• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

new idea on how to get MAX flow outa your cpu fans without big noise.

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.


Jun 1, 2001
hey guys, im sorry i dont post here as often as i should but i been buisy trying to get a computer working so i can stay online. but heres my idea.

i took a styrofoam cup and cut the bottom off it. then put the lower part of the cup on my fop-38. what this does is create a velocity stack in which air is compacted more dense and forced into the fan. i know its working becaue i would hear the RPM's of the delta 38 go down a little bit which means the fan has to work harder because more air (more dense air) is being pushed through the fan blades. un fortunatly i do not have rpm or temp monitoring equipment or i would test this on my system. anyone willing to give it a go on theirs?
velocity stacks are often used in high speed racing boats and cars where air flow through the engine is critical. i seen ppl use them all the time to get more air flow into their engines, and i just thoght about trying it on my delta 38.
the styrofoam seems to dampen noise and as a result, you can still hear the delta and its just as loud, but the whine seems to be at a lower much more tolerable LOWER frequency.
someone pelase test this for me as im really hoping its a good idea. hey maybe joe or ed will make an article out of it??

CryptokiD, wishes he had temp monitoring hardware.
i bought a compunurse thermal probe for 20.00 on line but u could probably find one cheaper at a local computer shop.
I have done some experiments along this line. The premise was based upon the latest genration of water blocks, where the inlet is positioned so that the jet is onto the surface immediately above the core. It reinforced that fact that the majority of heat in an HSF is in the area directly above the core.

I rigged a variable taper duct that concnetrated the airflow from my cohesive air blower (Hoot Chute) into a smaller diameter, higher velocity stream. Using my 80mmx80mm MC-462A as the test bed, I found that having the airflow narrowed to 50mmx50mm made the MC-462A run cooler at slower blower speeds, but at higher speeds, the resultant backpressure made the blower work harder but netted no additional cooling improvement.

I would guess that since axial fans are considerably less tolerant of backpressure than centrifugal blowers, you could not utilize such an extreme taper. Still it is an area worthy of some experimentation. It could buy you a few degrees of performance.

Your fan probably slowed down because the tapered funnel represents a constriction to airflow and the back pressure interferes with the axial fan's ability to turn at rated speed (ie. slowing it down), thus silencing it somewhat (besides acting as a damper to absorb vibrations) at the cost of cfm. I bet if someone were to use centrifugal fans (NOT those crossflow blowers with thin, window blinds-like blades) you'll get more airflow across -but what about the noise then? The velocity stacks may only be useful in channeling much-neede air right over the hotspot(s) rather than to quiet down fans. Unless some major improvement is made in designing fan blades, the voltage speed-control of fans and the usage of noise-absorbant foam in casings may be the only way to truly combat excessive noise. Maybe we'll use just slow, big fan just like in the iMacs?

I can see why this would work. Not on an automotive theory though. Your not creating better velocity like you would by putting a spacer under a carbeurator. What you are doing is directing the air flow. Fan's blow out at an angle. What you did was trap that air and direct it right onto the heatsink. This will create more pressure, and will slow your fan as it works harder to overcome the pressure, but you are limiting the amount of air that escapes from the side of the HS without cooling much of anything.

How much did you gain in cooling though. Probably not too much if any at all. While you are directing the air more efficiently, you are, at the same time, lowering the flow of the fan. You're probably preaking even in the long run. But hey, anything that quiets that hair dryer is great!
teh way i see it. if you put your hand over a fan. its RPM's get higher because of less airflow and load. so when i put the velocity stack on it, the rpm's went down. which can only mean its pushing more air.
Not necessarily. You will experience the same thing if you raise the pressure the fan has to work against. You raise the pressure, the fans slows down because it has to work harder, and when the fan slows down, it will move less air.

That is what you are doing when you put a "velocity stack" on the fan. It channels the air into a smaller area (remember fans are designed to blow out at an angle, not in a round tubular shape), hence increasing the pressure, slowing the fan down, and blowing less air.

This is why I don't believe in positive pressure cases. It causes the intake fans to work harder against the pressure, blow less air, and cause extra wear on the fans. With a negative pressure case, all your fans work at their rated speed, blow their rated amount of air, and are more efficient, and they last longer. But dust is a draw back.

You have an automotive look at this. And here's a good example. Sure an open velocity stack on a carbeurator will add horsepower, but they are the same size as the flow of the bottom of the carb. If you make it smaller than the flow pattern of the carb, you restrict the air....hence the "restrictor plates" used in NASCAR. Which, in the end, lower horsepower

.It's also the same technique that was used in the gold mining days in the wild west. They would channel water down a trough. As the trough went down the mountain the trough got smaller and smaller. This ended in a spout. This was used to pressure spray the mountain and cause the dirt to run off.The water gained pressure from getting smaller and small and going faster and faster. But the flow coming out was less than the flow going in.

Anytime you increase pressure or resistance it will always result in less flow.
well to all you non beleavers, i have confrmation that it works! a guy i know tried it and got 5 F reduation in cpu tempature. and at only the cost of a styrofoam cup.

so HA! :)
an you guys are talking weird i dont know if your understanding what exactly im talking abot so ill attempt a little picture

\_/ < velocity stack
|_| < fan
|_| < heatsink

that is the principle and it works! and this is an early design with more time maybe even lower temps could be reached.
ok i have a second conformation from someone who tries this and it works.

here is some pics of velocity stacks in more common configuration
<a target=new href="http://valkyrie.safeshopper.com/images/bu0odzxq.jpg">one</a>

<a target=new href="http://pedhead.safeshopper.com/images/bl15l0wy.jpg">two</a>

<a target=new href="http://www.oldbritts.com/image/40_800106b.jpg">three</a>

<a target=new href="http://www.oldbritts.com/image/40_800106a.jpg">four</a>

<a target=new href="http://www.oldbritts.com/image/40_800106c.jpg">five</a>

<a target=new href="http://www.chromebillet.com/500_mount.jpg">six <- the best one</a>
If that diagram is indicative of the setup and the test was conducted with the case upen, to accomodate the cup, then all that was accomplished is preventing the fan intake from drawing warmer air into it. If you take a parallel walled piece of ducting, no matter what it is made from and put it on the intake of the fan, it will accomplish the same thing.
Ponder this for a moment. The fan is blowing air down into the HS and it washes out the bottom warmer (hopefully) than when it went in. What do you think that air does? It deflects off of components and barriers around the socket area and a good deal of it bounces back up. Then, that warmer air is sucked back into the fan. The process repeats with the air getting slightly warmer each time, up to a point. By extending the intake upward, futher out of the case, the air up there is cooler than the air down in the case. Cooler air into the HSF yields a reduction in temps. I have a 3 inch square piece of duct 1 foot long. When I'm running my MC-462A with the Delta fan and the case open, if I put the stack on the delta intake, the cpu temp drops about 3-5C. Not a miracle, just some common sense. If you put the cover on the case and have adequate throughcase airflow, the gian from an intake stack is lost. An intake blowhole will return the effect and yield a temperature drop.

Heh, I was expecting to see some pics of a cup crammed on a HSF. I think we know what automotive velo stacks look like though :D
come on u guys. theres already enough negative remarks on this. i dont need mroe. what i need is TESTERS.

you can tell me you think it doesnt work. fine. but i already have 2 people proving it works, and no one that can prove it doesnt.
so either test it out to try and prove this wrong. an it will sink to the bottom like so many others. but you cannot just say you think it wont work.
I think your missing the point Hoot says it works and then told you why and i have to agree with him, i'm running a duct and i get great results because i'm drawing cooler air in to my fan your cup idea is doing the same thing so we know a certain % of your temps dropping is just the ducting away to draw in cooler air, what you need to do now to prove your point is to try a square duct the same length as your cup and then test for a temp difference between the two.
Yeah, I am in support of using velocity stacks (or fandapters or whatever) for better cooling CPUs but NOT as a bona fide way of quieting fans. So the thing is:

1. Mass flow rate of air IS lessened somewhat due to constriction to axial fan. Try centrifugal fans.
2. Air is channeled directly to the hottest part of the HS just right above the CPU core.
3. A tall stack prevents hot air recirculation.
4. 'Soft' materials used as stacks dampen out fan vibrations (one major contributor to noise).
5. It isn't so hard to build at home. Just make sure the stack isn't so tall as to push the fan up against the side panel (another constriction to flow) and that the stack is durable enough or it'll fall apart and the fan will slaughter all the tiny mobo capacitors and stuff.

So it'll definitely reduce temps while giving you a reason to get that 80 or 92mm fan you've been itching about for the past months.
uhm actualy u guys i have 2 buddies of mine from other forums telling it it works.

one guy gets a 5F reduction and the other gets 300rpm less spinning and the same air volume which leads to alt less noise.

bolth say they are happy,.
i guess there is diffrent watys of making it for what you want to do. because theres 2 diffrent outputs on this from 2 diffrent people. damnit i wish i had rpm and temp monitoring . my new 8k7a+ has it but my tbird is about 400 miles away right now in a fed x truck.

i think the ideal velocity stack for a computer is as tall as you can make it but not more than 6 inches and with a nice expansion in the width as it gets longer.

p.s. i tried ur theory on a square velocity stack i made it out of cardboard and duct tape and i couldnt get a diffrent rpm out of it.
i know human ears are not good at detecting rpm changes but on my coffee cup fan i can tell when i take it on or off. but with square one i hear no change except its a bit quieter.
when my 1.33G tbird comes in im doing a full reivew of this.
I'm sorry Crypto, I was under the impression that you were putting the cup between the fan and the heatsink. Now that you put your graphic up I understand better.

Yes, I can also see this lower temps. For the exact reason that Hoot put up, your sucking clean cool air.

The only thing that I don't buy is the 300 less RPM and same air volume. It is impossible, the air moved is proportional to the RPM. The faster the fan spins the more air it pushes, the slower it spins the less air it moves. The only way to get more air from the same RPM is to alter the fan blade pitch.