Use of Inline Centrifugal Fans In Computers?

[Anoolios]

Does anyone know if inline centrifugal fan designs have been used in computer cooling?

An "inline centrifugal fan" differs from a normal bladed "axial" fan in that they can develop significantly higher static pressure. Inline centrifugal fans are extremely efficient and can also be extremly quiet.

There is some info on inline centrifugal fans at these manufacturerers, however they only make large house-sized units, I used one of these to build an air filter and it is a truly amazing piece of engineering:
http://www.fantech.net/fr.htm
http://www.elicent.it/inglese/axcmet.html#

If you take a look at the cut-away diagrams on those sites you can see the construction using a bladed, radial impeller that has a close fit inside the casing so that air cannot get around it.

I have seen an implementation of "cross flow" ("squirrel cage") fan design by Cooler Master:
http://www.coolermaster.com/index.php?L...B-V81&other_title=AAB-V81Aero%20Blower

However the cross flow fans are significantly less efficient and more noisy than inline centrifugal fans, especially under high static pressure.

Has anyone seen inline centrifugal fans used in computers or have any ideas about this?

I'm not an engineer believe it would be possible to efficiently and quietly cool an entire computer system (with an appropriately designed case) with a single extremely efficient and tiny inline centrifugal fan...

 

[Howard]

I have never seen the application of the fans you describe to cool computers. Not that that means they've never been used for that purpose...

 

[IEC]

Inline centrifugal fans would be great for localized, high-velocity airflow, but that would require a system of pipes/ducts to maximize the efficiency...

At present I think 120mm fans at 5-7V work best with the cooling/noise ratio. No problems here with 120mm exhausts on both my systems.

 

[mindwreck]

I would think the reason why they're not used is because if they size and theres just no reason to require that much pressure.

 

[Anoolios]

Part of my thinking is that while axial fans are decent for moving air around a nearby heatsink, they are not very good at moving air through a case.

The volume of air that a fan can move in a given length of time is generally referred to in CFM units (cubic feet (or meters) per minute). The misleading thing about CFM measurments as they relate to axial fans is that their efficiency drops dramatically when they operate under static pressure, as is the situation when used for case cooling. Inline centrifugal fans are able to operate efficently even under large static pressures.

A small inline centrifugal fan operating at the same wattage as an axial case fan would be able to push or pull many times the volume of air per minute. If the case was sealed and the airflow directed over the major components (maybe some type of manifold?) the whole thing, including power supply, could be passively cooled.

The only things I can think of that may prevent usage of inline centrifugal fans in computer cooling:

Cost; these fans can be quite a bit more expensive to construct and require higher quality components.

Noise; while inline centrifugal fans are comparatively much more quiet than other fans, the sound of such copious air movement needed to cool a computer may be significant. House- sized centrifugal fan units sometimes use baffles to mitigate air movement noise.

Anyhow, I hope I can find a small inline fan somewhere and try this out.

 

[Kakumba]

ok, without using those links, my understanding is this:

These fans offer high CFM, but quite directional? thus, while not so good for moving air all through a case, they would be HIGHLY effective for using on a radiator in a watercooling setup?

If I am going off on a wild-ass tangent, yell at me, otherwise, im going to look into this a lot more when I get home from work....

 

[Howard]

Originally posted by: Kakumba
ok, without using those links, my understanding is this:

These fans offer high CFM, but quite directional? thus, while not so good for moving air all through a case, they would be HIGHLY effective for using on a radiator in a watercooling setup?

If I am going off on a wild-ass tangent, yell at me, otherwise, im going to look into this a lot more when I get home from work....

To my understanding, air is about as directional as water, not light. e.g. a case with only 1 fan, exhausting, will pull air from all entrances of the case to blow out.

 

[0roo0roo]

higher static pressure sure, but if you want lots of air, it'll get noisy very quick. i once in the pentium 3 days had a 5 inch bay drive cooler that stuck out part way cuz it had such a fan. was cute for kicks, but it wore out, maybe mine was defective, but it was mostly a gimmic for me. you just don't benifit from the higher static pressure of those fans. and well tests of those smaller ones on cpu hs's show they don't have any benifit there either, besides for the interesting look. generally they actually under perform which is why most don't have them

 

[0roo0roo]

oldie but goodie. fans 101 on the side column has a bit on the difference http://www.dansdata.com/hx45fan.htm

 

[Anoolios]

Kakumba, an inline centrifugal fan will move air in a more directional flow than a standard axial fan and may be useful for highly directed spot cooling, I think this is also worth looking into.

However, my main point here about inline centrifugal fans is that they are the best suited design for environments that are airflow-impeded and under static pressure, as computer cases are . I'm fairly certain that in a non-pressurized, airflow-unempeded environment that axial bladed fans will move slightly more CFM per Watt than inline centrifugal fans, but when there is even the slightest static pressure or airflow impediment, the inline centrifugal fans will move far far more CFM of air.


0roo0roo, the bay cooler fan you had was likely not an inline centrifugal design, there have been lots of cheap (and a few very well made) "squirrel cage" type fans that are significantly different from inline fans. The squirrel cage fans tend to be very noisy and are not efficient at high static pressure.

The Master Cooler units I pointed out ( http://www.coolermaster.com/index.php?L...B-V81&other_title=AAB-V81Aero%20Blower ) are one type of axial squirrel cage, delivering a directed but diffuse airstream without the central void of airflow present when using a standard axial fan.

Pansonic and Antec make another type of squirrel cage blower that delivers a more directional airflow at higher static pressure than standard fans, but much noiser and less static pressure than inline centrifugal: http://www.antec.com/us/productDetails.php?ProdID=77094 This is probably the type of fan that was in your drive bay unit. Industrial inline centrifugal fans are extremely reliable and quiet. My 8" AC powered inline centrifugal fan has been in almost constant operation since 1991 and moves more air than any similar sized axial fan I've ever seen (granted it is extrodinarily well engineered by Fantech). The thing sounds like a jet airplane taking off if you start it up without ducting attached, not because of the fan motor making noise, just the huge movement of air, its quite quiet with insulated ducting attached.

 

[Anoolios]

Great link 0roo0roo, I love the smoke test he did! The spirit of what he did is similar to what I'm suggesting, move a friggin' ton o' air though the case! :-D Only he went after it with brute force (strong AC powered axial fan) while I'm going for highly efficient, sleek, and sexy.

If I cannot find a small inline centrifugal fan I may try using the Fantech FX 4 model ( http://www.fantech.net/fx2.htm ) in an experimental project. The FX 4 has a 4" duct connection size and runs on only 19 watts max which is really tiny for one of these large fans. I could mount the FX 4 in an external box with an household air filter over the inlet, then connect it with flexible duct to a sealed computer case built of plywood. Remove all fans from the computer components, place a few exit channels in strategic spots and voila! Combination room air filter and passive cooling solution

 

[Kakumba]

Hmmm, this could be interesting for some extreme air cooling projects, having had a bit of a read. Im more of a watercooling person, but this could be interesting....

 

[Zap]

I believe that in a well-designed case there won't be enough static pressure to give centrifugal fans that much of a benefit over a decent axial fan. Centrifugal fans use numerous smaller blades with tighter tolerances. Those two aspects help it increase static pressure, but also help increase noise.

BTW, I believe that the inline centrifugal and crossflow/squirrel cage fans are actually really similar. Both draw air in through the center of the cage/centrifuge and expel air perpendicular to the intake. The squirrel cage directs all this air into one duct still perpendicular to normal exhaust while the inline makes the exhaust turn another 90º turn to expel air inline with the intake, meaning the air makes two 90º turns.

I actually have a fondness for squirrel cage fans, having owned some Nidec Gamma series fans in various sizes (26/28/30), some similar Delta fans in two sizes and a really "unique" large centrifugal fan with no housing (never found a use for it though, waste of money ).

 

[Anoolios]

Zap, I'm not sure if its possible, but I'm thinking about a much higher rate of airflow through the case than what is normal or would be possible with regular axial fans. :evil:

And yes, the squirrel cage designs are also centrifugal fans and very similar in construction to the inline ones, however the operating properties differ significantly between the two. Inline centrifugal fans are often used where long duct lengh and/or airflow obstruction/pressure occur, I think squirrel cage are more suited to moving a diffuse stream of air under lower pressure (though greater pressure than a normal axial fan). I looked at those Nidec fans and they have some nice airflow info on the site, what did you use those for? I've got a large squirrel cage fan on my roof (swamp cooler) that keeps my house nice and cool.

 

[trexpesto]

Small blower fan

Small 5.25 laminar-flow type fan
- Better pics at the manufacturer. They claim 21 db at 100 cfm

Also the size and material of the filters you use can make a big difference.
And put on a front fan blowing into the case.

 

[Zap]

Originally posted by: trexpesto
Small blower fan

LOL, my larger Delta blower looked like that in shape and size, but was higher RPM. It was a MONSTER!

Anoolios, I'd think even if the fan was 100% silent itself, the airflow would cause noise. Again it comes down to noise versus cooling.

What else did I use the Nidec fans for, but cooling? I used them for additional cooling on my Voodoo2 video card, extra exhaust spots in my case, cooling on my overclocked Celeron 300A... eventually I used the smaller Delta (really reasonable noise levels) for cooling on a socket A setup. Haven't used them since. I think I still have some in the garage somewhere.

 

[Anoolios]

trexpesto, all those fans you linked are squirrel cage types, I think the technical term is axial centrifugal. I'm looking for small (on the order of 80-120mm) inline centrifugal fans that are constructed similar to the Fantech and Elicent models I linked to above.

Zap, I definitely think airflow noise would be an issue if no component fans are used, that would take a lot of airflow! But if we are talking about just replacing the case fans with a single tiny inline centrifugal fan, then the airflow would be of a similar rate.

 

[trexpesto]

I always thought it would be funny to mod out one of those window-filling box fans..

edit: how about modding your house to install a hose for suction, then you can use a full size unit mounted remotely?

 

[Howard]

Originally posted by: Zap

Originally posted by: trexpesto
Small blower fan

LOL, my larger Delta blower looked like that in shape and size, but was higher RPM. It was a MONSTER!

Anoolios, I'd think even if the fan was 100% silent itself, the airflow would cause noise. Again it comes down to noise versus cooling.

With enough elbow-grease and ingenuity, one could DIY a noise-suppressing baffle like the ones they make for home applications.

 

[wwswimming]

thanks a lot for the Links.

centrifugal blowers are better for moving air through a small duct. the plenum from the fan output
to the duct inlet is a little easier to build.

i've seen centrif. blowers mounted on PCI cards. i think they just used the PCI slot as part of the
supporting structure, like a bracket.

one project we worked on, we needed a huge amount of airflow in a 4 inch diameter to cool
10 kilowatts with 10% duty cycle, radar stuff. because we had limited space, the only solution
i could find cost $5K. i was the mech. engineering supervisor - boy did i catch heat for that !

in hindsight, the $100 centrifugal blower would have been comparable, one made by Dayton
that has a 4 inch diameter outlet section. only problem is, it used up about 4 times as
much space as the $5000 blower. the system was for South Korea. getting a spec change
on the box would have used up months of engineering time, about a $million. sometimes
it costs less money to use the expensive part & stay within spec.

plus the VP of engineering was a heavy tobacco smoker, and so were the customers,
filterless cigarettes. talk about smoke-filled rooms. when they were having meetings, say
if i needed a signature, i'd go in, talk as little as possible, and split fast.

wanting to avoid that smoke-filled room scene was another reason we engineers avoided
spec changes. this was in 1989. in hayward. near the epicenter of the "World Series"
earthquake. that would be another story.

i think that, per CFM at inches of mercury or water (a measure of pressure), the vaneaxial
fans use less space.

if you were going to attempt an OC record on air-cooling and could attach all the fans
you want, i'd use the $100 centrifugal. i have a similar fan at home - $15 surplus,
works very similarly.

 

[Zap]

Originally posted by: Zap
I actually have a fondness for squirrel cage fans, having owned some Nidec Gamma series fans in various sizes (26/28/30), some similar Delta fans in two sizes and a really "unique" large centrifugal fan with no housing (never found a use for it though, waste of money ).

Here's a link to that unique squirrel cage fan that doesn't have a housing. I think I pretty much got all my squirrel cage fans from All Electronics except for the first couple of Nidec Gamma 28 units from Radio Shack.