Sign in with
Sign up | Sign in
Your question
Closed

Bottom mounted PSU - fan facing up or down, a test

Last response: in Components
Share
February 3, 2010 7:21:22 PM

There seems to be endless debates about whether a bottom mounted PSU should have its fans facing up or down, so I decided to do a test. Here is the air flow pattern in my case:
- one top 120mm exhaust fan
- one rear 120mm exhaust fan
- one side 120mm intake fan (situated where the video card is)
- 120mm CPU fan blowing up

with the PSU fan facing down, the temps are:

CPU idle: 31c, load: 45c
GPU idle: 52c, load 78c

with the PSU fan facing up, the temps are:

CPU idle: 37c, load 45c
GPU idle: 51c, load 77c

So there seems to be no significant temp difference whether the PSU is mounted facing up or down. The video card is actually 1c cooler when it is facing up, but I guess 1c is not significant enough to make one prefer it one way or the other.

The only real big number here is the CPU idle temp, it went from 31c to 37c, but the CPU load temp remains the same. My theory is that PSU fan facing up resulted in increased air flow around the video card but reduced airflow around the CPU, and hence the increase in CPU idle temp. But when CPU is under load, the CPU fan kicked into higher gear which was able to overcome the reduced airflow, and hence the load temp remained the same.
a b ) Power supply
a b U Graphics card
a b à CPUs
February 3, 2010 8:14:59 PM

Don't know about you, but I'd consider a 6C difference between the CPU temps a significant.

At anyrate, it could be that the PSU fan is affecting the CPU airflow in some way in test 2 (ie. vacuum affect).

A few questions:
1. Was the CPU fan always constant (as in with in 1-8 rpm%) between the test?
2. What CPU fan?
3. What PSU?
Score
0
a b ) Power supply
a b U Graphics card
a c 106 à CPUs
February 3, 2010 8:30:38 PM

Should be down to draw fresh cool air into psu which then exhaust out the back.
Score
0
Related resources
a b ) Power supply
a b à CPUs
February 3, 2010 9:31:45 PM

I am actually pretty sure the PSU sucks air in, in this case the lower GPU temperature would make perfect sense because the hot air that is exhuasted out of the bottom GPU grills is sucked out by the PSU, however by doing this the PSU gets no "cool air"...just heated air exhuasted from other components making your PSU hotter in my theory.
Score
0
a b U Graphics card
a b à CPUs
February 3, 2010 9:46:09 PM

I had this dilemma when I put together current rig. Antec 300, antec ew650.
http://www.antec.com/pdf/manuals/300_EN%20manual.pdf with pic
Quote:
Installing the Power Supply
1. With the case upright, place the power supply on the bottom of
the case.
Note: Power supplies with fans on the bottom of the power
supply will need to be mounted so that the fan is facing the top
of the case. Three Hundred provides mounting holes for power
supplies with standard mounting layouts to be installed upside
up or upside down.



I read the manual, for this case, fan up, there is a 1/16' gap under the psu, but no bottom case vents. Must not be enough to let the psu 'breath' ? I read other cases have large vents and even mesh filters to draw air from the bottom, but not all.
Score
0
February 3, 2010 10:12:54 PM

notty22 said:
I had this dilemma when I put together current rig. Antec 300, antec ew650.
http://www.antec.com/pdf/manuals/300_EN%20manual.pdf with pic
Quote:
Installing the Power Supply
1. With the case upright, place the power supply on the bottom of
the case.
Note: Power supplies with fans on the bottom of the power
supply will need to be mounted so that the fan is facing the top
of the case. Three Hundred provides mounting holes for power
supplies with standard mounting layouts to be installed upside
up or upside down.



I read the manual, for this case, fan up, there is a 1/16' gap under the psu, but no bottom case vents. Must not be enough to let the psu 'breath' ? I read other cases have large vents and even mesh filters to draw air from the bottom, but not all.


My NZXT case has PSU at the bottom too, but it has a PSU vent at the bottom and a mesh filter over the vent. My only beef is that the vent is not as big as my PSU's 120mm fan, so some of the fan is covered up. Not sure how big of a deal this is.

Based on my test above, mounting the PSU facing up has little effect on the components, except that CPU idle temp is 6c higher in this case, (that's significant, but CPU load temp remains the same).
PSU may run hotter b/c it has no access to cool air, only hot air exhausted from other components, but so this is no different from any PSU mounted on top of the case. Majority of PSUs are mounted on top, and they do fine, so I don't think mounting the PSU facing up will hurt the PSU itself.
If possible I would still prefer to mount it facing down, but with your Antec 300 it seems impossible without major modding. In this case I would just mount it facing up and not worry about it.
Score
0
February 3, 2010 10:17:55 PM

Shadow703793 said:
Don't know about you, but I'd consider a 6C difference between the CPU temps a significant.

At anyrate, it could be that the PSU fan is affecting the CPU airflow in some way in test 2 (ie. vacuum affect).

A few questions:
1. Was the CPU fan always constant (as in with in 1-8 rpm%) between the test?
2. What CPU fan?
3. What PSU?


yes I think the 6c difference is significant, but even more significant is the fact that load temp remained the same. When heat is a issue it is usually the load temp, not the idle temp.

during the two tests, CPU fan speed was the same when it was idle and when it was under load. I have a Xigmatek S1283, PSU is a OCZ StealthXStream.
Score
0
a c 248 ) Power supply
a b U Graphics card
a b à CPUs
February 4, 2010 1:56:08 AM

yiplong - Which case do you have?

All modern power supplies exhaust air out the rear of the psu case.

I have a Lancool Dragon Lord case for my new system. I have a Coolermaster HAF 932 for my emergency backup system. I've also had other cases with bottom mounted cases. You have two choices with a bottom mounted power supply.

First, if your case has sufficient ventilation, airflow, and cooling, then you could mount the psu fan side down so it draws in cool air from the opening in the bottom of the case and exhausts warm air out the rear of the case. The idea is that by drawing in its own supply of cool air rather than drawing in warm case air the power supply will be easier to cool and will last longer. However, do not place the case on a carpeted floor, especially deep pile carpeting. The carpet fibers will block air flow. A few years ago I made my own pc stand with casters.

Second, you could install the power supply with the psu fan on top. The concept is that the psu would help cool the interior of the case by drawing in warm interior air and exhausting it out the rear of the case. This concept goes back many many years when Intel started making cpu's that ran hot. The cpu's did not have cpu heatsinks or fans. In addition pc cases did not have case fans. In the original atx standards Intel specified that psu's should be placed near the cpu so the psu fan could help cool the cpu. Obviously this standard no longer applies. In modern tower cases we now have mutiple exhaust fans near the cpu.

The choice is yours to make based on your own specific ventilation, airflow, and cooling situation.

The ventilation, airflow, and cooling in my new Dragon Lord case and my older HAF 932 are excellent so I mounted my power supplies with the psu fans on the bottom.
Score
0
a b à CPUs
February 4, 2010 2:15:08 AM

yiplong said:
yes I think the 6c difference is significant, but even more significant is the fact that load temp remained the same. When heat is a issue it is usually the load temp, not the idle temp.

during the two tests, CPU fan speed was the same when it was idle and when it was under load. I have a Xigmatek S1283, PSU is a OCZ StealthXStream.



Ahhh... young and naive... hehe... you forgot something...
Like you said, yes load temps are the issue for you, so... what about your PSU temps?????

Theere you gooo.... see when you had your PSU intake drawing air from your case, it already draws hot air over its components, so it becomes even hotter, reducing efficiency, increasing consumption and reducing its lifespan, along with potential overclockability of the components.

Hence if there is an intake at the bottom of the case it is much better to let your PSU draw colder air in and the rest of the system as you found out is not getting any major differences.

Nice experiment and a good start! Congratulations on your attempt!.. now, on for some more! hehe... ;) 
Score
0
February 4, 2010 2:55:10 AM

darkguset said:
Ahhh... young and naive... hehe... you forgot something...
Like you said, yes load temps are the issue for you, so... what about your PSU temps?????

Theere you gooo.... see when you had your PSU intake drawing air from your case, it already draws hot air over its components, so it becomes even hotter, reducing efficiency, increasing consumption and reducing its lifespan, along with potential overclockability of the components.

Hence if there is an intake at the bottom of the case it is much better to let your PSU draw colder air in and the rest of the system as you found out is not getting any major differences.

Nice experiment and a good start! Congratulations on your attempt!.. now, on for some more! hehe... ;) 


Probably a vast majority of PCs have top mounted PSU drawing in hot air right off the CPU area and they all seem to function and OC just fine. A bottom mounted PSU even fan facing up will draw in far less hot air than a top mounted one b/c hot air rises. Since all the top mounted ones are doing fine, I am sure a bottom one is also fine.
So your idea sounds plausible at first glance, but it is actually nonsense when you think about it. Fan facing up will reduce efficiency, reduce lifespan and limit OC? Those are pretty serious consequences, show me some results and I will believe you.
Score
0
a c 144 ) Power supply
a b U Graphics card
a c 172 à CPUs
February 4, 2010 3:23:03 AM

I have three Antec 900 cases, all with PSU's mounted with the fans facing up. The PSU exhaust air has never been higher than 3 C. over ambient (as measured with an electronic cooking thermometer). One of my systems has 3 hard drives (a WD 640 GB Black and two 1 TB Greens) mounted in the lower drive bay. I have never seen HD temps higher than 29 C.

Did I experiment with mounting the PSU with fan facing downward? No. Did I experiment with cutting a hole in the bottom of the case? No. "Good enough" is frequently good enough.
Score
0
a b à CPUs
February 4, 2010 10:46:39 PM

yiplong said:
Probably a vast majority of PCs have top mounted PSU drawing in hot air right off the CPU area and they all seem to function and OC just fine. A bottom mounted PSU even fan facing up will draw in far less hot air than a top mounted one b/c hot air rises. Since all the top mounted ones are doing fine, I am sure a bottom one is also fine.
So your idea sounds plausible at first glance, but it is actually nonsense when you think about it. Fan facing up will reduce efficiency, reduce lifespan and limit OC? Those are pretty serious consequences, show me some results and I will believe you.


Look, don't take things so dramatically. It always depends on the case you are talking about. If there is ample space between the PSU intake and the graphics card (assuming the PSU is at the bottom and the intake is towards the VGA) then there will not be a big problem. If on the other hand we are talking about a cramped case and a VGA that outputs 150W of heat into the system then yes the PSU will suffer.

Also your definition of "seem to function and OC just fine" is far from my definition of fine. Fine for you could be a PC doing a word processing document and the occasional Solitaire and an overclock of 5% and the occasional BSOD. Fine for me is running a server based application 24/7 with no crashes or memory corruptions and OC is also a typical 30% with again no problems at all.

You correctly pointed out that a bottom mounted PSU even fan facing up will draw in less hot air than a top mounted one b/c hot air rises. But that again depends on the case airflow. If there is no airflow in the case, the air will have the same temperature everywhere, hence it will not matter where your PSU is, as long as the intake is looking inside the case it will draw the same hot air.

Regarding the temperature rise the the PSU efficiency and lifespan it is very long and can't be bothered explaining in a forum. You can simply Google it and find out for yourself. You can contact PSU manufacturers like Corsair, Enermax, etc or even email this guy http://www.jonnyguru.com who knows a thing or two and see what he says.
Score
0
a b à CPUs
February 4, 2010 10:49:25 PM

jsc said:
I have three Antec 900 cases, all with PSU's mounted with the fans facing up. The PSU exhaust air has never been higher than 3 C. over ambient (as measured with an electronic cooking thermometer). One of my systems has 3 hard drives (a WD 640 GB Black and two 1 TB Greens) mounted in the lower drive bay. I have never seen HD temps higher than 29 C.

Did I experiment with mounting the PSU with fan facing downward? No. Did I experiment with cutting a hole in the bottom of the case? No. "Good enough" is frequently good enough.


Can't blame the guy for trying. He is experimenting (we need more people like him). Curiosity did not kill the cat, it improved the man and that is what he is doing. His system probably runs fine as well but since he had some time in his hands why not try some stuff and find out for himself?
Score
0
February 4, 2010 11:29:07 PM

jsc said:
I have three Antec 900 cases, all with PSU's mounted with the fans facing up. The PSU exhaust air has never been higher than 3 C. over ambient (as measured with an electronic cooking thermometer). One of my systems has 3 hard drives (a WD 640 GB Black and two 1 TB Greens) mounted in the lower drive bay. I have never seen HD temps higher than 29 C.

Did I experiment with mounting the PSU with fan facing downward? No. Did I experiment with cutting a hole in the bottom of the case? No. "Good enough" is frequently good enough.



Yeah that's exactly what I was saying, and my little experiment there probably confirmed that to some degree. Mounting the PSU fan facing up or down doesn't seem to make that much of a difference.
I can't believe some folks would, without any actual experience claim that doing it one way or the other will "shorten the components' useful life", or "limit its OC potential" or other crap like that. What a load.
Score
0
a b à CPUs
February 5, 2010 1:22:03 AM

yiplong said:
Yeah that's exactly what I was saying, and my little experiment there probably confirmed that to some degree. Mounting the PSU fan facing up or down doesn't seem to make that much of a difference.
I can't believe some folks would, without any actual experience claim that doing it one way or the other will "shorten the components' useful life", or "limit its OC potential" or other crap like that. What a load.



hahahahahaha
Score
0
a c 248 ) Power supply
a b U Graphics card
a b à CPUs
February 5, 2010 1:47:56 AM

It actually did a very long time ago when power supply fans were the only means of exhausting hot air generated by the cpu. That was Intel's solution to their cpu heat problem. As I pointed out in my original post, that is no longer the situation.

That was long before the Internet. I distinctly remember searching through the advertisements in "Computer Shopper" for pc cases that had some ventilation. I found a full tower case that had small ventilation slots along the bottom of both side panels and the front panel. It also had a spot on the rear panel for either a 40mm or 60mm fan. I felt very lucky to have found a case like that. I used the case for every upgrade all the way up to the Intel 486 cpu.
Score
0
a b à CPUs
February 5, 2010 2:23:24 AM

yiplong said:
Yeah that's exactly what I was saying, and my little experiment there probably confirmed that to some degree. Mounting the PSU fan facing up or down doesn't seem to make that much of a difference.
I can't believe some folks would, without any actual experience claim that doing it one way or the other will "shorten the components' useful life", or "limit its OC potential" or other crap like that. What a load.



To finish off dealing with you (probably a 10 year old) http://www.frozencpu.com/help/h32/Power_Supply_FAQ.html

And read the part where it says:

"Does temperature affect your PSU's output efficiency and life span?
Temperature greatly affects the life span of many components in your PSU, including your capacitors, resisters and fans. Generally speaking, when temperature decreases 10°, your PSU’s life span will increase by about double. For example, if your output capacitor normally has a 3,000-hour life span at 105°, at 95° it will last 6,000 hours. However, for output efficiency, the same rule of thumb does not apply. Some PSU components achieve higher efficiency at lower temperatures, while others are the opposite."

Now assuming you are old enough to read English... i rest my case (not my computer case!)

Grow up and then come back to the forum to ask some serious questions and listen to some of the stuff people tell you about.
Score
0
February 12, 2010 9:46:47 AM

darkguset said:
Look, don't take things so dramatically. It always depends on the case you are talking about. If there is ample space between the PSU intake and the graphics card (assuming the PSU is at the bottom and the intake is towards the VGA) then there will not be a big problem. If on the other hand we are talking about a cramped case and a VGA that outputs 150W of heat into the system then yes the PSU will suffer.

Also your definition of "seem to function and OC just fine" is far from my definition of fine. Fine for you could be a PC doing a word processing document and the occasional Solitaire and an overclock of 5% and the occasional BSOD. Fine for me is running a server based application 24/7 with no crashes or memory corruptions and OC is also a typical 30% with again no problems at all.

You correctly pointed out that a bottom mounted PSU even fan facing up will draw in less hot air than a top mounted one b/c hot air rises. But that again depends on the case airflow. If there is no airflow in the case, the air will have the same temperature everywhere, hence it will not matter where your PSU is, as long as the intake is looking inside the case it will draw the same hot air.

Regarding the temperature rise the the PSU efficiency and lifespan it is very long and can't be bothered explaining in a forum. You can simply Google it and find out for yourself. You can contact PSU manufacturers like Corsair, Enermax, etc or even email this guy http://www.jonnyguru.com who knows a thing or two and see what he says.



Sorry to bump an old thread, but there is no such thing as a case with no airflow. Hot air will rise pushing cold air down, regardless of whether or not air is flowing through the case, and this movement is airflow. Consequently, the air won't have the same temperature everywhere.
And +1 nitty_gritty.
Score
0
April 7, 2010 3:42:23 AM

An interesting thread, thanks Yiplong for sharing.

I too have been thinking about this since I got myself a new PSU. I got the 600 Watt Nesteq ASM X-Zero which is an heat pipe cooled semi fanless PSU.

I have the Lian Li PC-P60 case which has very good airflow and the PSU has to be mounted on the top.

Now the interesting thing about this PSU is that the manufacturer recommends you install it with the fan facing up. I quote the manufacturer

"Reverse Install Possible - Stack Effect Cooling Solution

By turning the power supply unit upside down (fan facing upwards), the created low pressure can draw cold air from outside to enhance the passive cooling effect up to 42% at 0dB(A) ASM-mode. This leads to even less fan operation and a prolonged period of absolute silence.
"

At first I thought that this PSU would be better installed at the bottom of the case rather than the top, but that would mean the heat that comes off the PSU would hit the GPU and CPU thus raising the temps of these components.

My Lian Li P60 case is designed so that the PSU is installed on the top with the fan facing down, but I have tried it this way and the PSU fan keeps kicking in which annoys me but that does mean the PSU is reaching more than 65 degress (at this temp the fan kicks in).

So I tried it the way the manufacturer recommended it with the fan facing up, the space between the top of the case and the fan of the PSU was very tight, I can barely get a finger to fit into that space, but the P60 case has a huge 140mm fan on the top sucking air out, so this gave me some re-assurance to go ahead it and do it.

Boy was I pleased! Now the fan never kicks in under normal use! In fact the only way to get the fan running again was to use Pime95 to push the system to the max! After stressing the system for well over 2 hours the PSU fan finally kicked in just before the third hour! I was very pleased but still wanted more so I turned the system off and connected the back exhaust fan to a 12volt connector and ran the test again.

Now with the back exhaust fan doing 1500rpm and making a hell of a lot of noise, the system ran for just over 5 hours without the PSU fan kicking in at all! :bounce: 

I got tired after 5 hours and was convinced that with the exhaust fan running like hell, there was no way this PSU was going to reach 65 degrees and kick start its fan.

Money bloody well spent and I never thought I would love a PSU but this is something special and to me my most important component inside my PC case.

Just in case you are wondering, here is my spec

Lian Li PC-P60 Mid Tower
Nesteq ASM X-Zero 600 Watt Semi Fanless PSU
Core 2 Quad Q9550 (not OC'd)
Arctic Cooler Freezer 7 Pro V2 CPU Cooler
XFX GTX260 Black Edition (OC'd out of the box)
4GB Kingston Ram
1 HDD
1 DVD Writer
3 Front intake fans (120mm)
1 Back exhaust fan (120mm)
1 Top exhaust fan (140mm)


I'm replacing the GTX260 for the Sapphire ATI HD4670 for a more HTPC experience.

By the way that Antec 300 case looks sweet, I wonder how the Nesteq PSU would perform in a case where it is fitted right at the bottom with the fan side up.

Cheers
Score
0
a c 144 ) Power supply
a b U Graphics card
a c 172 à CPUs
April 7, 2010 8:18:54 PM

Quoting a frozenCPU FAC:
darkguset said:

...
Some PSU components achieve higher efficiency at lower temperatures, while others are the opposite."

That's either misleading or an oversimplification. The PSU, as a whole, will run better at lower temperatures. Otherwise, why derate at higher temperatures?
Score
0
a c 144 ) Power supply
a b U Graphics card
a c 172 à CPUs
May 13, 2010 7:26:50 PM

agent, look at the date of that article.

I submit to you that cases have evolved somewhat (:) ) in the past two years. And what are known as "gaming" cases have become much more common.
Score
0
October 31, 2011 12:38:11 PM

My novice experience. Just purchaased a Cooler Master HAF 932 and loved the space. Installed my power supply (Ultra LSP750 750-Watt Power Supply) on the bottom, fan side down, as suggested and two days later smelled the smell of burning components. Opened the case side and found the top of the power supply too hot to touch. This power supply has 135mm fan on it. During this time I found my graphics card to be running 10-15 deg hotter than it did in my old case. I flipped the power supply so the fans were on top and the power supply drew air in from the case. Graphics/CPU both started running 20-25 degress cooler than before and no more smell. I have tried to figure out a way to check my actual power supply temps, but haven't found anything that works, but given thet smell is gone, I dont hear the power supply fan or CPU fan kick on near as much and all of the temp readings I can see are much improved, I am assuming all is happy inside the case.
Additional note... The case was NOT on the casters, but on the original "feet" and was on a hard surface, my desktop.
Score
0
!