Did anybody get one of these working yet?
I think I'll give it a shot. I have a 1/2 hp motor that spins at 1725 rpm. It has a reduction on it, which brings the speed down to about 2 rps (120 rpm). 120 is a bit too slow, so it would have to run at 1725. I've been looking at RC helicopter swash plates for tilting the blades. I haven't seen one in person, but from online pictures, it looks like it would be the ideal size for this application. I'll have to look at one in person, and maybe pick one up. I also have an old 12 inch subwoofer with a deteriorated surround, so I can use the magenet/VC from that to move the swash plate. I was thinking put the VC in front of the fan instead of between the fan and the motor for simplicity. Now onto the hard part--the center piece to hold the blades and the linkage. A friend of mine has a lathe and a mill, so I would be able to machine whatever pieces I need. What kind of bearings would I need to allow the blades to rotate effortlessly while withstanding quite a bit of radial force?
I'm just throwing ideas out there for now. I'll have to look at the swash plate to see if it can be used. Any comments/suggestions?
I think I'll give it a shot. I have a 1/2 hp motor that spins at 1725 rpm. It has a reduction on it, which brings the speed down to about 2 rps (120 rpm). 120 is a bit too slow, so it would have to run at 1725. I've been looking at RC helicopter swash plates for tilting the blades. I haven't seen one in person, but from online pictures, it looks like it would be the ideal size for this application. I'll have to look at one in person, and maybe pick one up. I also have an old 12 inch subwoofer with a deteriorated surround, so I can use the magenet/VC from that to move the swash plate. I was thinking put the VC in front of the fan instead of between the fan and the motor for simplicity. Now onto the hard part--the center piece to hold the blades and the linkage. A friend of mine has a lathe and a mill, so I would be able to machine whatever pieces I need. What kind of bearings would I need to allow the blades to rotate effortlessly while withstanding quite a bit of radial force?
I'm just throwing ideas out there for now. I'll have to look at the swash plate to see if it can be used. Any comments/suggestions?
For the blades I'd be inclined to use the sort of bearings used in in-line skates. They are cheap, easily available, and they run smoothly. You could also use the smaller sort, like for RC cars. You could calculate the maximum force the bearings will need to handle, but it isn't going to be very much. Depending on how large you make the blades you shouldn't have to deal with more than a few pound at the bearings.
I have a functioning four bladed fan assembly based on a small 5W driver, but I have yet to mount it on a motor. I forgot to drill the back plate before mounting the blades 🙄 I'll have to be creative before mounting it. Also, since I'm spinning the entire motor I'll need to either fabricate some slip rings for the signal.
I have a functioning four bladed fan assembly based on a small 5W driver, but I have yet to mount it on a motor. I forgot to drill the back plate before mounting the blades 🙄 I'll have to be creative before mounting it. Also, since I'm spinning the entire motor I'll need to either fabricate some slip rings for the signal.
need pictures
Requesting photos--even at this early stage, CodeS.
I'd love to see the arrangment you have of a functioning four bladed fan assembly. Pics will totally help with this.
Also, I am having a hard time picturing what you mean by "spinning the entire motor" by your description. Are you referring to spinning the voice coil and needing some brushes and concentric contacts to get the signal to it? Not spinning the main fan motor, right?
If this is the case, I think your design is better than the alternative of a spinning/stationary linkage, like the swash plate on a helicopter, as the slip rings will possibly induce a bit of high frequency noise, but this will tend to be filtered out by the large mass of the blades. The swashplate type will tend to generate more low frequency noises.
Photos! (please)
[finally someone is really making one of these... leave a trail of bread crumbs for the rest of us]
CodeSuidae said:
Requesting photos--even at this early stage, CodeS.
I'd love to see the arrangment you have of a functioning four bladed fan assembly. Pics will totally help with this.
Also, I am having a hard time picturing what you mean by "spinning the entire motor" by your description. Are you referring to spinning the voice coil and needing some brushes and concentric contacts to get the signal to it? Not spinning the main fan motor, right?
If this is the case, I think your design is better than the alternative of a spinning/stationary linkage, like the swash plate on a helicopter, as the slip rings will possibly induce a bit of high frequency noise, but this will tend to be filtered out by the large mass of the blades. The swashplate type will tend to generate more low frequency noises.
Photos! (please)
[finally someone is really making one of these... leave a trail of bread crumbs for the rest of us]
xplod1236 said:
There are two different radial force components here, and one may be able to be completely eliminated. Maybe even both.
Basically, a symmetrical flat fan blade can receive an equal force from the front and back. There will be a twisting force pushing the blade back to a neutral position (towards no airflow), but this would not present a load on a bearing really.
There is a sort of propeller force pushing against all of the blade stems, and this is at a maximum when the blades are at max pitch (in either direction). This force is the same as being placed on fixed blades, but is generally not that strong. Imagine the fan trying to lift the motor, helicopter style. It likely is not strong enough to lift the motor, but if you can lift it by the blades (using your hands, with the motor off please) without overstressing the bearings, then they are strong enough.
Re: need pictures
Ok, I took some pictures and uploaded to a Picasa album. Be sure to read the comments.
What I've got (or rather, had; I had to take it all apart to drill the back plate) is not very practical for actual operation. The center post has to be long in order to give a good angle for the push rod to operate the axle. I've got a couple of different designs I'd like to try for this, I'd like to use a rack and pinion drive (with nylon anti-backlash gears if I can find some) instead.
See the pictures and the video I posted, it'll be clear. Yes, on the fan motors output shaft I will need a couple of copper contact rings that spin with the shaft. These would be connected to the speaker inputs to drive the voice coil. Brushes would ride on the rings to transfer in the signal.
Another option would be to snag the rotary transformer from the head assembly in a VCR. Through this you could couple the drive signal with no sliding contacts. I'm not sure how much current those transformers can handle, probably not much at low frequency. For best operation you'd want to couple power in at high frequency, say 400Hz, along with a low-level audio signal, then use a small amp on the rotating parts to drive the speaker. That's a whole project in itself 🙂
I agree. A simple first-order low pass filter will take care of any brush noise. With a mechanical swash plate setting operating at audio frequencies there just isn't much you could do to remove any noise. In a well-engineered assembly I wouldn't expect much noise, but the better the engineering is the more expensive it'll be.
Well, as you can see, this is a very rough prototype. At this stage I'm just trying to check what ideas might work and which are just stupid 🙂 It's already clear that the blades need to be very light, once the cone gets them going it needs to be able to stop them again. Any extra weight will make the suspension resonance worse.
I'm not sure if I'll bother reassembling all four blades, to make it work I'd need to sink some more effort into improving the push rod mechanism. I'd rather move on to a more compact and linear design.
An advanced version of this thing would probably benefit from a voice coil position feedback circuit to provide better position control. Along this line, it's probably also quite reasonable to use hard-disk head positioning motors to rotate the blades. It might be necessary to rewind the voice coil for the necessary power, but the force available and the speed of operation are both very high. Each blade could be directly driven by one coil and magnet set. This obviously would require either a mechanical centering mechanism or a feedback system for positioning. Again, a project in it's own right.
neededandwanted said:
Ok, I took some pictures and uploaded to a Picasa album. Be sure to read the comments.
What I've got (or rather, had; I had to take it all apart to drill the back plate) is not very practical for actual operation. The center post has to be long in order to give a good angle for the push rod to operate the axle. I've got a couple of different designs I'd like to try for this, I'd like to use a rack and pinion drive (with nylon anti-backlash gears if I can find some) instead.
See the pictures and the video I posted, it'll be clear. Yes, on the fan motors output shaft I will need a couple of copper contact rings that spin with the shaft. These would be connected to the speaker inputs to drive the voice coil. Brushes would ride on the rings to transfer in the signal.
Another option would be to snag the rotary transformer from the head assembly in a VCR. Through this you could couple the drive signal with no sliding contacts. I'm not sure how much current those transformers can handle, probably not much at low frequency. For best operation you'd want to couple power in at high frequency, say 400Hz, along with a low-level audio signal, then use a small amp on the rotating parts to drive the speaker. That's a whole project in itself 🙂
I agree. A simple first-order low pass filter will take care of any brush noise. With a mechanical swash plate setting operating at audio frequencies there just isn't much you could do to remove any noise. In a well-engineered assembly I wouldn't expect much noise, but the better the engineering is the more expensive it'll be.
Well, as you can see, this is a very rough prototype. At this stage I'm just trying to check what ideas might work and which are just stupid 🙂 It's already clear that the blades need to be very light, once the cone gets them going it needs to be able to stop them again. Any extra weight will make the suspension resonance worse.
I'm not sure if I'll bother reassembling all four blades, to make it work I'd need to sink some more effort into improving the push rod mechanism. I'd rather move on to a more compact and linear design.
An advanced version of this thing would probably benefit from a voice coil position feedback circuit to provide better position control. Along this line, it's probably also quite reasonable to use hard-disk head positioning motors to rotate the blades. It might be necessary to rewind the voice coil for the necessary power, but the force available and the speed of operation are both very high. Each blade could be directly driven by one coil and magnet set. This obviously would require either a mechanical centering mechanism or a feedback system for positioning. Again, a project in it's own right.
I don't have a prototype yet, but I did (partially) draw up some ideas. The speaker magnet/VC will sit around the shaft (stationary). The swash plate will ride on the shaft in between the VC and the blades. I'm planning to use rulon bearings to reduce rotating mass on the fan blades. I will also use a rulon bearing between the swash plate and the shaft to align and support the VC. The bearing will be fixed to the shaft, and the swash plate will slide over the bearing. This will once again reduce the moving mass. I will be using the motor from a 12 inch woofer. I will post more drawings as I get them done.
Congratulations! Of the hundreds posts on this thread, you are the first to post pictures of an actual "work in progress". You deserve a pat on the back!
xplod1236 said:
I look forward to seeing more of the design. Will the connection of the voice coil to the swash plate provide support to keep the voice coil moving linearly? As I understand it the speaker surround provides a significant amount of the support needed to control the movement of the voice coil.
Hara said:
What, you didn't like the video I posted?
I'm not sure I understand what you mean by the voice coil moving linearly, but I'll try to answer. The swash plate will keep the VC aligned (as the surround does), so it won't be able to wobble and scrape within the gap. So I will have the spider and the swash plate controlling the sideways movement of the VC (spider and surround on a speaker). The in and out movement will only be controlled by the spider. I should have a drawing of the whole assembly by tomorrow, and hopefully I'll have it built within a week or two.
To eliminate the play in your design, look into using RC car tie rod ends. This is what I'm using for my design. I picked up a 6-pack for $4 The overall length is just under 3/4". They're plastic and fairly lightweight, and if they don't break on RC cars, I don't see why they would break on a fan.
To eliminate the play in your design, look into using RC car tie rod ends. This is what I'm using for my design. I picked up a 6-pack for $4 The overall length is just under 3/4". They're plastic and fairly lightweight, and if they don't break on RC cars, I don't see why they would break on a fan.
xplod1236 said:
It sounds like you understood exactly what he was asking here. (As I understood it...) Basically the cone+spider system keeps the voice coil from leaning sideways (or exhibiting greater excursion on one side than on the other, etc.)
In my experience, the spider tends to do nearly all of this correction, and drivers with cones removed completely do not tend to have their voice coils "wobble" when driven.
Removing only the surround, but leaving the cone in place can cause problems due to the uneven distribution of mass of the cone.
We can always add a second spider if necessary, or even build a push-pull version where two drivers are cannibalized and two voice coils oppose each other to push and pull on the center rod in a complementary fashion. This would buy us a lot of stability (and an increase in power handling capacity).
xplod1236 said:
Yes, precisely what I meant. There shouldn't be any side loading to twist the VC against the spider, but I don't know if maybe any asymmetry in the magnetic fields might cause problems at higher power if the support from the surround is removed. Probably not, but either way it won't be a problem if the swash plate is restraining movement to a single axis.
I was thinking about those too, I agree they should be more than adequate to the task, considering the shock loading they endure in their designed use. Alas, the only set I've got are currently serving their intended purpose 😀 They are manufactured to fairly close tolerances and run pretty smoothly, so I don't think they'll generate any noise. In my prototype the only concern would be the friction they introduce, they can tend to be pretty stiff when new, my little driver will have trouble making them move. Maybe I can spin the ball with a drill to free them up a bit.
I did some research into blade design for reversible fans. Pitch is mostly determined by the blade size and power available. A pitch of 40 degrees is about the maximum, it turns a lot of the power into turbulence. Longer blades increase the power required from the fan at a given pitch, in a non-linear relationship (the longer the blade, the faster the outer rim moves, so the more resistance it encounters).
Some designs use a symmetrically curved blade rather than a flat blade. Evidently this is more efficient, but I suspect that it may affect the force required to hold a given pitch. I don't know whether it would increase or decrease that force though. That might be something to investigate later to improve the efficiency of the device.
Normally the input to a speaker serves only to move the cone away from center, with the power of the signal working against the suspension which works to center the VC. In a normal speaker the cone has no force working to hold it out, in fact the air pressure does the opposite, making it increasingly more difficult to increase the excursion of the cone.
For a fan blade, if the aerodynamic force on the blades is such that it tends to make the blade increase in pitch rather than decrease, then the blades may not return to center very well (or at all) until the input signal begins to drive the VC in the opposite direction. The result would be a tendency for the blades to stay pitched up, giving poor accuracy. The blade design, therefore, should include some bias to zero pitch such that a zero input signal will cause the blade pitch to immediately return to zero. A cupped design as describe above might contribute to this behavior.
The VC from the driver that I'm using scrapes even with very little power input. The speaker had a deteriorated surround and probably sat for a long time (before I acquired it) with the cone twisting the VC out of alignment. So maybe now the spider is twisted a bit. The swash plate should straighten that out though.
The ones that I picked up came in 2 pieces. I had to snap the hollow ball in place. The ball moves around so freely, I thought there would be play, but there's none to be found. But if yours are a bit tight, spinning them to loosen them up should do the trick.
The ones that I picked up came in 2 pieces. I had to snap the hollow ball in place. The ball moves around so freely, I thought there would be play, but there's none to be found. But if yours are a bit tight, spinning them to loosen them up should do the trick.
CodeSuidae said:
I look forward to seeing more of the design. Will the connection of the voice coil to the swash plate provide support to keep the voice coil moving linearly? As I understand it the speaker surround provides a significant amount of the support needed to control the movement of the voice coil.
What, you didn't like the video I posted?
Ok. Well, you two posted one after the other I was kind of referring to both of you. And he was first to post a picture ON the forum. Anyways keep up the good work!
Here's an update for my project. The shaft will rotate the whole assembly. The swiss-cheesed swash plate will rotate with the housing (as should be obvious). The flanged bearing (black) combined with a thrust bearing (tan) and the sleeve (cut away for clarity) will make up the rest of the swash plate. The bearing and sleeve will be stationary (not spinning but free to move up and down the shaft) while the rest spins around it. The sleeve will be connected to the voice coil. In my opinion, this should work, but I'm always open to criticism and suggestions.
CodeSuidae said:
It's pretty easy to make a blade self-centering, so there is a restoring force to drive the blade back to flat (zero pitch - no push or pull). If it is symmetrical about the axis of rotation, if can actually be at rest at any pitch. Basically one edge is catching air and trying to force the blade to go to a higher pitch,and the other edge is being pushed back to center.
Remember the "my hand is an airplane wing" experiments with the car window? You may feel a strong lift (up or down) but almost never feel ay torque trying to twist your hand to a higher or lower angle of attack.
If the majority of the blade is being dragged behind the axis of rotation, it will tend to be self-centering. This is like driving a car forward and letting go of the steering wheel. It tends to straighten the wheels.
If the majority of the blade is before the axis or rotation, it will tend to flap open. The force it towards a higher pitch (in either direction). This is like driving a car in reverse - it tends to exaggerate the turn. Let go of the steering wheel and you will end up going in a small circle backwards.
If a fan has blades that are driven towards lower pitch, and then the fan is spun backwards, the blades will tend toward higher pitch (and vice-versa)
I believe that most blades are designed for no net torque. for a variable pitch blade this becoes more important, as the suspension is not rigid.
Likely, a voice coil has enough restoring force from the spider to settle back to center and take the blade back with it. This will also depend on the friction being overcome to turn the blades (at rest even).
If a blade is designed to balance on its axis of rotation and have no restoring force or exaggerating force, it will work the same at any fan speed. Highly recommended to strive for this.
As far as blade geometry, it gets a bit counter-intuitive. first fnd a hobby shop and tell them you want to see a helicopter blade fr a helicopter that can fly upside down. They will show you the "flat" blade. This can be pitched to have positive or negative lift, and has a relatively flat cross section.
One last bit on blade geometry:
The tips of the fan are travelling much faster than the part closer to the center. Any given pitch angle will tend make the tips push air much faster than the inner parts. This may not be desireable and may cause problems, either in sound, or just making the blades flap.
Propellers attempt to acheive uniform lift along the blade length by having a much higher pitch towards the center and almost zero pitch toward the blade tips. We cannot do this with a variable-pitch blade.
One method of handling (here's the counter-intuitive part) is to taper the blades from very wide near the center to very skinny near the tips. This provides more lift near the center and less at the tips (which balances the huge difference in linear speed).
Hard to describe in words. Here's a picture of a common variable-pitch blade:
http://www.towertechinc.com/service/images/fan_1.gif
An externally hosted image should be here but it was not working when we last tested it.
See? The tips have less surface area to scoop less air at a given speed (but they are travelling much faster) than the blade area closer to the center. The blades also stop short of the center and the fan has a larger hub area. This keeps the blades out of each other's way when varyingthe pitch thru the center, but also to avoid reverse air flow (leakage) thru the empty space near the center.
Here's a huge one, which has variable pitch, but not enough to completely reverse the direction of airflow like we need:
big variable pitch fan showing large hub
xplod1236 said:
It's totally OK to think outside the Swash Plate model, by the way. In a helicopter, it is necessary for banking left and right and for tilting forward-downward and forward-upward.
What a helicopter needs to do that we don't need to do:
A helicopter needs to make the blades have different pitches at the same time. If you want to lean to the left, the swash plate tilts and the blades rotate to a greater angle of attack on the right and then lower on the left. So each blade bites more air on the right, then flattens as it comes around to the left. In this way, more lift is generated on the right side than on the left consistantly, and the the copter leans to the left.
We don't need this. We need all blades to advance and retreat together. We should never need to have the fan blowing on one side and sucking on the other.
This doesn't invalidate the swashplate concept at all; it's just that we don't need all of the flexibility of a helicopter mechanism, so don't need to stick with that model completely.
If the linkage from voice coil to blade twisting starts to get heavy, we will get reduced performance. You don't want to use all of your power wiggling masses that don't contribute to the sound.
If the moving mass of the linkage starts to get heavy, why not just put a voice coil at the base of each blade? Coil goes up, pushes a lever and twists one blade (times 4). The four magnets will not be moving masses, just the voice coils.
CodeSui, your picture reminded me of this. Actually one of my earliest ideas was simply to put a servo motor on the axis of each blade (where your bearings are) and then send these the twisting signals. No voice coils and huge magnets, just 4 small servomotors--one on each blade. That's what your picture reminded me of.
Just to clarify: I'm calling it a swash plate, but it will only move back and forth. The bearing it's riding on will not allow it to tilt like a helicopter swash plate does; so all the blades will tilt together. Is there a better name for it?
neededandwanted said:
I considered that early on, but decided against it, at least until I decided that a voice coil version isn't workable. I like the simplicity of driving it with a simple amp instead of a servo controller. To run a servo controller I'd need to convert the input signal into movement commands for the servo controller. I don't know of any off the shelf equipment to do that.
xplod1236 said:
In your design it appears that the arms attaching the rotating part of the swash plate to the arm on the blade is a ridged X shape. If so, how does that part of the swash plate track the rotation of the arm on the blade?
It's not rigid. The arms on the blades have hollow balls where they connect to the swash plate. I just didn't take the time to draw it up.
In case nobody has posted this yet, there's an informative thread at AVSForum in which the maker participated
http://www.avsforum.com/avs-vb/showthread.php?t=786803
http://www.avsforum.com/avs-vb/showthread.php?t=786803
- Status
- Not open for further replies.