I'm building an Arduino controlled busker organ. The organ needs a reservoire of air with constant pressure. I'm reusing lots of old stuff. I have a pump for inflating small boats, which should be effective enough. But the air flow from the pump is not constant. And the air consumption of the organ is not constant. Therefore I have this design:
I use a bellows from an accordeon. I pump air into it. I let air go out to the pipes (the air goes to a wind chest, a chamber where the foot of the pipes comes through the ceiling. Each pipe end has its own servo valve.)
This bellows should have a constant pressure. Unfortunately the bellows acts as a spring. You need a bit force to extend the bellows, and there's a recognisable spring constant. That means that without a weigh on the top, the bellows would still produce a pressure, the higher the bellows extends, the higher pressure. My hope is that with a counteracting spring I could minimize the spring effect. I'd place a spring inside the bellows. Without a weigh on the top, the bellows would find an equilibrium. Extending the bellows from that equilibrium would need some force, pressing down the bellows would likewise need some force. After that, when I place a weight on top of everything (the red brick in the picture), I could add the pressure I need for the pipes to sound correct. The added pressure would be constant. But the spring effect of the bellows and the added counterspring would still be there. When pumping air into the bellows untill it reaches the no weigh equilibrium point would mean the bellows teoretically contains the pressure of the added weight. Pumping further would increase the pressure, using air and having the bellows go below the point will decrease the pressure. My aim is to minimize the difference. For the bellows to be of any use, I want the added weight to be the major part of the produced pressure and the spring effect to be as small as possible. And my big question is (besides is there any other flaws in my thinking) will the counterspring help at all? The spring effect of the accordeon bellows really is remarkable (I might be able to do something to it, it's from an 80 year old trashed accordeon), it extends some 30 cm until the force is way too big, and the equilibrium is just som 3 cm before fully pressed. My hope is that if I use a counterspring and get the equilibrium at say 15 cm, I could have a range from 10 cm to 20 cm, where the bellows would provide a pressure suitable for the pipes. The magnitude of the pressure would be determined by the weight and the variation in pressure between 10 cm and 30 cm would be as small as possible relative to the total mean pressure. My intuition says that the counterspring would minimize the ratio of variation / mean pressure. Am I wrong?
You may say it's just for me to test and find out. And I will. The whole project is being designed as I go on building it. This bellows part I haven't figured out yet, I just hope to find flaws in my thinking before I make crucial decisions.
Or use the air pressure tank from a water well system. I see then advertised on Craigs list, sometimes for free. Most are still usable, but have been replaced as a precaution. Some have steel tanks, some fiberglass, but ALL have a rubber air bladder.
Johan,
Are you still considering this project? Going for a steam-punk look or something like that?
There are a ton of ways to do what you want depending on budget and the look you are going for. I have some experience in this area and am guessing you want something unique and cheap. The counteracting spring is a good idea except that the bellows is probably a progressive spring and that might be hard to do. Something utilizing the concept of the old Nautalis gym equipment could work. Let me know if you want some more info. It would be mechanical, self regulating, easy, and cheap.
Yes, still working on it. Not very steam punk. Just an old cupboard I bought as second hand. I might coat the PVC pipes with tin foil.
The bellows is working as a spring. I try my best to soften it, by stretching to extreme and maybe infusing with some leather softener, as long as there are leather parts that can be softened.
I have no idea yet. I need to adjust the pipes to get the best sound. If the pressure is too low, I can add weight on the bellows or I can widen the opening at the pipe foot. I could measure the air pressure and find out flow rates etc, but that would add a lot of complexity to the project. It's a shortcut to just build everything and then adjusting the pipes to what I have. But before that, I need to learn more about springs.
I used to play a harmonium - it uses a foot pumped bellows to provide air.
The bellows was very soft and weighted so the idea is sound.
However a spring internal to the bellows may not be suitable, as it would be working over too wide a range of movement. (see fig)
You may need to consider a pulley and clock spring mechanism to suit the range of movement.
I love the harmonium. I have two, one at home and one at job, which is mechanized with 61 servos. The harmonium has free brass reads like a harmonica or accordeon, which produce the sound, and they are not that dependant on constant pressure. I'm afraid organ pipes are more prone to lose the correct sound and pitch due to wrong pressure.
Johan, This sounds like a fun project. I have attached a crude concept sketch that should actually do what you want. (I am new to the forum and am learning the mechanics of posting.)
If you look at the sketch of your bellows there are some pressure and height numbers that I guessed at. Lets say that 1 psi of air will extend the bellows half way. But the pressure needs to increase to 8psi to get full extension. In the little graph of psi (vert axis) to extension, you can see that I think the force to extend the bellows will increase exponentially. That indicates a progressive spring. A normal linear spring would show a straight line in the graph. Finding a progressive spring to fit inside your bellows to give a neutral resistance throughout the entire travel might not be possible, plus there would be a bucking problem because of the length.
Instead, I propose a wooden wheel whose diameter is about 2 times the travel of the bellows travel. The wheel is fixed above the bellows with a rope attached to, and wrapped around, the wheel. The other end of the rope is attached to the top of the bellows. The wheel should be sized so that as the wheel is rotated thru approximately 1/5 rotation (72 degrees) the rope will pull the bellows through its full range of travel. There is a stick with a weight on the end of it attached to the wheel such that when the bellows is all the way closed, the stick is maybe 15 degrees past vertical. If you push down on the stick so that the rope pulls the bellows up, the effective radius of the weight increases, giving the weight more leverage so that it pulls harder and harder on the bellows.
You will need to experiment with wheel diameter, stick length and angle,and weight, to get a smooth extension of the bellows. When correctly balanced, just pushing on the stick with your finger should easily extend the bellows.
Then as you said, putting a brick or weight on the top of the bellows will establish the actual operating air pressure.
Making one that works will be easy. Making one that works well, will not.
That's a very good idea, lloyd-ss! I will treat the bellows to lessen the spring effect (heat, oil, stretching), and if I can't achieve a setup with an even sound from the pipes through the whole bellows travel, I will definitely try this method. The wheel could be just a quadrant. It could be elliptical or whatever strange spiral, as long as its pull would at each point be equal to the spring force of the bellows.
The reed organ is an ongoing project. Right now I need to make it play one tune independent of any computer. To fit the whole tune in a microcontroller board (Leonardo) was tricky. I have to place the tune in PROGMEM. But that's another story.
The rotating wheel will be an interesting thing to add to the bellows. Too bad it needs the space that I was hoping to use for the air pump and its crank. I'm running out of space in my cupboard.
Johan, I think the bellows counterweight mechanism can be kept fairly compact, and would function something like a weight in a grandfather clock. If the bellows were in the left side of the cupboard and collapsed it was 15 cm tall, and fully extended it was 70cm, the rope from the top of the bellows would go through a pulley directly above it, and across to a variable diameter drum (roller) at the top-right of the cupboard. The drum would be on a fixed axis of rotation but would have 2 cables wrapped around it. The "power" end of the drum would be a constant 9cm in diameter and one of the cables would be attached to it and wrapped 5 times around the drum. A weight at the end of the cable would apply the torque to the drum to rotate it. A second rope would be attached to a variable-diameter section of the drum and would go thru 2 pulleys on the ceiling of the cupboard to transfer the force from the drum to the bellows. The variable diameter portion would have a spiral groove cut in it for the rope such that when the bellows was fully collapsed, the rope would be pulling from a 5cm radius. As the bellows became more extended and needed more force to pull it open, the radius of the groove for the rope would become smaller and smaller to maybe a 2cm radius, so that the pull-force would progressively increase. That should still leave some space for the pump and crank. What do you think?
I do have components for setting up a wheatstone thing, but that probably is overkill. But I do have these barometric sensors, too, which give digital values in Pascal, I think. And I do have a few of that kind of pressure sensors you mentioned. I never thought they would react to air pressure, too.
Overkill or not, it's overfun, and that's what counts.
Do you really need to know what the pressure and distance are? Be careful not to overthink this (I do that too, LOL) I think just a simple mock up where you can vary the leverage on the bellows will get you all the answers in an afternoon of experimentation. Simple trial and error.
But if you really do want to read the pressure, consider one of these Flexiforce thin pressure sensors by Tekscan. They are cheap, thin, and you could slip the end of it into your bellows and seal it back up with very little effort. It is really just a variable resistor (the circle responds to being squeezed), but also it responds to air pressure if there is a pressure differential to be detected. There is a tiny air passage in the film from the round circle out to the end where the 3 pins are. That allows it to sense a pressure differential. An RC time function on the Arduino will sample the RC discharge time at a fast rate and return a time constant that can be converted (calibrated ) to give you the pressure. I have used these and they are much simpler than a wheatstone bridge type of transducer where you might need an instrument amp and then do an AD conversion. Just a thought. I can tell you like problem solving and take an idea and run with it. That is refreshing to see. BTW, on the Flexiforce sensor, only the 2 outer pins have connection. the middle pin is just for mechanical stability.
Lloyd
Johan,
My apologies if I appear to be pushy or rude. It is really just enthusiasm for an interesting project. This sort of problem solving with simple solutions is what I did for 30 years and I loved every minute of it! Thank you for your patience, but let me know if I am getting to be a bother.
Lloyd
Look for a CPAP machine, you should be able to get an old one for not much money. There volume changes to maintain a constant pressure. They also move a fair amount of air.
