfortunately, that is something I can't afford for the time being.
Don't know if you're going to find a 780g mb that is much cooler. I have the gigabyte version and the sb chip is hot also. It might just require some creative case cooling.
I know what you mean about the bios. It seems that mb manufacturers release the alpha/beta versions and let the public test.
I've come to realize that Tricools are not good, but it's all I have right now.
Maybe you could use a tricool positioned on the inside end of the middle drive bay, to cool the gpu and mb. Also, for testing, I'd remove the gpu and use the 780g igp to check on the gpu as the heat culprit.

Your sig indicates a pwm tricool. I didn't know that antec made any 4 wire pwm fans.

You might be able to locate a cheaper yate-loon d12-sm fan. At 5 volt, it spins at 880 rpm. There are people here that can help with setting it up for 5 volt if needed.

HAHA
Posts: 20
Joined: Tue Apr 26, 2005 4:13 am

Post by HAHA » Thu Jul 31, 2008 1:49 am

The ambient temperature in the lab where I have the P182 and the Antec Solo workstations is between 25 - 30 Celsius at the moment.
While I wouldn't go to say the P182 is magic, I must say I haven't tried anything that is as efficient. It turns out the new Antec Solo build can't keep quite the same low temperatures in spite of a cooler CPU. I notice the differences mainly with the GPU that stays about 8 degrees higher in the Solo. Could be some sample variance accounting for some difference though.

I still have plenty of spare cooling capacity in both machines because the Nexus fan is operating at only 880 rpm (9V I think).

No matter what benchmark or burn-in software I try to run on these machines, I never get anywhere close to high temperatures. Even if I would raise ambient to 35 Celsius, I would still be safe without even raising the rpm of the single exhaust fan.

It's all in the taping though...
I didn't get anywhere near these results before taking care of the air leaks.

kazaroth
Posts: 6
Joined: Tue Aug 05, 2008 4:24 am
Location: London

Post by kazaroth » Tue Aug 05, 2008 4:36 am

HAHA wrote:The ambient temperature in the lab where I have the P182 and the Antec Solo workstations is between 25 - 30 Celsius at the moment.
While I wouldn't go to say the P182 is magic, I must say I haven't tried anything that is as efficient. It turns out the new Antec Solo build can't keep quite the same low temperatures in spite of a cooler CPU. I notice the differences mainly with the GPU that stays about 8 degrees higher in the Solo. Could be some sample variance accounting for some difference though.

I still have plenty of spare cooling capacity in both machines because the Nexus fan is operating at only 880 rpm (9V I think).

No matter what benchmark or burn-in software I try to run on these machines, I never get anywhere close to high temperatures. Even if I would raise ambient to 35 Celsius, I would still be safe without even raising the rpm of the single exhaust fan.

It's all in the taping though...
I didn't get anywhere near these results before taking care of the air leaks.
Do you have any pics of the taping (or pics/detailed description in another thread?)?

I've just bought a P182 and plan on using it with a Thermalright HR-01 and Accelero Rev. 2. I have 4 Nexus 120mm 1000 RPMs plus some bits and bobs, though happy to buy other bits if needed.

(I also have an Enermax Modu82+ and the system will be a Q9450 + HD4870)

I realise this is the wrong thread - but you posted here about such remarkable results!

Alternatively, what's the very best thread about P182 configurations? :)

nwrigley
Posts: 18
Joined: Wed Aug 22, 2007 7:30 pm

Post by nwrigley » Thu Aug 21, 2008 8:01 pm

kazaroth wrote: Do you have any pics of the taping (or pics/detailed description in another thread?)?
I would also like a description or pics of how to tape an Antec Solo case.

frenchie
Friend of SPCR
Posts: 1346
Joined: Tue Jul 08, 2008 4:53 am
Location: CT

Post by frenchie » Fri Aug 22, 2008 12:37 am

Hi,
I'm interested in that too !!
So far, I have the back bottom holes and the empty 5.25 bays blocked off... I'm wondering if I missed any other major spots.

Back on topic :
Best way to know the answer is to do your own tests :) Airflow patterns differ so much from one setup to the other that it's hard to tell which works better...

Post Reply