Does "Evacuate the heat" role for the PSU work?

PSUs: The source of DC power for all components in the PC & often a big noise source.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee, Devonavar

Posts: 46
Joined: Sat Feb 28, 2004 1:11 am
Location: New York, New York

Post by Wintermute » Tue Mar 02, 2004 12:34 am

What's the best practical way to use a 120mm PSU in a hot system to maximize cooling without allowing the fan to speed up beyond minimal?
Not sure about this: Instead of running the 120mm fan off of the PSU's fan controller, attach it to the mobo's fan controller. That way the speed of the 120mm fan is dependant upon the internal case temperature. Basically, the PSU's 120mm fan becomes the exhaust fan. This has the benefit of allowing you to run the fan at a higher voltage than a traditional 120mm exhaust fan because it is inside the case (i.e. less noise) as opposed to being attached to the back of the case. What about the problem of cooling the PSU with warm air from inside the case? Simply open a hole on the front of the PSU (where the wires are coming out from). Attach a 80mm fan over the hole and connect it to the PSU's internal fan connector (i.e. the one the 120mm fan was initially connected to). Run a duct from this fan to the front of the case (through the 5 1/2 bay) or to a top mounted blowhole (which then becomes an intake hole). I believe the front bay duct has been covered a few time in the forums. The top intakehole idea may work a little better as there are cases that alright employ this idea for exhaust. Modding a case for this would also be fairly easy. The 80mm fan would draw in the colder outside ambient air to cool the PSU whenever the PSU complained about being too hot. I believe the Nexus NX-3500 would be ideal for a mod like this.

This setup wouldn't be any louder than a 120mm exhaust fan/80mm PSU fan setup because both fans are controlled by the same fan controllers in both setups. The ability of running the 120mm fan at a higher speed because it is internally mounted I believe would compensate for the additional work required to push air past the PSU internal components vs. blowing directly out of the case. As well, the 80mm fan is now on the inside too, pushing cold air over the PSU vs expelling warm air out. As a result, this fan could also be runner a little higher without the expense of increased noise.

The biggest issue in this setup would be the "collision" between the air that the 80mm fan was blowing over the PSU with the air being blown upwards by the 120mm fan. If there is anyone here who knows something about fluid mechanics then perhaps they could shed some light on this.

Posts: 25
Joined: Wed Feb 25, 2004 7:59 am

Post by strogian » Tue Mar 02, 2004 1:49 pm

If there is anyone here who knows something about fluid mechanics then perhaps they could shed some light on this.
fluid mechanics! that's what it's called!

i've been trying to think of that so long

Posts: 38
Joined: Wed Nov 19, 2003 5:53 pm
Location: Nor Cal

Post by acompeau » Fri Apr 09, 2004 9:36 pm

Wintermute wrote:If there is anyone here who knows something about fluid mechanics then perhaps they could shed some light on this.
Well, I know something about fluid mechanics, but it's not going to shed a ton of light on this. Adding that second fan shouldn't pose a big problem as long as the exhaust path is sufficiently open. If it's not, then those two fans might be working against significant back pressure. Most low-flow (i.e., quiet) fans do very poorly indeed with any significant back pressure. That's why we all bother to cut out grills, or use wire grills, etc.. A more likely problem is that the particular power supply you try this with "blocks" air flow where you put that 80mm fan. Considering the designers do pay some attention to air flow, they may very well have used that part of the PSU to mount components, etc., since they didn't expect any air flow there.

So while I think the idea could work ok conceptually, the devil is very likely in the details.



Ice Czar
Posts: 20
Joined: Tue Nov 16, 2004 9:22 pm
Location: [H]igh Altitude Lab

Re: Does "Evacuate the heat" role for the PSU work

Post by Ice Czar » Sat Nov 27, 2004 12:24 pm

MikeC wrote:The whole dual fan PSU idea seems to have been promoted by Enermax & AMD around the time the original Athlon appeared. [Some of may recall how the Athlon pushed the then-standard 250W PSUs (especially no-name generic ones) to overload, causing all kinds of mysterious instability issues that were eventually traced to inadequate PSUs.]
Athlon Thermal, Mechanical and Chassis Cooling Design Guide

went out and bought an Dual Fan Enermax to conform to that
had a perfectly good PP403X at the time too :p

Patron of SPCR
Posts: 634
Joined: Wed Mar 10, 2004 5:01 pm
Location: Saginaw, Michigan

Post by Sizzle » Mon Dec 27, 2004 5:30 pm

I've been doing some digging here while searching for a new PSU for my Antec 3000b. This thread I think may have the answer I am looking for.

Two people mentioned the possible scenario I am worried about, but none of our experts here gave a direct answer.

If you have a 120mm exhaust fan (such as in a Antec SLK case), does the addition of a 120mm bottom mount PSU create more noise or disrupt system airflow?

I would think that surely a 120mm case fan is fan enough to cool the system. By that thought, a 80mm rear mount PSU would only have to cool itself, and possible the optical drives during use.

Just to throw it in there because I am thinking about a passive PSU as well, in a SLK type case, with a passive PSU, would a front intake 120mm(or any other size I guess) fan need to be added to cool the system with the rear 120mm exhaust fan?

This was a great thread btw, the more time I spend at SPRC, the more anal I am about my pc :lol:

Posts: 1076
Joined: Sun May 25, 2003 8:29 am
Location: USA

Post by DryFire » Mon Dec 27, 2004 6:58 pm

Maybe a PSU like this one: ... 002&DEPA=0

something like this seems to be a very good idea. You can fit a large slow moving blower (15cm) in thier and still have air taken in from the back. This also leaves extra space to put more in the rear of the psu (see pictures)

Posts: 296
Joined: Fri Dec 03, 2004 12:38 pm
Location: France, Europe Folding for SPCR

Post by JanW » Wed Mar 02, 2005 7:10 pm

MikeC wrote:What's the best practical way to use a PSU with a CPU-facing (bottom) intake fan (such as a 2-fan Enermax, Antec, etc or 120mm fan Seasonic, Fortron, Nexus, etc) in a hot system to maximize cooling while keeping the noise level as low as with the best setup using a quiet single 80mm rear fan PSU?
I see three possible approaches, all three aiming to prevent the PSU from coming after the CPU in the thermal chain:
  1. Reverse the direction of the 120mm fan, blowing (more or less) fresh air onto the CPU. As for the rest of the case: put a second intake fan close to the HD's (Bluefront needs a fan on the HD's, too), open up some exhaust somewhere (without fan) and run positive pressure. This must have been suggested before!?! ...but I missed it? Oh wait, you specified it must be an intake fan for the PSU. Why?
  2. Re-locate the PSU further into the case, so it's not above the CPU (see my silent rig). This is the opposite of what motivated Bluefront to his PSU relocation, and hopefully decreases the fraction of CPU heat that escapes the rear case fan.
  3. Put the PSU inside the case. Duct the PSU exhaust to the CPU HS intake, and duct the HS exhaust to promptly exit the case. That way, the PSU helps exhausting air, but doesn't get hot. Practical? You be the judge.
My silent rig is pretty low power (by todays standards anyway). Still, case exhaust temperature when running fans at a rated flow of 37CFM total (the real flow is probably much lower) is 18-20C above ambient. Dropping flow further by reducing exhaust to only the slow spinning PSU fan (Nexus@7V, 22CFM theotical flow, still audible) would increase that temperature proportionally. This would be too much for comfort if the PSU were the last element in the thermal chain. Same goes for higher system power. Either there must be other ways for heat to leave the system (that's the "take the PSU out of the thermal chain"-class of suggestions, see #2 above), or air flow must be higher (so now it gets down to how silent we aim it to be. I'm currently limited by wind noise through the SuperTornado heatsinks and exhaust grill).
I don't have a clue how Bluefronts setup can work. The energy needs to go somewhere. If the PSU is the only way for heat to leave the system, exhaust (and thus PSU) temperature rise above ambient will be simply proportional to total power draw and inversely proportional to airflow. If we assume the same low airflow I'm talking about (the goal is lowest noise) the only constant left to play with is intake air humidity to increase its heat capacity... So if I were to go that way in my rig, either it wouldn't be silent, or the PSU would be frying.

Posts: 4
Joined: Mon Sep 05, 2005 1:02 am

Post by hsuchiaching » Mon Sep 05, 2005 3:29 am

Hello guys,
I am a newbie in this forum and in the field of silencing computer. This thread caught my attention since the result of discussion will help quieting my rig. However, I am quite surprised that there seems to be no definate conclusion after one and half year. Anyway, here's my rig.

2 x Xp-m 2800 in Gigabyte MPX with 2x Zalman 7000Alcu at lowest speed, temp at 50~60 (yes, I did fit them in even tho the fins kinda intervening with each other)-----quiet

4x Maxtor 160G PATA, 2 optical drives, and floppy.-----quiet

Radeon 9250 with passive cooling+ audigy----quiet

Along with antec truepower 550, which is the loudest thing in my rig, are set in 3700bqe with 2 x 120mm (one in, one out).

While the rear stock antec 120mm exhaust running at 12v (still quiet) and the front maddog 120mm at 5v (quiet as well), the 9mm fan at the button of antec 550 produced high frequency noise at full load consider other components are farily quiet. When I covered up the 9mm fan, the noise dissappeared but the rear 8mm fan still rams up high.

Alright, here's my questions related to this thread.

1. Is it necessary to have a extra fan (either 9mm or 120mm) at the button of PSU to serve to purpose of sucking out hot airs from CPU(s)or cool psu?


purpose 1-take away CPU heat

a. 9mm fan sucks away cpus heat? True?
b. If true, how significantly effective compared to the rear 120mm fan while both competing for airflow?
c. If 9mm fan is only marginally effective (I meant rear 120mm fan actually did the most work), then it shall be took off and the lost airflow from 9mm could be compensated by 120mm from the orginal competition airflow lost? right?

purpose 2- cool psu

a. If 9mm is effective taking away heat from CPU, on other hand, those heat were brought into psu at a right angle airflow, does it really cool psu at full load?
b. If so, how significantly effective compared the rear psu 8mm fan?
c. If the above question were all negative (hopefully someone can answer my qestions), then we could conclude that there's no need of such extra fan at the button of psu? And yet, any suggestion to impove airflow or decrease noise?

*Lifetime Patron*
Posts: 5316
Joined: Sat Jan 18, 2003 2:19 pm
Location: St Louis (county) Missouri USA

Post by Bluefront » Mon Sep 05, 2005 3:48 am

You are right....there is no definative conclusion to these questions. I've had much success using a single bottom 120mm PSU to cool the whole computer. But I don't think such a setup can work quietly without modification to the std ATX PSU form factor.

When you have a rear 120mm fan and a bottom 120mm fan in the PSU, both sucking air from the same space....the end result is not optimal from a sound stand-point. But many people use that setup successfully.....

It's an area wide open for new ideas..... :lol:

Welcome to SPCR.

Posts: 647
Joined: Sun Apr 17, 2005 6:09 am
Location: UK

Post by vertigo » Mon Sep 05, 2005 10:30 am

My remedy is to stop hot cpu air from going into the power supply. I use a conventional heatsink (XP/90) with a sucking fan. That keeps the hot air concentrated. Then I use a partition along the top edge of the cpu, which extends to the back fan and the side panel. The cpu air pretty much can't reach the power supply, it goes out the back fan.

To reduce the negative pressure I leave a PCI slot open, this also brings some air to the video card. The only other intake is in the front.

Posts: 187
Joined: Wed Jul 27, 2005 3:16 pm
Location: Travelling worldwide

To what extend is this a desrable goal ?

Post by jjr » Mon Sep 05, 2005 6:25 pm

Maybe it's a silly question, but to what extend do we really want to prevent the air from going into the PSU ?

I mean : most PSUs with an intake fan are quite "closed" otherwise : few grills. I guess what they are meant for is tunelling. Makes sense to me.

Now if we prevent the PSU from getting air from inside the case, with another fan starving the intake fan (be it exhaust or CPU fan ducted to the outside) : where will it get its air from ?
The PSU will have no or little air, warm up, and ramp up even more until it can eventually overcome the compeeting fan.

For me it would always make sense to have a intake dusted CPU fan to keep a positive pressure with air as cold as possible. Then exhaust through the back AND the PSU.

On the other hand, if you have a PSU with grills, then you can duct out as the PSU will take it's air from elsewhere.

At the end, I guess the question is whether one should have a PSU with intake fan at all (vs. open grills). Or as suggested before in the thread you manage to put it somewhere else or have it get air from an area away from the CPU.
Ducted from the outside for instance but that's going to be 2 ducts in 1 case ...

Posts: 187
Joined: Wed Jul 27, 2005 3:16 pm
Location: Travelling worldwide

Post by jjr » Mon Sep 05, 2005 6:34 pm

NB : or alternatively put the intake somewhere else.
Maybe on the front like Tagan does.

Friend of SPCR
Posts: 160
Joined: Fri Mar 03, 2006 11:50 am
Location: Gothenburg, Sweden

Post by kentc » Tue Mar 07, 2006 12:46 pm

thread been silent (no pun...) for a while but i couldn't help but think about this thread when i read a seasonic product catalog:


Posts: 376
Joined: Mon May 15, 2006 12:24 pm
Location: CA, USA

Post by toronado455 » Mon May 15, 2006 12:40 pm

I would want a PSU *with* a 12cm because it is more quiet and efficient - as long as that is true for the application.

But looking at the close position of items in the case and the location
of the 12cm fan at the bottom of the PSU, I am concerned that it might
"compete" with the case exhaust fan which is at a right angle to the
bottom of the PSU. I'm also concerned that the CPU HSF will at least
partially obstruct airflow to the bottom of the PSU.

If OTOH, there was sufficient space and a sufficient gap between these
items and the PSU, then the 12cm fan starts to look like a great idea.
I would definitely feel comfortable about using a PSU with 12cm fan in
a case like this one where there would be a fairly large gap between
the bottom of the PSU and the other items: ... 6811125471

But in a case like this one... ... 6811125486

...I would be concerned about the close proximity issues.

Any opinions?

Posts: 1066
Joined: Sun Jan 16, 2005 2:59 am
Location: Santa Cruz

Post by ronrem » Sat Mar 03, 2007 9:39 pm

There's an undervoltedSempron 2800 you could cool easy,often these hardly warm a full size HS,so a PSU fan can suck that off. A PSU EXPECTS incoming air at over 30C,and some of the more silence-friendly chips have a typical core temp that's below that in operation.

I've been sketching a plan for a inlet duct/HD enclosure. This would be a wood box about the size if a 1/2 g milk carton able to house 2 or 3 HDs. The whole housing would be suspended with,at the case floor,a rubber/plastic collar to seal it to the whole in the floor. Under the floor is 3" clearance,a chamber open only to the rear. Near the top of this duct box is a 140 mm Aerocool that blows toward the rear,aimed midway between the chipset and a Ninja (passive) Just above the Ninja is a PSU with a 140 mm set-speed fan.
This could be an Xclio,a Coolmax,an nMedia... A 120 fanned PSU would suffice. It would be pretty close to a Ninja so most of its draw would be pulling between the HS's fins.

On the back I'd have a low rpm 120,perhaps a Scythe,something that at max is 800 rpm or less. This plugs in as CPU fan so Smart Fan/Q Fan/? adjusts speed based on cpu heat. Ideally it's off a lot and seldom over 500 rpm. Most 140's in PSU's are Yate loon's with a setting at 1000 rpm,a higher setting and Auto. The Aerocools run at about 1000rpm,could be dialed down.

While this has 2 140 mm fans all togather,both are mounted internally-so it's a very indirect soundpath. I'd expect this to handle a 3600 Brisbane easy. I doubt the fans or HD's would be very audible. This box-duct would be oversize on most cases-though I have an old jumbo server box that could do this easy. More likely I'd build a mostly wood case to specs.

A low power Sempron with a mild Vid card,not a lot of HD's, I'd feel could be vented with the PSU,though I'd tend to add a low RPM case fan for "worst Case" moments. Having a one fan rig,like Bluefront is do-able and ideal for most uses.

Mr Spocko
Posts: 229
Joined: Sun Jul 10, 2011 1:56 pm
Location: UK/Eire

Re: Does "Evacuate the heat" role for the PSU work?

Post by Mr Spocko » Tue Aug 09, 2011 3:55 pm

I support the thinking that a PSU should assist in removing heat from a case. Placing it at the bottom is of no use in this regards.
The original post is quite old however fast forward to 2011 and my view remains the same despite the increase in TDP of processors PSU's are "designed to work" and happily so in removing hot air from the case.

I see no evidence of PSU problems regarding top mounting nor do I note any significant noise issues on modern decent quality PSU's which seem more than able to cope with modern case temps.
Put them at the bottom and all the work on a typical case goes to the back fan which could mean an additional fan or a faster one required which in turn will add "more noise" ;-)

Posts: 50
Joined: Wed Sep 29, 2010 8:37 am
Location: Central Europe

Re: Does "Evacuate the heat" role for the PSU work?

Post by Tetreb » Wed Jan 04, 2012 2:57 pm

I put a Corsair AX750 to suck air out from the case (PSU placed at the bottom). The GTX480 above it seemed to heat it up because after short time at high load the PSU fan would spin up rather high. I turned the PSU upside down and with its own air intake it stays silent, other temps seem unaffected. There are two large intake fans at the front too.

Posts: 10
Joined: Sun Aug 28, 2011 5:20 am

Re: Does "Evacuate the heat" role for the PSU work?

Post by Zustiur » Mon Jan 30, 2012 12:40 am

Ultimately it's a question of the cooling efficiency of the PSU.
You can't really make judgements on the best solution without knowing how much cooling/airflow the PSU actually needs.
I'm not clue-y enough about the terms and numbers here to make a qualified statement, but here's what I'm thinking:

(All of these numbers are just made up, I'm not using real world figures)
If the PSU (with no fan at all) plateaus at 60c, but with the fan running plateaus at room temperature (or slightly above), then the question is; how much air flow is required to achieve this? I don't think you can just boost the fan to the point where the PSU ends up being below room temperature... So you only need to work out how much air flow is required to keep the PSU at room temperature.
If it takes 500 RPM from your fan to achieve this, but you can leave it at 800 RPM without hearing it, then you want it to be drawing roughly 300 RPM worth of heat from inside the case.

In other words, if the fan providing more cooling than the PSU requires while running at your preferred RPM, then you should also draw some air from the case.

I'm not sure I'm explaining myself well, so look at it like a program:

Code: Select all

IF FanRPM is Inaudible Then
    IF PSU at RoomTemperature Then
        Draw X air from Case
        Do not draw air from case
    END IF
    Bugger. Buy a quieter PSU/fan.

Posts: 238
Joined: Sat May 11, 2013 8:33 am

Re: Does "Evacuate the heat" role for the PSU work?

Post by QUIET! » Sat May 11, 2013 9:30 am

The heat in the PSU issue really died when 80+ PSUs became widely available.

Unless you've got dual GPUs, lots of high rpm HDDs and a really hot CPU, the internal case temp isn't going to be very high as long as your air flow is decent.

The place where things can go wrong is in the air flow.

PSUs are almost always setup with exhaust fans, they blow air out of the case.

Hot air rises so your exhaust should always be at the top.

If you put a powerful exhaust fan near your PSU fan, that can reduce the amount of air that your PSU fan can flow, that is not good. That's not so much of a problem with the silent PC crowd but be aware.

If you add an intake fan to your case, that is going to make more air flow out through your power supply, that is good.

Once again, hot air rises so to avoid recirculating hot air, keep your intake well below your exhaust and probably on the opposite side of your case to avoid dead zones with no air flow.

Your case might look like swiss cheese with fan grills and vents and most of that is 100% useless or worse.

Blocking off unnecessary vents can help in many cases.

The end result of setting up a PC that way is that I took a $30 cheapo case, swapped the exhaust fan to intake position, blocked off some vents and now my PC is nearly quiet (the case fan isn't the greatest) and coming from the exhaust of the PSU is a strong breeze cooler than a warm fart. My system temps are excellent too.

To be fair my computer does not generate a lot of heat and the 450 watt 80+ gold power supply is gross overkill but my next project will have a ~250 watt 80+ silver PSU with a HD7790 GPU (but only an ssd instead of 3 HDD, an ssd and a bdrw). If that is successfull I've got my eye on a 240 watt 80+ gold small form factor PSU and a custom case.

I guess the point is that good PSUs run cool and if you have decent air flow, the computer can't put enough heat in the air to warm it very much so your PSU probably will not be throttling its fans up and down making excess noise unless its responding to load.

Post Reply