Sign in with
Sign up | Sign in
Your question

Case fan setup - confused

Last response: in Overclocking
Share
March 22, 2006 12:20:32 PM

I have an Antec case - a performance AMg one (660 I think). It has the following 80mm fan slots -

One on the side panel, pretty much directly over the AGP slot.
Two at the front - one at the bottom, with a nice filter in front of it and one above that which is in front of the removable HDD cage (containing two HDDs currently).
One at the back.

The CPU cooler is not stock, but it's just a fairly standard uprated heatsink with a 80mm LED fan.

I currently have the side one as exhaust, plugged into a fan-only connector from the power supply (temperature-controlled), the front bottom one as intake on a fan-only connector, and the rear set as exhaust on standard power, e.g. full power constantly.
I also have a PCI exhaust fan but I took that out.


My case seems to get quite warm inside, particularly around the RAM and CPU as the graphics card (BFG 6800 Gs) cuts off the airflow quite a bit from the front intake. Also, I'm wondering how these fans interfere with each other, for example will the side fan take out cool air pulled in from the front fan? How does the fan on the graphics card affect everything?

I've also heard that having airflow from back-to-front can work well.

Do I even need an intake? won't the exhaust fan be enough to get rid of the hot air? And won't the intake fan just bring in hot air in the summer? (Although I live in the UK, so this might not be a big issue!)

And if I use the PCIexhaust fan won't that conflict with the one on the grpahics card as they are both trying to pull air away from each other???

I'm not that bothered about noise, it's the most efficient cooling with what I've got that I'm most interested in. Help?
March 22, 2006 12:52:25 PM

You put fans in front of the tower where are your hrd discs to pull air in(disc cooling),and all others out.
Get your self one of these or similiar http://www.zalman.co.kr/upload/product/zm_mc1_c.jpg
And fill all your free fan slots if loudness is not your enemy:) 
March 25, 2006 12:22:41 AM

The side panel fan will work better as a cold air intake, not as an exhaust. You need at least one fan intake in the front to blow over the hard drive(s). The PSU fan and rear case fan(s) should be exhaust. No PCI slot fan is needed or desired. As long as the number of exhaust fans > the number of intakes, this setup will work well. If your room is hot in the summer, well, then the air the case uses for cooling will be hot too...
Related resources
March 25, 2006 12:25:48 AM

The side panel fan will work better as a cold air intake, not as an exhaust. You need at least one fan intake in the front to blow over the hard drive(s). The PSU fan and rear case fan(s) should be exhaust. No PCI slot fan is needed or desired. As long as the number of exhaust fans > the number of intakes, this setup will work well. If your room is hot in the summer, well, then the air the case uses for cooling will be hot too...
March 25, 2006 1:16:05 AM

The most common way of doing things is, in the front, out the back. If a side fan is present, in the side.

So go in the front and side out the back.
March 25, 2006 1:52:35 AM

Quote:
As long as the number of exhaust fans > the number of intakes, this setup will work well.


I'm curious about your logic behind this comment. My aerodynamics engineer buddies tell me that you can have all intakes, all exhausts or a mix without much difference. I remember one of them saying that total flow is a little higher if you only have intake or exhaust fans because you raise the cross-sectional area of the moving air (assuming exhaust vent area ~ intake vent area). Velocities decrease if you only have fans on one side but that isn't necessarily a show stopper.

Quote:
If your room is hot in the summer, well, then the air the case uses for cooling will be hot too...


No doubt one of the harsh realities of ambient air cooling.
March 25, 2006 2:30:18 AM

Quote:
...I'm curious about your logic behind this comment. My aerodynamics engineer buddies tell me that you can have all intakes, all exhausts or a mix without much difference. I remember one of them saying that total flow is a little higher if you only have intake or exhaust fans because you raise the cross-sectional area of the moving air (assuming exhaust vent area ~ intake vent area). Velocities decrease if you only have fans on one side but that isn't necessarily a show stopper...
I normally don't argue with engineers, but in this case I will try. There is a difference between measured fan airflow and case cooling effectiveness. We are not trying to create either a vacuum or a pressure environment. Heat naturally rises, so bringing cool air in from down low in the case, (either actively or passively) is good. Any airflow created at the upper rear of a case should be exhaust. Blowing or sucking air in from the top/rear makes no sense. Many OEM builders use only the PSU exhaust fan at the rear for cooling, so it is true that all the fans can be in one direction only. The more cooling power you need, the more important it is to look at 2 issues:
1) Specific component cooling: CPU, GPU, northbridge and hard drives typically need forced airflow over them. Aerodynamics and high pressure areas do come into play here. That's why cool ambient air is blown over the component and never sucked through it.
2) Internal case ambient air cooling: Flow of cooling air without pressure. A slight vacuum works OK, but we do not want to create hot pockets of dead air that a pressurized system would cause. Think about the airflow in terms of where it will speed-up and slow down, and the argument for equal intake and exhaust flows makes sense. I hope this helps.
March 25, 2006 3:07:38 AM

Quote:
I'm curious about your logic behind this comment. My aerodynamics engineer buddies tell me that you can have all intakes, all exhausts or a mix without much difference. I remember one of them saying that total flow is a little higher if you only have intake or exhaust fans because you raise the cross-sectional area of the moving air (assuming exhaust vent area ~ intake vent area). Velocities decrease if you only have fans on one side but that isn't necessarily a show stopper.

First of all, a slightly more correct notation would be Airflow in <= Airflow out.

We're cooling, not working on best air current.

More intake fans means that the hot air isn't venting out as quickly, and therefore the hot air stays in the case which will ultimately raise the temps a little bit. It will eventualy reach an equalibrium, but it will be at a higher temp.

Meanwhile, too much exhaust and not enough intake will cause a lower pressure, which could impact cooling negitively as not enough air is being pulled into the case to obsorb the heat. However, it does not need to be equal, just "enough".

Intake in the front and side, exhaust in the back and top.
March 25, 2006 3:07:46 AM

Quote:
1) Specific component cooling: CPU, GPU, northbridge and hard drives typically need forced airflow over them. Aerodynamics and high pressure areas do come into play here. That's why cool ambient air is blown over the component and never sucked through it.
2) Internal case ambient air cooling: Flow of cooling air without pressure. A slight vacuum works OK, but we do not want to create hot pockets of dead air that a pressurized system would cause. Think about the airflow in terms of where it will speed-up and slow down, and the argument for equal intake and exhaust flows makes sense. I hope this helps.


Equal intake and exhaust flow is a GIVEN, unless you are either pressurizing your case or pumping a vacuum. For a typical desktop case, the pressure differentials are relatively small (or exceedingly small on a case like a Coolermaster Stacker or ThermalTake Armor since much of those case's surface area is vent)

I'm not so sure you understood my point. Think of it this way:

Example 1: The case has a two 12cm fans pushing air from the front. The case also has a 12 exhausting near the CPU and another out the back of the PS. Let's assume each fan delivers 60cfm. Total airflow through the case is ~120cfm (a little higher because non-fan vent surface will pass some air).

Example 2: The case has four 12 cm fans pushing air out on the back and top panel but none pushing air in at the four 12cm diameter vents along the front of the case. Assume each fan delivers 60cfm. Total airflow through the case is 240cfm.

With a large vent surface area, your pressurized/vacuum argument is really far from reality. My point of view doesn't worry about CPU, NB or GFA fans that service individual heat sinks - I'm talking about choices in dealing with how one deploys fans along the exterior surfaces of the case. If the CPU, GFA and NB fans do their job, the total flow through the case becomes key, especially when you configure large surface area inlets and outlets as described in #2 above. Hope this helps.
March 25, 2006 3:22:00 AM

Sure, that might maximize flow, but since there is not clear intake and exhaust direction, the air ends up being recycled more, meaning that you aren't removing heat from anywhere.

If all vents exhaust, ultimately the intake has to come from somewhere, and since you're surrounding the case with hot exhausted air, you can't help but pull it in.

If you do all intake (not possable with today's active cooling PSU's), where is the hot air going to go?
March 25, 2006 3:53:18 AM

Quote:
...Example 1: The case has a two 12cm fans pushing air from the front. The case also has a 12 exhausting near the CPU and another out the back of the PS. Let's assume each fan delivers 60cfm. Total airflow through the case is ~120cfm (a little higher because non-fan vent surface will pass some air)...
If the 2 exhaust fans were turned off airflow through them would be the 60 cfm created by the intakes. That air has to go somewhere. What is the flow when the exhausts are running?
March 25, 2006 3:55:54 AM

Quote:
Sure, that might maximize flow, but since there is not clear intake and exhaust direction, the air ends up being recycled more, meaning that you aren't removing heat from anywhere.


Air in from the front. Air out is to the top and back. Just like in the push/pull example, except with nearly twice the flow. The hot air going out the blowhole will not magically jump to the intake vent in an all pull arrangement any more than it will in a push/pull arrangement.

Quote:
If all vents exhaust, ultimately the intake has to come from somewhere, and since you're surrounding the case with hot exhausted air, you can't help but pull it in.


I never said that all VENTS exhaust. I said that all FANS blow out the exhaust. The intakes are still in front, just as in push/pull.

Quote:
If you do all intake (not possable with today's active cooling PSU's), where is the hot air going to go?


Maybe you should go back and re-read the examples...
March 25, 2006 12:48:59 PM

Quote:
...Example 1: The case has a two 12cm fans pushing air from the front. The case also has a 12 exhausting near the CPU and another out the back of the PS. Let's assume each fan delivers 60cfm. Total airflow through the case is ~120cfm (a little higher because non-fan vent surface will pass some air)...
If the 2 exhaust fans were turned off airflow through them would be the 60 cfm created by the intakes. That air has to go somewhere. What is the flow when the exhausts are running?

I think you're finally beginning to understand my point! (but that should be 60x2=120cfm) In a push-pull arrangement, both intake and exhaust fans move the same air, for the most part. Some air will enter or exit from additional vent surface that does not have a fan blowing through it. But with push-pull, (if all else is kept equal) you will get higher air speed per fan. With all fans in either push or pull, there will be some non-fan vent surface resistance, so air speed per fan will decrease. BUT!! With an equal number of fans, an all-push or all-pull will have more airflow volume than a push-pull. That's the point!!! Remember that my underlying asumption is that the case has a large vent surface so there isn't alot of airflow restriction.

So to more specifically answer your question above, you're making my previous point: when you take a push-pull and turn off the exhaust fans, the bulk of the air flow does not change in terms of direction of flow. Velocity will be reduced and local to the exhaust fan, turbulence will be reduced.

Don't get me wrong - I use push-pull myself, but I know that first and foremost, you want a clean airflow path past the heat sources and out of the upper part of the case. So I have two 12s pushing air across the HDs. I have two 12's on the back, one exhausting the air off the CPU HSF and one out the back of the PS. I also have a 9.2 alongside the PSU (and therefore up high) and one 9.2 out the blowhole. So two 12s in, two 12s and two 9.2's out. Much of the front of my case that isn't covered by fans is filtered vent space and I have one 5" bay set up as a vent between my two opticals to allow air to move heat they generate. This setup guarantees that air is forced past my major heat sources (PS, CPU, GFA, RAM, HDs, NB, SB, opticals) and also that rising hot air is exhausted.

But I don't do push-pull to avoid a vacuum from forming as you have eluded to. I push air over the HDs to generate a turbulent condition around the them - I have four and they generate a fair bit of heat. The HDs I use don't have any heat dispersion apparatus like a heat sink, so it helps to have some turbulence around them. I hope to eventually build a case that does not rely so much on air flow volume for cooling. With ambient air, you're kinda strapped to local heat sink fans on the CPU, VGA and NB/SB, but I'm sure other designs will work. With all of the efforts to sell liquid cooling, I'm surprised that there hasn't been more diversity in design and engineering of gaming cases.
March 25, 2006 1:10:07 PM

Quote:
Sure, that might maximize flow, but since there is not clear intake and exhaust direction, the air ends up being recycled more, meaning that you aren't removing heat from anywhere.


Air in from the front. Air out is to the top and back. Just like in the push/pull example, except with nearly twice the flow. The hot air going out the blowhole will not magically jump to the intake vent in an all pull arrangement any more than it will in a push/pull arrangement.The point is that in the example he has, all the intakes/outtakes are where the fans are mounted. So all exhaust fans means no intake vents other than through places where the case isn't air tight. This was in reference to my next paragraph anyway, so taking it by itself puts it somewhat out of context
Quote:
If all vents exhaust, ultimately the intake has to come from somewhere, and since you're surrounding the case with hot exhausted air, you can't help but pull it in.


I never said that all VENTS exhaust. I said that all FANS blow out the exhaust. The intakes are still in front, just as in push/pull.
Then I think this is where your example was confusing then, since all the fans are mounted in the vents, so your 2 in and 2 out vs 4 out is using all vents either way.
Quote:
If you do all intake (not possable with today's active cooling PSU's), where is the hot air going to go?


Maybe you should go back and re-read the examples...Perhaps clearing them up a bit might help. If you had mentioned you had sufficient intake in the 4 fan case, and were still doing front in back out, just with no intake fans, that would have been a different scenerio.

Also, of note, if you get 4 fans pulling air from the same "reserve" (the closed case), without an intake fan, each fan will not perform up to it's normal spec, so you'll probably talking 220 CFM in the 4 fan example, if that (Which I'll admit is nitpicky, but it's true)

So, if your 4 out vs 2 in and 2 out theory is correct, 2 out is the equivilent airflow to 2 in and 2 out. So explain why there are examples of case temps and cooling getting better with intake fans vs without them? (Including my own case, in which the HD drops several degrees)

I'm not saying that your engineer friend is wrong, however, perfect theory and functional practice sometimes don't always match. One of the articles I read about a Scythe booth mentioned this, where a simple change to a heatsink that would make you think it might work better worked worse (increasing fin size).
March 25, 2006 2:01:07 PM

Arguments for what/where are somewhat valid.

It has been proven by a number of reviewers with thermal monitors and readers that there are areas that get hotter then others, and that by causing a "pressure" variant type, versus a "vacuum" variant, the pressure designed type system always cools less, and sometimes to the hardwares detriment.

The ATX and BTX specifications take into account the airflow within the system, and how it relates to cooling and case design. Generally, airflow should be from coolest areas, across heat producing areas, and exhasted out in as linear a path as possible.

Antec engineers designed your case to have the two front fans to blow in, populating the HDD cage first; the lower front as an optional (depends on air restriction accross HDDs (How many) and cooling needs at bottom of case); the side/door fan to blow directly on the card area; the rear fan to exhaust the air from the CPU and VRM areas; and the power supply the CPU and VRM areas to a lesser degree, along with the cases upper areas.

Some power supplies do not cool the upper area of the case well, and or numerous cables inhibit good airflow accross "burners" and such, thus the addition of the infamous top "blow-hole".

Not just fan-flow direction needs to be accounted for, fan's cfm rating needs to be known also. Again, strive for an equilibrium, to a slight vacuum. Using to powerful of a fan, especially in the side intake, can disrupt the airflow from the front to back design, causing "eddys", that hurt your cooling more then help it. IDE/SATA/power cables all play into the airflow stream too. That's why good cable management is as needed as thought out fan useage and placement. More is not always better then fewer!
March 25, 2006 7:35:28 PM

Quote:
Maybe you should go back and re-read the examples...


Quote:
[Perhaps clearing them up a bit might help. If you had mentioned you had sufficient intake in the 4 fan case, and were still doing front in back out, just with no intake fans, that would have been a different scenerio.


Way back at the beginning, I did when I wrote:

"I remember one of them saying that total flow is a little higher if you only have intake or exhaust fans because you raise the cross-sectional area of the moving air (assuming exhaust vent area ~ intake vent area). "

So to make things even more anally clarified, my original scenarios assumed that the case's vent surface area was approximately that of EIGHT fans because I mentioned four fans and then assumed that exhaust vent area was ~ to intake vent area. I wasn't trying to be devious or tricky - I put it there in my first reply.

Quote:
[Also, of note, if you get 4 fans pulling air from the same "reserve" (the closed case), without an intake fan, each fan will not perform up to it's normal spec, so you'll probably talking 220 CFM in the 4 fan example, if that (Which I'll admit is nitpicky, but it's true)


Except that the original asumption I made was that the fans would deliver 60cfm. That ignored that actual flows per fan will vary somewhat between a push-pull and an all-pull (and I already admitted that anyway). I understand the flow dynamics but made the 60cfm/fan assumption to simplify. Regardless, a 4 fan rig set up all-out will flow MUCH more air volume than a 2-in + 2-out. No contest, again assuming a free-flowing case like the Armor.

Quote:
[So, if your 4 out vs 2 in and 2 out theory is correct, 2 out is the equivilent airflow to 2 in and 2 out. So explain why there are examples of case temps and cooling getting better with intake fans vs without them? (Including my own case, in which the HD drops several degrees)


I think you are mistaking my argument entirely. I was talking about two different four-fan setups: one all-pull, the other with 2 push and 2 pull. Have you done the experiment with an equal number of fans? That is, have you looked at data from 4 fans out with suffient inlet vent to serve 4 outs vs. 2 fans in + 2 out? Because if you're talking 2 fans out versus 2 out plus 2 in, then that's a different scenario. Remember, I got into the discussion in reply to newf, who wrote:

"As long as the number of exhaust fans > the number of intakes, this setup will work well."

And along the way, I made it clear I was talking push-pull vs all-pull or all-push in a highly vented case. I specifically mentioned the Coolermaster Stacker and the ThermalTake Armor. If you aren't familiar with those cases, I can tell you that a large percentage of their surface is ventable.

Quote:
I'm not saying that your engineer friend is wrong, however, perfect theory and functional practice sometimes don't always match.


Uggh, I never claimed that ANY theory is perfect. In fact, I'm an experimentalist with over 30 years of experience and Iwill go to my grave believing that most scientists fail to grasp the true foundation of the scientific method. But in the scenario I described, there is adequate data to support my friend's hypothesis as described. Having said that, I'll remind you that my case has 4 fans out and 2 in. So I guess I believe in the "Some In-More Out" philosophy. Well, I don't really believe in it, I'm just using it till I can build something more elegant (and much quieter).
March 26, 2006 3:52:08 AM

Quote:
Stuff...

I've never had a case that had enough outs to do 4 outs and no ins without sacrifcing intake areas, so I've not been able to try. My current case initialy had 2 outs (PSU and 1 120MM case fan) and no ins, so I could always just turn the exhaust fan off and try 1 in and one out (I also have another 120 MM fan as an intake in the front now), but I KNOW that will be less effective, since the PSU fan moves significantly less air than the case fan, and the Zalman 9500 relies on the case fan. However, I wouldn't expect temps to rise more than a degree or two, because the Zalman is still venting out air even if the case fan is off (Since it's aimed at the hole).

As for trying with 4 fans, the airflow may add up (More or less, as we've both said, and I understand the simplicity sake addition you did), but there are other benifits of an intake fan, like forcing airflow over parts that could benifit from increased airflow. However, in that regard, 8 exhaust fans wouldn't make up for it. But, that's not what you were trying to say.

So, yes, more fans blowing a single direction will get more airflow than mixing fans doing both intake and outtake (Your point), but not nessesarily better airflow overall (My point).

Anyway, sorry to frustrate you. I think we both sort of missed what the other was trying to say.
March 26, 2006 4:39:19 AM

Quote:
Anyway, sorry to frustrate you. I think we both sort of missed what the other was trying to say.


No frustration here. Talkin' it out is what it's about.
March 31, 2006 1:09:21 PM

Ok.... so it appears there is slight disagreement in the perfect fan configuration! (I had kind of noticed that from other threads here and other websites).

It seems the most common and agreed upon layout is one/two inlet at the front, with at least one exhaust at the back (not including the PSU exhaust).

However, I intend to do some experimentation, seeing as I have some spare time and a somewhat anal attitude towards things like this. :twisted:

I think that in reality case design, component and cable size and layout, airflow within the room containing the PC and many other factors probably mean that there won't be one answer that fits all, but I intend to find the best configuration for my PC.

So, with some brand new fans (the current ones are a big clogged and may adversely affect results) on order, I'm going to check a range of different configurations and see what works the best, using all of the temperature sensors on the mobo etc, plus a digital thermometer.

I couldn't find any examples of this being done previously, so if anyone knows of any similar experiments please let me know.
March 31, 2006 2:24:08 PM

I have an X-Navigator fully populated w/ 13 ide devices (2 optical are at the way top) and all the drives and case ambient are completely cool. here is my setup:

intake:
-for every ~3 hdds a 60mm fan blows in at front of case (the x-nav has removable trays w/ a fan mount there) and I mounted one in the lower section of the 5.25" bays.
-the 120mm side fan is a cool air intake (originally it was mounted as exhaust, and caused some heating issues)


exhaust:
-2 x 80mm in the rear above the ATX plate (behind the CPU, originally were intake and caused some issues as well)
-2 x power supply fans
-80mm top blow hole


Works well, hope it gives you some ideas.
April 3, 2006 3:59:12 PM

Ok, so I did some basic testing on Saturday.

The PSU, CPU HSF and rear 80mm exhaust ran at full throughout the tests.

The three fans I played around with were the front and side ones. After each change of fan configuration, I let the PC idle for about 5 mins and then ran continous burn-in tests in SiSoft Sandra for 5 mins. Due to potential issues with my temperature sensors (see http://forumz.tomshardware.com/hardware/Odd-temps-repor...) I'm not actually sure which temp is which, but I monitored three which before any messing about were as follows (all celcius)-

15-22 (Case?)
28-30 (CPU?)
constant 68 (Power supply?)

The software used to monitor these are Sisoft Sandra, Speedfan and MBM5.

The 15-22 reading is reported as being the CPU in two of the three programs I used, but I now think it is the case temp. Anyway, I'm going for overall effects across the board.

All fans are 80mm. There is a kind of division horizontally in the case made by the grpahics card coming right across to meet one of the HDDs. This line is directly in between the two front fans, the higher (ff1) being sited to blow across the HDD.

Initial setup is side fan (sf) set to exhaust, lower front fan (ff2) set to intake on fan-only power supply (I'm using the Antec Truepower 330 power supply).

Initially, I tried changing the sf to intake as that seemed to be the general consensus here for side fans. The (presumed) CPU temp went up to about 35 on idle, and then rocketed to 50 under any kind of load. i then tried unplugging the side fan which had the effect of lowering the presumed case and CPU temps by about 2 degrees.

Next the intakes - I won't describe all the different settings, but I have actually found the lowest temps with no intake at all. However, after chekcing HDD temps under load I did find that the HDD in front of ff1 warmed up a lot, to a max of 55, so I plugged that fan back in on a fan-only supply. this seemed to solve the issue at the expense of maybe 1 or 2 degrees on the case/CPU temps.

The biggest drop was that strange 68 - at one point it went as low as 40, and currently sites at around 45-50. SiSioft Sandra reports this as a power/AUX temp, and the fan at the back of the PSU seems to run very slowly and there's quite a bit of heat coming out of there. Maybe an issue with the power supply???

So that is the condensed result sof my tests - i actually like my front intake fans as the leds make a nice glow through the front grill of the case, and so I wanted to justify having them there but it turns out I couldn't get them to lower any of the temps except that HDD. And the HDD is crammed in with loads of cables etc so that probably doesn't help.

I also did soem research on the internet looking for people who have done this and found surprisingly few. The main results from the ones I did find were like mine - little or no intake, and one, maybe two exhausts at the back. One site did get better results with a top blowhole, and actually i've found a nasty hotspot at the top of the case in front of the PSU and behind two opticle drives. I'd like to try a fan there but it means cutting my steel case :( 

Anyway, like I've said before, everyone's PCs are different and some may well get great results with front intake fans, but if you havn't tried it without them I'd recommend giving it a go.
April 3, 2006 6:47:56 PM

Quote:
Ok, so I did some basic testing on Saturday.


Cool! Thanks for reporting your work. I need to re-read and digest it a bit.

OK, I re-read. I think you're probably guessing right about the high temp being the PS. Hope to God it wasn't your CPU! 8O Do you remember what change caused it to drop to 40? Maybe having the front fan and the side pushing air in? I'm not surprised about the front fan's effect on the HD temp. And about the blowhole - it's probably worth putting one in since you have located a hotspot there. If you end up having to swap the PS, I'd recommend biting the bullet then. And I'd empty the case to do it, not as big a job as it seems and in the process, you might be able to tidy up cables a bit. Thanks again!
April 3, 2006 7:44:18 PM

Right, I dug out my notes. The previous post was done from memory.

With f1 and f2 on fan-only supply, temps were 27 (board?)/33(CPU?)? from Speedfan and 28/31 from Sandra.

Changing f1 fron fan-only to full raised board(?) temp (the lower one) up by two degrees. No other change.

Adding the side on full power as intake, CPU was up 2 degres, board went down to 28, an average of 4 degrees less.

All of them off, initially the CPU went to 35 under load, 18(!) on idle. At one point it hit 15. The board temp went to 28, and under a full benchmark run (CPU,RAM,Filesystem) went up to 30.
The HDD hit 46 under filesystem benchmark after I moved it lower (it was right next to the RAM and the side of the CPU heatsink). I'm not sure what a good HDD temp is, or how much a game/software will use the HDD compared to the benchmark load, but I presume the benchmark would be the maximum use it will get.

To create a kind of tunnel along the bottom, from the lower front fan across and under the graphics heatsink (doesn't vent outside :(  ) I added a PCI exhaust fan into the lowest PCI slot. As the heatsink/fan on the GPU is on the bottom of the card, and as I explained earlier there is effectively a partition across the case caused by it I thought this might create flow across the bottom of the GPU (I really should write htis up properly with some diagrams!).

The effect of this with the front fan on fan-only power was to raise the CPU/Board temp by about 1-2 degrees, and I didn't get time to check the GPU temp under load but I'm not sure if it will work. The PCI exhaust pointing up towards the GPU fan will fight against the GPU fan's direction, possibly reducing its effectiveness, but it will also remove the heat from this potentially hot area....

I will mess around with that on Saturday.

As an additional conclusion on the previous post, I need to get some more results and move HDDs around but I think if you have space for airflow from the front intake up to the CPU area running one front intake fan on fan-only/half power maybe could help. I don't think (in my case anyway) side fans help, as intake it disrupts the font-back flow too much and as exhaust it take sair unecessarilly away from this main flow.

Top blowholes remain to be investigated.

If you're interested I'll post the links to the similar experiments I found.
April 5, 2006 1:57:10 AM

Quote:

If you're interested I'll post the links to the similar experiments I found.


Hell yeah.
April 6, 2006 10:55:44 PM

Quote:
...If you're interested I'll post the links to the similar experiments I found.

Me too. :) 
April 8, 2006 8:51:03 PM

so in lamens terms for us n00b's whats the best setup:

for front, back and top??

thanks in advance
April 9, 2006 2:01:14 AM

I think there is some consensus here in what works in a vertical case:
1 intake fan down low in the front of the case flowing cool air over the hard drive and into the mobo area.
1 exhaust fan blowing out the upper back of the case. At a minimum this will be the power supply fan.
If you have an opening in the lower side of the case, an intake fan placed here really helps keep the chipsets and videocard cooler.
For every intake fan you should have an exhaust. An extra exhaust is OK.
Down low use intakes. Up high (back or top) use exhaust. Heat rises.
The more heat a system puts out, the more airflow you need (more fans).
Carefully reading each poster's comments, we all seem to agree that this works well.
April 9, 2006 10:36:51 AM

Quote:
Down low use intakes. Up high (back or top) use exhaust. Heat rises.


I almost forgot that theory, thanx.
!