A better fan design - the search continues

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I had this idea for el-cheapo cooling for class A amps with fans. I've seen fans mentioned a lot of times, but i'd like to know if someone has tried something like this... it basically a square metallic tube:

---------------------------------
....|------------------------|....
....|------------------------|....
....|------------------------|....
---------------------------------


The vertical lines would be a couple of small 12vdc PC fans running at 9v, both blowing to the same direction; the horizontal lines strips of Al running through the tube (any decent Al extrusion would do here), and i could place foam on the inside of the extreme sections to damp the fan noise even more (those pretty dots). I think i could make it real small and have it removing a LOT of heat quite silently, the thing is that dust accumulation is a concern for me.
What do you guys think? I'm running out of ideas, and i've already ran out of cash :bawling:
 
Disabled Account
Joined 2003
when I was in college, I did some fluid dynamics simulations for a group that designed a low-noise fan that is similar to your but not quite.

it is essentially a rotating pipe, with fins on it. Think of your diagram. and imagine the horizonal lines are a negal pipe, except that has openings along its axel (?). the pipe is driven by two electric motors on its two extreme ends. Each motor also has a conventional fan in them blowing air towards inside of the pipe.

so air would go into the pipe from the two ends, and being thrown out throw the openings on the pipe.

It has extremely low noise.
 
There are lots of extrusion profiles designed for just that use. www.apexjr.com has some that are drilled for 8 TO-3 devices. My Hafler DH-500 has the fan inside the cabinet blowing out through the heat sink. this is more efficient than sucking the air through the heat sinks, and therefore a little quieter. at its low speeds, the airflow is inadible. There is a problem with mechanically transmitted noise. Use the largest diameter fan and heat sink arrangement you can to minimize the velocity for a given airflow. Of course, there is no free lunch - larger profiles cost more and lower velocity airflow doesn't increase the efficiency of the heatsink as much as higher for a given surface area.

If I were designing a chassis of this sort, the fan would be mounted on the heatsink, blowing air out of the cabinet. The heat sink would be mechincally isolated from the chassis. Isolators should be designed for the mass of the object to be isolated and the desired isoalaiton frequency. You could probably get satisfactory results using an isolator simply rated for the weight/number of isolators. Manufacturer's websites may help you choose the right one. Lord (US), Barry (US) and Hutchinson (France) are three manufacturers that come to mind.

I'd also damp the case panels, to prevent radiated mechanical noise. using the same stuff you would use in a car or use MDF and veneer it. Use a very soft gasket between the heatsink exit and the case. You might want to try a flared exit, using one of the available speaker port flares. Use multiple computer inlet filters to minimize the dust intake.

Good luck and let us know how it works out. I plan something similar at least for the first cut of my Aleph-Xs.
 
PULL vs. PUSH

From modeling and simulation, the best cooling solution would have a large area inlet with a grating to start laminar flow, a taperring tunnel to the heat sinks and transistors, a slightly expanding tunnel to a large multi-speed fan that pulled out the heat. Heat and Mass Transfer....the largest intake air mass plus laminar air flow across each fin for the greatest contact

Putting the fan in front creates turbelant flow which reduces the amount of air that makes physical contact along each heatsink fin.

Ideally you would also want to take advantage of rising heat energy and have the air flow up. This can lead to dust problems when the amp is on the floor.
 
The things that worry me are noise and dust. I was considering watercooling, but for "only" 200w it beats the purpose if an air cooled system can do it just as well and (i beleive) as noiseless. Two fans at 9vdc should move more than enough air, and be reasonably quiet, even more wih damped entrances.
Dust is a necessary evil with fans - i've seen air conditioners that have a soft cloth on the air entrance, so probably it can be taken care of without restraining airflow too much. In any case, if it gets dusty a vaccum cleaner should clean it easily.

As for the mechanical vibration, i suffered them on my PC; it happens when fans get dusty, specially in the axis. I found that rubber washers (and cleaning the damn things ;) ) do the trick just fine. Those i will be using for sure.

Now that i think of it, are fans electrically noisy? I was planning to feed them more or less directly from the supply rails of the amp.

By the way, as i recall it the laminar/turbulent airflow is a tradeoff; laminar flow gives you better air flow but turbulent flow removes heat much better IIRC. In any case, the output of a fan is quite turbulent in itself...
 
Panaflo

The quietest fans seem to be the Panaflo-brand fans by Matsushita (Panasonic). When the L1A series fans (popular with PC silencers) are run at 5 volts, they are practically inaudible - you could leave them on indefinitely. Also, on the silentpcreview.com forums, there has been some talk about a 120mm Panaflo fan sold on eBay that was part of an OEM order for some company, and the fans are even quieter than the L1A 120mm for the same operating voltage, while seeming to move more air.

Some NMB (Nippon Micro Bearing, Minebea Electronics) fans are also fairly quiet when run at less than 12 volts. Dell uses them in the back of their minitower chassis.
 
I haven't tried Panaflo, which seems highly regarded by some
PC people. However, I have some of the Papst low noise fans,
and replaced the original fan in my PC PSU with one of these.
What I did find out, however, was that the Papst motor has
1/4 the power of the original fan, and running the original on
on 6V, which should give about the same power seemed to
make that one about as silent as the Papst. Besides, in the
case of Papst I had to return severl of them because of some
clicking or ticking noise, most certainly caused by bad
manufacturing quality (and I have heard several other having
the same problem). And I always thought Papst were supposed
to be high quality!!! Lesson, runing a cheap fan on low voltage
may be just as effective and silent as a special silent fan, at a
lower price.
 
EBM/Papst used to be made in Germany, but then they started moving their manufacturing to China, and now their products suck. I suppose they will soon wonder why no one keeps paying the $25 for one of their fans like they used to, and it is because their manufacturing has turned to crap. Panaflo is very good, however. I use those in my own computer, and at 5v they are silent.
 
BAM said:
EBM/Papst used to be made in Germany, but then they started moving their manufacturing to China, and now their products suck. I suppose they will soon wonder why no one keeps paying the $25 for one of their fans like they used to, and it is because their manufacturing has turned to crap. Panaflo is very good, however. I use those in my own computer, and at 5v they are silent.

Yes, Papst used to have a very high reputation for reliability
at least.

Actually, I think the silent models I bought were made in Hungary.

Another interesting thing with "silent" fans is that they often
use sleeve bearings (I think that is the english term) rather
than ball bearings. However, I understand that they wear
out much quicker so after a while they are as noisy as a
ball-bearing fan and continues getting noisier.
 
Christer said:
Another interesting thing with "silent" fans is that they often
use sleeve bearings (I think that is the english term) rather
than ball bearings. However, I understand that they wear
out much quicker so after a while they are as noisy as a
ball-bearing fan and continues getting noisier.
If you peel the label off the back and then remove the rubber plug underneath it, the bearing and shaft are exposed. Just put 1 drop of oil in here once a year and it will last just about forever. Put some adhesive tape over the plug where the label used to be.

IME, AC fans are quieter than brushless DC ones. The DC ones have a torsional vibration from the switching of the current to the stator. AC fans are much better in this regard.

To reduce the speed of an AC fan, IIRC AKSA suggested using a fluorescent tube ballast inductor in series.

Both types of fans are a possible source of radiated (pulsating) magnetic field that can get into low level circuitry and cause noise problems. The DC ones, because of their switched current drain can induce noise into other stuff connected to the same supply rail. Use electro+film bypass caps close to the fan.
 
Originally posted by Circlotron Both types of fans are a possible source of radiated (pulsating) magnetic field that can get into low level circuitry and cause noise problems. The DC ones, because of their switched current drain can induce noise into other stuff connected to the same supply rail. Use electro+film bypass caps close to the fan.

That's what i feared - i'll have to experiment with that. I have to experiment there, but i imagine a pair of caps would be more than enough.
 
Disabled Account
Joined 2003
Circlotron said:
Both types of fans are a possible source of radiated (pulsating) magnetic field that can get into low level circuitry and cause noise problems.

that's correct in theory. EM interference for DC motors can be largely cured with brush receipies and commutator designs. and shouldn't be that big of a problem if you are talking about line-level signals.

I'm running two fan'd heatsinks now, one for a JLH1969 and another is a Citation 12. plugging and unplugging the fan socket make zero (perceived) difference on the sound or noise.
 
Electrical noise:
Noise from a fan can easily find its way into other circuitry. So perhaps it would be best to keep it on its own power supply, afterall.. a fan only draws a measily about of current.

Dust:
As for dust if you are using heatsinks with alot of height you can easily set up a duct to enclose the airflow of the fan.

Noise:
Mounting the fan on rubber stoppers, washers or boards will reduce the amount of low end noise emitted by the fan, making sure there are little restrictions in the airflows path will keep the noise down too, polished surfaces will cause less of a restriction in airflow also. Another way you can drop the noise level is by reducing the voltage obviously... Any decent fan (Panaflo, Pabst, maybe sunon..) will be quieter, and will spin at a lower voltage.
Cheaper fans are going to neede even the full whack of 12v(120v/240v) to spin, and then bring its voltage down to something quiet..

Types:

Well, theres 2 main types.. axials and blowers... axial fan will be quieter. they weigh less and there are less moving parts..
But... there is a new type of fan that is alot quieter.. its the TMD fan where the coil windings for the motor are placed in the chassis, and not the impeller.. meaning the impeller assembly is going to weigh less :)

tmdnolid500.jpg



As for size... the smaller you go the more noise you are producing for the same CFM, but at higher pressure. 80mm/120mm is usualy the best

Also as to the original post 2 fans in series with each other is going to produce ALOT more noise.
 
The idea was to use two small fans rotating slowly, but i imagine a slightly larger one would do just as well. 80mm was my original bet.

By the way, how much can i expect the °C/W rating to drop with a fan blowing air over it? I know it depends on a millon things, but then again, i gotta know how much of a surface area i'll be needing.
I think someone mentioned it improved the rating by a magintude of 4 to 6, which would be a gift to me :)
 
Lisandro_P said:
By the way, how much can i expect the °C/W rating to drop with a fan blowing air over it?

Check the manufacturers' web pages. for example an AAVID flatback in a size appropriate for class A amps http://www.aavidthermalloy.com/bin/...=10.000&LengthUnits=in&SButton1=Change+Length natural convection is about .5 degrees C/ Watt, drops to .205 at 200 feet per minute airflow, .08 at 1000 feet per minute. Not that you would want that kind of airflow in your listening room.
 
Lisandro_P said:
I think someone mentioned it improved the rating by a magnitude of 4 to 6, which would be a gift to me :)
Yep. I posted the results of some experiments a while back (can't quite find it at the moment) and IIRC with the fan running at 6 volts it improves 5:1, at 15 volts :devilr: about 6:1. A 100mm square heatsink could dump 100 watts at about 40 deg rise.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.