Possible to determine PC fan needs without experimentation?

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

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
El Doug
Posts: 103
Joined: Sun Jan 01, 2006 6:32 am

Possible to determine PC fan needs without experimentation?

Post by El Doug » Sun Jan 01, 2006 12:48 pm

Airflow is a large component in the cooling, but also noise, of a PC.

To minimize this source of sound, lower-speed fans both in the case and PSU can be utilized. Is there any way to ball-park how much flow Ill need based upon computing components within the build?

Id rather not have to spend 3 days experimenting with various setups and fan speeds while performing temp and stability tests... please tell me there is a quick answer?

teknerd
Posts: 378
Joined: Sat Nov 13, 2004 5:33 pm

Post by teknerd » Sun Jan 01, 2006 12:58 pm

Well before most people here can really help you, we need to know the components that will be going into this system.
It is definately possible to ballpark (and even almost pinpoint) the airflow you'll need for your system (especially if you are using compenets that other peopl here are using).
So anyway, post the components and then we can get to work :-)
[img]http://tinyurl.com/rqlff[/img]
[size=75]Main Rig: X2-4200+ | XP-120 | A8N-SLI Premium | 2GB RAM | XFX 6600GT w/VF700 | 74GB Raptor & 320GB 3200JD | P180 | SmartPower II 500W | 2x Dell 1704FPV (Digital)
Media Center: Athlon XP-M 2500+ @ 2.2GHz | ALX800 w/92mm Panaflo | 1GB RAM | 40GB WD | SLK3000B | Modulated, Fan-Swapped 380S PSU
[/size]

El Doug
Posts: 103
Joined: Sun Jan 01, 2006 6:32 am

Post by El Doug » Sun Jan 01, 2006 2:09 pm

X2 4400+
7700cu w/ fanmate
7800gtx 256mb w/ AC Silencer
A8N-SLI Premium (hence passive northbridge)
2x 200gig SATA2 Samsung


this is a horizontal desktop case, but it is full atx and has more than enough headroom. the case has mounts for 2x80mm fans rear, and 1x80mm fan BELOW the HDD cage

teknerd
Posts: 378
Joined: Sat Nov 13, 2004 5:33 pm

Post by teknerd » Mon Jan 02, 2006 12:20 am

First, i would recommend changing the CPU cooler. A better cooler (the Scythe Ninja if it will fit or the XP-120 if the case is not tall enough). Equip either one with a Nexus 120mm fan for a great CPU cooling setup.
Second: 80mm Nexus's in all of the case fan mount points should provide more than adequate airflow. Make sure you softmount these fans to reduce vibration.
Third, What case specifically are you using. If possible you might want to suspend the hard drives. Some people here might be able to give you some good modding tips for the case.
Fourth, Pick up a multi channel fan controller as you will want to be able to control the fan speed for the case fans, cpu, and AC silencer (you will definately want to undervolt this).
Fifth (and perhaps most importantly), what kind of budget are you working with here. This will greatly affect some of the advice you get.
Also, are these components already bought are are you considering them?
[img]http://tinyurl.com/rqlff[/img]
[size=75]Main Rig: X2-4200+ | XP-120 | A8N-SLI Premium | 2GB RAM | XFX 6600GT w/VF700 | 74GB Raptor & 320GB 3200JD | P180 | SmartPower II 500W | 2x Dell 1704FPV (Digital)
Media Center: Athlon XP-M 2500+ @ 2.2GHz | ALX800 w/92mm Panaflo | 1GB RAM | 40GB WD | SLK3000B | Modulated, Fan-Swapped 380S PSU
[/size]

El Doug
Posts: 103
Joined: Sun Jan 01, 2006 6:32 am

Post by El Doug » Mon Jan 02, 2006 12:38 am

ive been using this setup in a mid-tower for 6 months - it just sounds like the exhaust of my ninja (bike, not HS)

the case is an x15e. i have been toying with some elastic-mount HDD ideas, but thats not my real concern. cooling is no problem, i just want to minimize noise with the correct proportions of case/psu/vga fans.

the idea behind the AC Silencer is not to super-cool the vga. I simply want to maintain the same temps as stock while running silent with severe undervolting - i hope that i get some minor perks as well with this cooler doubling as a case exhaust.

i see no point in altering my HSF setup - the proc puts out very little heat as it is, and spending $70 to lower by 1C and 1dB seems a bit rediculous. If you truely believe Ill see significant performance by switching the CPU cooler, then by all means Ill go for it. budget is a null-issue

if I even run an intake, it would be on an in-line resistor or something - i would want it to be litterally silent, and doubt that increasing it with any kind of controller would be beneficial.

as for the 2 rear exhausts, is a fan controller necessary? theres no way I can convince my computer to simply rev them up by itself when certain thermal loads are surpassed? i run the CPU cooler with the fanmate, and almost always have it going silent (except in rendering/encoding and gaming situations), so this also seems to indicate to me that a fan controller would be of little value

I really just want to figure out how much airflow would be necessary to keep this computer silent at idle/minimum web-browsing loads. I dont care if it gets loud during heavy use. Elastic HDD mounting, fan gromits, and perhaps even acoustic insulation and a quiet-rated optical drive are all part of my plan, and will be executed - i dont need help with that. My only concern is maximizing efficiency with the fan setup

McBanjo
Posts: 671
Joined: Thu Dec 08, 2005 1:40 pm
Location: Sweden
Contact:

Post by McBanjo » Mon Jan 02, 2006 12:46 am

Buy a new chassi with 120mm fans ;-)

El Doug
Posts: 103
Joined: Sun Jan 01, 2006 6:32 am

Post by El Doug » Mon Jan 02, 2006 12:49 am

not happening :P

im using this rig as is - just want some help with new PSU choice and hence what fans to use at what power

dukla2000
*Lifetime Patron*
Posts: 1465
Joined: Sun Mar 09, 2003 12:27 pm
Location: Reading.England.EU

Re: Possible to determine PC fan needs without experimentati

Post by dukla2000 » Mon Jan 02, 2006 1:45 am

El Doug wrote:Airflow is a large component in the cooling, ... Is there any way to ball-park how much flow Ill need based upon computing components within the build?
Yes. For each component (CPU, PSU, vid card) as well as each compartment (or sub-compartment) in the case the good old law of air cooling holds:
Q = (1.76 * H) / dT
where
Q is airflow in cubic feet/min
H is heat in Watts
dT is temp rise in degree C

Just remember that your actual fan throughput will be a lot less than its theoretical output due to static pressure and/or not-optimum flow paths. Theoretically your 80mm Panaflo @ 9V is 18cfm: for calculation purposes assume it is 9cfm until you can actually experiment with your setup. (It may even be a lot less than 9cfm!)
[size=75]2009/Oct: Jetway JNC81-LF * 4850e naked under fanless Xigmatek Apache * Antec mini Skeleton w/Nexus 120mm PWM fan * Delta 90W brick w/Skeleton DC-DC board * WD2500BEVT 250Gb blue[/size]

El Doug
Posts: 103
Joined: Sun Jan 01, 2006 6:32 am

Post by El Doug » Mon Jan 02, 2006 1:52 am

dT? that sounds like calc. and if i want 0 dT, i seem to be dividing by zero... gah, I hated math. thank the Lord the bio degree only required 2 semesters of the crap

so lets say for example (and im pulling this off the top of my head), I have a total of 80watts of heat being produced. how much flow will i need theoretically?

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

Post by jaganath » Mon Jan 02, 2006 5:08 am

so lets say for example (and im pulling this off the top of my head), I have a total of 80watts of heat being produced. how much flow will i need theoretically?
I don't think you can work this out by math alone. The simplest way is to build it and test it yourself; I think it's an exaggeration to imagine that this will take 3 whole days. To my eyes it doesn't look like a ridiculously hot setup, the X2 4400+ and the 7800GTX appear to be the main sources of heat. Also the amount of flow you will need will depend on ambient temp, airflow paths, etc like dukla said, so to paraphrase Nike, just do it! :lol:

dukla2000
*Lifetime Patron*
Posts: 1465
Joined: Sun Mar 09, 2003 12:27 pm
Location: Reading.England.EU

Post by dukla2000 » Mon Jan 02, 2006 6:22 am

El Doug wrote:dT? that sounds like calc. and if i want 0 dT, i seem to be dividing by zero... gah, I hated math. thank the Lord the bio degree only required 2 semesters of the crap

so lets say for example (and im pulling this off the top of my head), I have a total of 80watts of heat being produced. how much flow will i need theoretically?
Relax :D dT is only "change in temp" and not calculus!

Lets say you are prepared for 15C temp rise (e.g. your room temp is 20C, you are happy for your case to run at 35C as all other components will be happy in an 'ambient' 35C), and inside the case you have that 80W. Then
Q = (1.76 * 80) / 15 = 9.4cfm
So that 80mm Panaflo @ 9V is probably a good place to start your experiments. If your case airflow is good (low static pressure) you may be able to drop the Panaflo below 9V.
Again (in this paper example) remember that the intake air temp to your CPU hsf will now be 35C. So say you are prepared for the CPU to run @ 50C and have an adequate heatsink on it. And that the CPU runs 50W. Then the fan on the CPU hsf should be
Q = (1.76 * 50)/ 15 = 5.9cfm
A fair chance you could get away with a Panaflo @ circa 5V.
etc.
[size=75]2009/Oct: Jetway JNC81-LF * 4850e naked under fanless Xigmatek Apache * Antec mini Skeleton w/Nexus 120mm PWM fan * Delta 90W brick w/Skeleton DC-DC board * WD2500BEVT 250Gb blue[/size]

El Doug
Posts: 103
Joined: Sun Jan 01, 2006 6:32 am

Post by El Doug » Mon Jan 02, 2006 11:10 am

well, thats really somethin! if i listen to my case manufacturer who says that it is "designed for maximum flow," then Ill just assume the flow is aweful. Ill calculate my real heat output and work from there. You think starting with twice the theoretical flow is a good idea? Should I start with more just to be sure?

as for that dT: im SO glad its not rate of change - you saved me a lot of headache :)

dukla2000
*Lifetime Patron*
Posts: 1465
Joined: Sun Mar 09, 2003 12:27 pm
Location: Reading.England.EU

Post by dukla2000 » Mon Jan 02, 2006 3:10 pm

El Doug wrote:You think starting with twice the theoretical flow is a good idea? Should I start with more just to be sure?
As per jaganath you still have to build it and test it. In my experience assuming a fan pushes only about half of its theoretical output is a good starting point - you can always slow it down some more if things are cooler than planned (or still too noisy!) In practice, around SPCR, the tendency would be to let the temps rise a bit in favour of slower/quieter fans.

The main use of these paper calculations is that if, for example, it is taking you 30cfm to keep 80W cool & stable you are probably doing something wrong. So back to the drawing board, add a duct or whatever ...
[size=75]2009/Oct: Jetway JNC81-LF * 4850e naked under fanless Xigmatek Apache * Antec mini Skeleton w/Nexus 120mm PWM fan * Delta 90W brick w/Skeleton DC-DC board * WD2500BEVT 250Gb blue[/size]

El Doug
Posts: 103
Joined: Sun Jan 01, 2006 6:32 am

Post by El Doug » Mon Jan 02, 2006 3:53 pm

well, yes - i just wanted a starting point for fan setup before I begin testing. I was almost afraid that 30cfm would be too little before I found this forum.

Thanks for all your help, much appreciated

Post Reply