I have a quiet fan application that doesn't directly involve computers. Mostly around here, we want very quiet fans that we manually adjust their RPM to be a balance between noise and empirically-determined heat (CPU/MB temps). However, my constraint is a minimum total air exchange, with a maximum of a one fan (really opening) for exhaust and one for intake. So, here are my thoughts and questions.

1. My understanding is that total air throughput is limited by the highest CFM fan in a push-pull system. The added intake fan mostly improves the efficiency of air exchange and/or provides additional internal pressure to minimize dust intake. Is this correct?

2. How can one actually calculate the air exchange rate? Simply taking the box's volume and dividing by the CFM seems straightforward, but somehow simplistic and possibly wrong.

3. RPM and CFM vary linearly (0.5 X RPM ==> 0.5 CFM), and are proportional to voltage from 12V base. Or have I got this mixed up?

4. The commercial box I'm looking at has a single 80mm AC fan for exhaust rated somewhere around 25-35 CFM and 36-46 dBA (!!!), but is spec'ed for a somewhat smaller box. I'm chatting with the vendor about modifications and would like to offer them some specific alternatives. SPCR standard 12V fans should be fine with a low-cost transformer (and possibly a Fanmate ). We can porbably build some baffles for larger fans. The box is approximately 3'x3'x4' (36 cu. ft). I think we want at least 2X complete air exchanges per minute--more if possible. Thus, a total minimum 72 CFM. Any recommendations for quiet high air-flow fans?

Thanks in advance. For the record, I use Enermax and Panaflo 120mm at ~800 rpm in most of my computer systems.

Cheers,
Tragus

1. The highest CFM fan in a push-pull system does limit the flow, but the static pressure will reduce flow even more. With both intake and exhaust fan the same speed you will have almost zero static pressure, so flow goes up slightly compare to same setup with intake fan removed.

2. Do you want to exchange all the air in the box? If that is the case things becomes very complex.

3. RPM and CFM seems to follow each other almost linear for most fans between 6 and 12V. If you go outside this range it depends what fan you are using etc.

4. To get 72CFM airflow with just 80mm fans will be very noisy. If you want to go with a single fan (on 1 intake+1 exhaust) you probably want to go with a 120mm fan.

1) IIRC it is limited by the lowest cfm fan (which will be creating static pressure for the highest one to overcome). But a push/pull does help overcome the static pressure of the enclosure between them.

2) You need to account for the effective cfm (as result of the static pressure), but otherwise volume/cfm should work.

3) Broadly correct. (Look for a post here by cpemma that explains the small (+- 1%?) offset in the voltage proportional to rpm argument due to the fans electronics.) This is only true for DC fans, BTW.

4) Forget quiet. Another one of my printouts says static pressure in the average electronic enclosure is 10Pa or more. A quick scan of some fan specs I have to hand says (for example) you would need a (97.7cfm, 43dB(A)) Papst 4312NGN 120mm fan to get >70cfm through a 10Pa static pressure enclosure.

Check out the fan folks web sites for Engineering info as well: Sunon have some good stuff, NMB as well (both of which I printed a while ago). In the meantime I can't find the same Sunon URL and NMB have 'sexed theirs up' here.
The NMB printout I have is
here.

ps - I was posting in parallel with silverarg: I agree with his 'complex' answer to 2 (compared to my simple) because it all depends on the air path through the enclosure. If you have dead spots by design but are not allowed them because of the airchange requirement, then things are non-trivial!

Thanks for quickness to both dukla2k and silverarg. My response:

1. Ah, static pressure! I forgot. So, for my application, having a push-pull would be a significant advantage.

2. Yeah, I'd really like as complete an air exchange as possible. The box could be considered essentially empty for air flow purposes. I don't know how much control I have over the placement of the vent holes, so will have to consider that non-triviality.

3. Thanks. I just needed to reassure myself.

4. I agree about the 120mm, though I've been frightened by some of the CFM v. noise specs as listed at Digi-Key. The NMB B20 looks promising, however. Perhaps it's time to ask Dorothy?

BTW, for what it's worth, the box itself is an acoustically isolated chamber (~50-70 dB attenuation) that will house some rather flat loudspeakers (Manger transducers), some LEDs, some water spouts, and a medium-sized critter for 1-3 hours at a time. It's really for the critter's comfort and health that I need the air exchange, not really heat per se.

This may go where you dont need it, but what kind of critters are those? You are going to have to strap them down with 72cfm blowing past them! (And dont forget the finger guards!

Cant lay my hands on my Heating & Ventilation textbook, but IIRC human occupied rooms are recommended around 2-3 changes per hour. (More for kitchens & bathrooms etc.)

Dukla, the critters are ferrets (mustela furo). Your comments has sent me to do some research about normative air exchange rates according to established institutional guidelines. The vendor, who deals primarily with apparatus for small rodents, gave me the specs. On reflection, his quoted "several air exchanges per minute" does seem rather excessive and actually not in keeping with his current product line. Thank you for the heads up, and I'll post back the results. Sometimes it takes an external head to point out the obvious.

NB: No critters will harmed in the course of these strictly behavioral studies, and all research protocols are strictly approved by oversight review boards. Details available on request.


[Edited for a P.S.]
P.S. A little poking around shows 2-6 ACH (air exchanges per hour) is recommended for humans (more==better for hospital situations), but current guidelines are 10-20 ACH for animal facilities. I'm going to contact the vendor about this apparent discrepensy, but clearly it makes my job much easier (nominal 12 CFM for the high end). Even the SPCR trusty 1LA can handle that!