• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

80mm rear case fan w/duct to CPU heatsink

mcvan

Member
With two hot computer buzzing away loudly under my desk, I am looking for some kind of solution for quiet without risk of overheating:

All thing being equal, the larger the fan, the more air it moves. So rather than take the already warm air in the case and shoot it through a whining 7k rpm 60mm fan, why not use a good heatsink (like the GW32/38) and a larger quieter fan to move cooler air from outside the case through a duct?

The rear case fan hole on nearly every case I've seen is designed to fit a 80mm fan, and it is only 6"-8" away from the top of any heatsink mounted on a socket CPU. All you'd need is some kind of flexible tubing and an interface for the heatsink (like the 60mm to 80mm adapters people are talking about on this forum).

The vaunted (screaming) Delta fan is only rated for 37cfm and is LOUD at 46dBA! Someone else mentioned Mechatronics (http://www.mechatronicsinc.com/products.htm) -- they have an 80mm fan that will do 46cfm at only 32dBA!(model F8025E)

Anyone try this before??
 
I agree, go large and concentrate the air flow where needed. Not an uncommon design from quality box makers. Small high RPM fans are just plain dumb. My rig is an old school InWin A500 sporting a BX board with FCPGA in Slot adapter so the PS fan (flipped to suck) is within 1/4" of the adapter. A CPU fan is usually not required anyway in this case it definitely is not. If possible try to direct the flow throught the PS so an additional fan is not even required - and make sure you are using a quiet one. I'm not familiar with the heatsink brands/models but the best would not be one designed for those stupid little fans - that is you want one with larger fins spread apart fairly wide without any obstructions (a real passive heatsink). You could prolly fashion some ducting if necessary out of thin paperboard and duct tape, else poke around Home Depot. Just some thoughts. Death to noise 😛
 
Thanks for the thoughts. Found a nearby electronic parts shop w/ lots of fans to poke at... You mentioned

<< Not an uncommon design from quality box makers. >>

-- which ones? Maybe I could pick up a few tips from their websites.
 
What are you cooling with this? I don't think a PS fan can cool both a PS and a cpu all by itself.

I recommend using the ducted case fan, and have it blowing OUT. Dunno why, but in real world performance, a single output fan always does better than a single input fan.
 
Dell, Gateway, Compaq, HP ... I think they all forgoe CPU fans when possible.

Mine runs at 825 (sold as 533). It helps to use a CPU that is sold at a frequency that is not pushing that line's design limits. That is, the same production is cranking out chips which they sell at higher freqs.

Remember, this stuff is geared towards the lowest common denominator like being able to stand up to 24/7 number crunching in the Sahara desert during spontaneous combustion season (okay, well maybe not but you get the idea). In a typical 72 degree fahrenheit environment most fans are just silly. Another thing that peeps seem to ignore is the fact that bothering to cool a CPU a few degrees with a fan is not going to make it run faster, it just allows the potential to run faster which depends on increasing the frequency. So if it runs without freezing at a given freq it is silly to add a noisy little fan.
 
Most of the discussion i've seen centers around Alpha 6035 sinks ducted to a rear exhaust fan. My p3e550 runs 787 set up this way with a PCP&amp;C 80mm Silencer in an Antec sx-1030. I fabricated adapter rings for the sink &amp; fan from thin aluminum stock and aluminum duct tape and connected them with 3&quot; aluminum dryer vent hose.

I'm sure that excellent results could be had using a side blowhole and ducted fan either pushing or pulling, depending on the sink.

To use really quiet fans, a big sink is necessary to transfer heat away quickly. Obviously, higher speed processors, particularly T-Birds, will require stronger fans- there IS a limit to sink size...
 
mcvan,
Your proposal is fundamentally very sound, some time ago Intel developed a &quot;Low Profile Cooling Duct&quot; which led cool air from rear panel to chipset and cpu area of ATX case design. Duct had an integral fan. We also duct cool air to the hs/fan from front panel inlet fan of an AT case design. So you are in good company.;-)

However, two issues to consider.....

1:- Is it wise to &quot;remove&quot; the hs fan?? Consider perhaps replacing existing fan with a slightly lower capacity/quieter one if noise really bothers you.

2:- More important...Consider the effect of your new &quot;inlet&quot; flow fan on the &quot;overall air flow&quot; patterns and throughput, don't have air flows fighting each other. And consider possibility of recirculating warm air exhausting from the power supply back thru your nearby new &quot;inlet&quot; fan...sucking local air. Intel recommended a baffle to prevent such recirculation in their LPCD design above. Then also consider that the best airflow design has inlet air flow approx = outlet air flow. So do a little design work/rework considering all the case fans and these guidelines.

Might consider mounting new fan in side panel instead of rear panel, might be better re air flow &quot;patterns&quot;. In any case, think &quot;overall&quot;,
not just &quot;add a fan&quot;.
John C.
 
I've got a pretty small desktop case with a 1.5&quot; by 3&quot; vent near the socket, but not big enough for even a 60mm fan. Using wire mesh and duct tape, I created a simple duct between the vent area, and the top of the heat sink, and put in a quiet generic 80mm fan (sucking air in). The fan is suspended at an angle inside the duct I created, that is the only way it could fit. I also put a deflecter on the PS exhaust - so all the air will go 'right' after leaving the case, and not make two lefts and re-enter. The duct isn't air tight, and not well attached to the heat sink so air leaks. But even the generic 80mm fan moves more air than the 60mm fan that was on the hs. The cooling seems about the same as before, but much quieter. I recommend you try it ducting and using larger fans. Wire mesh and duct tape worked well for me.

My next step is to remove the tiny, silly fan on my Radeon...
 
Thanks for all the input, fellas. Obviously, others HAVE done similar things before.

FIY, my main system:
T-Bird 850 at 140x7=980mHz
Abit KT7A Raid
Landmark 298 case w/a front case fan blowing in
SH-340W PS
GlobalWin FOP32-1 w/4600rpm fan
2 IBM 46G 75GXP drives for RAID 0 (These are fast &amp; quiet)

The internal MB temp sensors with MBM5 shows idlie temp at 28-30C, climbing to perhaps low 40s under heavy load. So the cooling IS quite effective, it would appear, and I have no random lockups or crashes of any kind in W98SE, multitasking with many programs.

The noise by itself is not TOO bad, mostly lower pitched, but quite noticeable. Most of seems to originate from the GW cooler fan. I am amazed at how much the case cuts down on the decibels, especially the higher pitched fan whine.

But I DO have a second machine:
Intel P3-550 (slot 1) oc'd to 792 (144FSB on a MSI BX-Master)
CO-P301 CPU cooler
generic ATX box
IBM 75GXP 30G drive

This machine is slightly noiser, with more high frequency content. The case is ok, but doesn't keep the noise in like the Landmark. Lighter panels, lower rigidity &amp; weight. The cpu temp measures a couple degress higher -- different MBs, CPUs -- who knows where the difference really comes from. Again, it seems cool effectively enuf, as the machine is very stable. The case has a bunch of holes punched on both of the side panels, presumably to keep air circulating, but I think all they do is let more of the noise out.

The bottom line is that while each machine on its own is ok, together, their combined noise can drive me nuts.

I know, I know, if I was a true blue enthsiast, I'd grin and bear it, right? But I work with these suckers all day, sometimes 10-12 hrs. I don't want to feel like I work in a factory.

I did pick up an 80mm Orion fan -- rated at 40cfm @ 33dBA. After poking around in a couple plumbing stores, I can see that my challenge is going to be the tubing to direct the airflow -- in or out -- effectively at the fan. The details of your rig, Jhhnn, will be very helpful for me. As I type I am thinking about looking through the recycling bin for cans of the right diameter...

The fan I got is considerably quieter than the YK fan on the GW, which is rated at 36dBA and 26 cfm. I just KNOW that if I can find or build the right kind of tubing to direct the airflow, reduced turbulence and even airflow will drop the noise by more than the 3 dB rating difference between the fans.

Another alternative is the new Thermatake Volcano II cooler, a conventional design which has got some pretty promising reviews already: http://www.ocaddiction.com/reviews/cooling/volcano_vs_vantec and http://216.122.74.171/reviews/cooling/volcano2/1.htm. The big deal about the Volcano is the fan: 36cfm at only 31dBA! Just replacing my GW fan for this might do the trick! And I've seen the whole unit advertised for as little as $17-18: http://216.157.14.244/ttvol2.htm

With the P3 slot-1, I think all I can do is find a quieter fan, block those holes and add dampening to the case.

BTW, johncar, you say

<< We also duct cool air to the hs/fan from front panel inlet fan of an AT case design >>

Who is &quot;we&quot;? Do you work in a case design/mfg company?
 
I think what you are talking about, I've seen in all newer Dell computers. They have just a heatsink on the processor with a plastic tube or tunnel that leads out the back vent.
 
mcvan,
>I can see that my challenge is going to be the tubing to direct the >airflow -- in or out -- effectively at the fan. The details of your >rig, Jhhnn, will be very helpful for me. As I type I am thinking >about looking through the recycling bin for cans of the right >diameter...

&quot;Adaptive engineering&quot;...anything cheap and easy that works. We used
thin sheet aluminum, (and cardboard near mobo to prevent electrical shorts),..cuts with cheap scissors...make/test templates out of cardboard...like making boxes out of sheet cardboard...fold to shape and tape together with masking tape. Or &quot;adapt&quot; miscellaneous hardware as you already suggest.

>BTW, johncar, you say
<< We also duct cool air to the hs/fan from front panel inlet fan of an AT case design >>

>Who is &quot;we&quot;? Do you work in a case design/mfg company?

No, but we sounds more polite and less egotistical than &quot;I&quot;..got into the habit long ago....and not the Royal &quot;WE&quot;.;-)
John C.
 
Motero, you're absolutely right. Just checked tech documentation at Dell and they're using what they call a &quot;fan shroud&quot;. (http://support.dell.com/docs/systems/dzuul/rr.htm#socketed_processor) The cpu and rear panel fan are aligned so there's no need for anything more than a simple 90 degree angle plastic molded tube. No such luck on my systems...

I think these guys (Dell, IBM and whoever else) are ahead of the &quot;high performance&quot; aftermarket. Nobody's come up with anything like this for a generic case solution. As a manufacturer, it'd be easy: integration of a quiet high cfm 80mm fan, flexible tubing, a good heatsink. How come no one's done it yet??

NAC, wire mesh &amp; duct tape, huh? That sounds worth exploring too. Thanks for the tips, everyone! 😉
 
mcvan,
> think these guys (Dell, IBM and whoever else) are ahead of the &quot;high >performance&quot; aftermarket. Nobody's come up with anything like this >for a generic case solution. As a manufacturer, it'd be easy: >integration of a quiet high cfm 80mm fan, flexible tubing, a good >heatsink. How come no one's done it yet??

You haven't described an &quot;integrated product&quot;...just 3 separate components that users would have to install in many &quot;different&quot; cases/systems. Whats new?? We and others have been ducting cool room air for years.

And likely the Dell system uses the airflow of the fan feeding the duct as one of its inlet case airflow fans...possibly eliminating the hs fan...remember OEMs don't oc/have temp problems...they're looking for the most &quot;cost effective&quot; cooling solution/s in &quot;standardized&quot; systems.
John C.

 
I've been planning this same sort of mod all week - a shroud or duct from my heat sink to an exhaust fan. No question in my mind that one of my 80mm (quiet) Panaflos will suck more air through the sink fins than the little high speed Deltec with the big whine. It also is mounted to suck. This arrangement should be close to ideal - quiet, effective, also draws cooler air over the chipset. The only thing to watch is making sure one is really pulling through the heatsink, and that there is enough air coming into the case.

My own question is about wiring the Panaflo, and I would appreciate anyone who knows something specifically about them clicking on over there, and giving me the benefit of their knowledge.
 
Success! 😀

With a bit of flexible 3&quot; diameter ducting, some wire mesh and duct tape, I managed to get my homebrew cooler working. When I first got the thing set up, the noise was horrific -- much worse than with the GW FOP32-1. Some experimentation led me to suspect turbulence at the &quot;grill&quot; of the case fan hole. I cut it all off with a small hack saw, vacuumed up the metal bits, then put it back together again. It works pretty good. The noise is only 3-4 dB lower than before, but that's all the fan specified anyway. The cooling seems a bit better than before, maybe by about 2C overall.

The ducting was a cool find -- think of a skin of slightly rubbery plastic that's been fused on a slinky toy in its most stretched condition. 3 meters cost me $4. I think I might have used 6&quot;.

Still can't help thinking that a Termaltake Volcano II with its 31dBA 36cfm fan would have avoided all this...

Sorry, Homer, can't help you with the Panaflo fan wiring. Have you found their web site?
 
Homer, you wrote........
I've been planning this same sort of mod all week - a shroud or duct from my heat sink to an exhaust fan. No question in my mind that one of my 80mm (quiet) Panaflos will suck more air through the sink fins than the little high speed Deltec with the big whine. It also is mounted to suck. This arrangement should be close to ideal- quiet, effective, also draws cooler air over the chipset. The only thing to watch is making sure one is really pulling through the heatsink, and that there is enough air coming into the case. End quote...

Just a suggestion, consider blowing &quot;cool&quot; room air into the &quot;reversed&quot;/downflow flow hs fan. Sure you lose the benefit of slightly better upflow performance, (Alpha hs???), but you gain by feeding the hs cooler air...and every degree cooler air to hs = same # of degrees cooler cpu internal temp. Plus, you are increasing the cool air flow thru the case....as you say &quot;enough air coming into the case&quot;. Think the tradeoffs will result in cooler chip.
John C.
 
Back
Top