2 Prop project needs help with magnets.
Wurlitzer
Posts: 237
I am seeking assistance from those on this forum who may have extensive experience in projects using both electro and permanent magnets.
This project uses one Prop to scan approx 500 inputs and generate a set of serial messages to a PC and to a 2nd Prop which currently drives approx 1000 outputs which activate some small electromagnets which are part of a pneumatic primary and secondary valve assembly.
It may be best to just communicate directly with me rather than waste the forum's bandwidth on issues which are only tangential to the propeller.
I would like to eliminate most of the pneumatic valve assembly as it is prone to air leaks and not always opening and closing properly. If I were to simply use bigger electromagnets that would work for some of the outputs but others would require some very large magnets and of course some serious current which I would have to handle with my output devices.
I am attempting to use electromagnets to repel permanent magnets thus increasing the force between them while at the same time reducing the need for very large and expensive electromagnets to control some of the larger valves.
If anyone would be interested please let me know at craig.whitley@roadrunner.com
We can certainly keep the forum in the loop for information that may affect the prop but more than likely this will not be of interest to the majority of members.
Thanks
This project uses one Prop to scan approx 500 inputs and generate a set of serial messages to a PC and to a 2nd Prop which currently drives approx 1000 outputs which activate some small electromagnets which are part of a pneumatic primary and secondary valve assembly.
It may be best to just communicate directly with me rather than waste the forum's bandwidth on issues which are only tangential to the propeller.
I would like to eliminate most of the pneumatic valve assembly as it is prone to air leaks and not always opening and closing properly. If I were to simply use bigger electromagnets that would work for some of the outputs but others would require some very large magnets and of course some serious current which I would have to handle with my output devices.
I am attempting to use electromagnets to repel permanent magnets thus increasing the force between them while at the same time reducing the need for very large and expensive electromagnets to control some of the larger valves.
If anyone would be interested please let me know at craig.whitley@roadrunner.com
We can certainly keep the forum in the loop for information that may affect the prop but more than likely this will not be of interest to the majority of members.
Thanks
Comments
I'm intrigued that you're controlling so many outputs. Wow!
As for wasting forum bandwidth, I wouldn't worry about that. We learn a lot from each other on this forum as problems are solved - or not.
What exactly are you working on? Some kind of giant pipe organ? Some details might help us zero in on some suggestions for you.
Here are the stats:
Approx 980 pipes/drums/percussion all which use a very small 120 ohm electromagnet which attracts a 3/8" round armature when the pipe is to be played. When that armature moves, it opens an air way to atmosphere which in turn collapses a primary pneumatic (approx 1" x 1" covered in very thin leather).
The organ is under 10" of lift (water) or about .3 psi. the normal pipe valve might be 1-3 sq inches but there are larger valves with maybe 6 sq inches of surface area.
That primary pneumatic is attached to a pair of circular leather/felt covered valves which when pulled down by the collapsing primary pneumatic, allow a much larger secondary pneumatic to collapse and that action opens a valve which allows air to enter the pipe or a 3rd pneumatic which might beat a drum or other percussion.
Wicks makes a direct electric assembly which draws a lot of current and is really only practical for the smaller pipes in the organ. At higher pressures or very large pipes these coils become too large and would draw way too much current.
The other major downside to the Wicks arraignment is that when you apply current to the magnet the armature/valve assembly is as far away from the coil as it ever gets and of course you need a ton of current to pull that armature/valve assembly closer all while air pressure is trying to push the valve in the other direction.
My thought was to use a reasonable size coil to repel a magnet. The plus side would be that when energized the magnet/valve assembly would be touching the electromagnet thus the repelling force will be at its maximum and that is exactly what is needed. Once the valve starts to move away from the valve seat the air pressure differential begins to decrease.
Below is a very crude diagram of what I envision for prototype one minus all the control for the magnet to prevent it from just flipping over. Assuming I can link to the jpg.
I am also entertaining using a cylindrical magnet as the armature of an electromagnet again with the intent of repelling the cylindrical magnet when the coil is energized. That might be the best way to go as it would also captivate the magnet.
-Phil
If you intend to design electromagnetic actuators from scratch, be aware that all the different industries seem to have their very own terminology for the very same things, so sorting out all the terminology can get very confusing. I have recently started learning about magnetic design and it's one of the most confusing things I've ever had to deal with for that reason. The mere conventions for how the various B and H vectors point whether inside or outside materials is enough to drive me crazy sometimes.
However, there are some nifty looking software programs that might help you out. I can't vouch for any of them because I haven't yet started learning any, but maybe you can have a look:
http://www.infolytica.com/en/products/trial/magnet.aspx
http://www.quickfield.com/free_soft.htm
http://www.integratedsoft.com/products/magneto/default.aspx
In any case, there are a lot of bright people on this forum and your application sounds really cool, so don't be shy about providing as many details, drawings, photos, etc. as you can - or are willing.
ericball: Given that the Wicks corporation has for years used just the electromagnet with a leather & felt valve surface for years on smaller lower pressure organs gives me an indication that their concept works but again IMO when the forces holding the valves closed are at their maximum the armature is being away from the coil is at minimum force so they have to use a lot of current (and copper) to get these puppies to work.
Phil: Given there are just under a 1000 of these in my organ dictates that the per unit cost be reasonable. An RC servo would work but speed and cost would be an issue. It is amazing given all the moving parts involved for each and every pipe how fast these pneumatics actually operate.
I have looked at K&J Magnetics and they have some good charts depicting the fall off of force per distance. The magnets I had contemplated using had about 6lbs of pull (way more than needed) when they were very close to a steel object. By looking at this differently that being pushing rather than pulling I can generate maximum push force when these valves are close and have the maximum air pressure resisting their opening. IMO that is exactly what is needed to reduce the size and current requirements of the 1000 coils.
I'll work up a couple of crude crayola 101 drawings to help explain things further.
The Wicks valve is at the bottom right.
On the top, this design I think has a fighting chance with the magnetic repulsive force being maximum when the opposing forces are at maximum. This design uses a good portion of the original pipe valve without all the primary and secondary pneumatics and their associated valves. NOT shown is the valve return spring which in conjunction with the air pressure holds this pipe valve closed.
Well, is it possible to find off-the-shelf items that can do this? or are you determined to design and/or build them yourself from scratch? Because you'll be in the market to buy ~1000 of these, I'm sure you could get something custom made and get a good price for them, too. Anytime a customer gets serious about buying 1000 of anything, magic starts to happen. But, of course, an already made item would save you even more money. If you just want to quickly tinker with something to get a feel for things, there are solenoids at McMaster that run $10 to $20 each, which indicates to me you could probably buy them for ~$2 to $3 each if you bought ~1000 of them from the original manufacturer. There are push and pull versions in a wide variety of forces.
Maybe check this out just to get some ideas:
http://www.mcmaster.com/#solenoids/=a9zmks
If you haven't done so already, it might be helpful to do some kind of power budget for your organ so you can size what sort of power supply you will need at a given voltage. For example, if you're operating solenoids that require 24 VDC, you'll need to know what sort of wattage and current the power supply will need to handle. I'm guessing the organ would never activate all the electromagnets at the same time, and that there's some maximum number that would ever be in use at any given time. If you can determine what that maximum number is, then you can figure out what sort of total current you would ever require. I suppose that might depend a lot on what sort of music or whatever you're performing. Of course you could always assume that all 1000 electromagnets are in use at the same time and take it from there, but that might be wildly impractical and insanely expensive.
And crude crayolas are very welcome here. The more drawings the better.
How much are you allowed to change? I mean, you could in theory replace everything with a PC running emulation software, but it would not have the same character. Is the problem the leaks or the solenoids not opening or not closing?
If hissing air is a problem, then will the noise of solenoids opening and closing be a problem too?
Besides some cushions and bumpers of some sort, might some sort of control of the solenoid voltage help with that? Could a PWM technique be used to vary the current in the solenoid on the millisecond time scale to help prevent hard clacking, etc.?
-Phil
The 2 props have eliminated a pneumatic matrix relay and created all the MIDI features to record and playback anything played on the organ.
The propeller in the console also controls what is called the combination action which is a series of memory settings for the pistons which allow for rapid changes in multiple stop tabs via pushing the proper piston.
It also is now playing a PC based Virtual Theater Organ and or the pipe organ.
Even organs that are professionally maintained have issues with cyphers and notes not playing due to all the moving components and small air ways which tend to attract pieces of dirt blah blah blah.
Some of this is mental masterbation but I have enjoyed doing so much with the propeller.
This is my organ in my house so I can do just about anything I want given the limited depth of my wallet.
Regarding the solenoids making noise, like the Wicks Direct Electric action simply placing little felt bumpers in strategic locations make these effectively silent.
Thanks again for the links.
I may take your suggestion to use something off the shelf for a proof of concept but for the final application I will wind the coils myself so as to create the proper size unit for each size of pipe. As you can imagine the sizes vary quite a bit as the pipes range from 6 inches to 16 feet with a corresponding toe hole size to admit sufficient air.
Insanely expensive? Yup! Given a new organ of this size might go in the range of $500K and I bought it used and non-operational for $12K, I have a lot to play with.
At full organ, playing 5 notes in each hand and 2 in the pedals with my feet the number of pipes that might be playing simultaneously could be ~~200 to 300.
Have you checked the available magnetic materials available now? Super-magnets as used in the brushless motors, etc. They allow for much smaller and superior motors so perhaps they could assist your job too.
Pipe organs have a slower response than electronic even with the fastest actions as it still takes a few milliseconds to get the air flowing in the pipe and create the tone.
Using the Wicks solenoids as a guide, IMO I could at minimum reduce the current required (amp turns actually) by half if the armature had a strong LIKE pole to be repelled by the coils magnetic field. I believe the forces are additive although they quickly decrease with distance.
I am trying to put my hands around, how the 2 Propellers are coming in other then communicating serialy.
Are you doing this. When you push key for the "G flat", the first propeller serialy communicates with the PC and the second propeller activates the "G Flat" pipe?
From what, I see if this is what you are going to do then you will need 40 propeller boards, 20 of them setup for inputs only and the other 20 setup for outputs to drive your magnets and leaving some I/O pins for serial comms to the PC.
This will be easy to do if you are going to do this way. This is, what I call basic one for one networking and easy to do.
Scan 2 keyboards of 61 notes each with 2nd touch so that is 122 edit: No make that 244 for both KBs (soon to be analog) using two 32 bit serial shift register chips
Scan 1 pedal board of 35 notes with 2nd touch again soon to be analog using the same two 32 bit serial shift register chips.
Scan 96 stop tabs using three of the 32 bit shift registers.
Any change in the items listed above from last scan to new scan will generate a standard MIDI message which is sent to a PC running a Virtual Theater Organ program called MidiTzer (great program) and also standard MIDI messages up to the 2nd propeller in the pipe chamber.
In addition, the console Propeller scans a series of Piston switches and creates a memory of which stop tab are on or off for each piston so while playing hitting one piston can change the setting of all 96 stop tabs. Each of these stop tabs have 2 solenoids which again I have the propeller using the 32 bit shift register chips to drive so when you push a piston you see the physical stop tabs move.
The console Prop uses the upper 32K of EEPROM to store a handful of these memory settings and a serial communication to the PC provides for almost unlimited memory settings stored on the hard drive.
Up in the Pipe chamber the 2nd Propeller receives the MIDI messages and determines which pipes/percussions should be playing. Then that data is sent out 16 bits wide in serial form to 16 pipe driver boards each having three of those same 32 bit shift registers which also have open collector outputs. Those outputs currently control the very small electromagnets which start that chain of pneumatic valves in motion.
To incorporate higher current magnets, I will use MOSFETS controlled by the same 32 bit shift register boards.
It is amazing what can be done with this chip.
Awesome. I admire the sheer purity of your obsession. Well then, if you're going to roll your own, then it will be important to learn about magnetic materials. For example, I think the core of your solenoid should be a low carbon steel. I'm certainly no expert in this area, but maybe an alloy 1018 would provide good characteristics for the cost and availability. Alloy 1018 is fairly common stuff and it's easy to work with. As for the permanent magnet, you can get NIBs (Neodymium-Iron-Boron, NdFeB) magnets fairly cheap and they are incredibly strong. I've bought some stuff off this guy on ebay and he seems reasonable to me:
http://stores.ebay.com/Neodymium-Magnet-For-Less
Keep in mind that there will be not only electrodynamics to consider but also inertial effects and mechanical dynamics. You don't want the NIB that actuates your leather flapper to weigh too much, otherwise the inertia will be too large to get the fast response times you want. You'll also have to consider how such magnets are attached to your flappers. I suppose epoxy might work, but I don't really know.
As I (think I) understand your intended design, your solenoid will produce a magnetic field that will oppose the pole of the permanent (NIB) magnet that is mounted to move your flapper valve. But keep in mind that when the solenoid is not powered, then the NIB is going to be attracted to the steel core of the solenoid like crazy. If you try to circumvent that problem by not using a steel core, then I think you'll need a lot more current to produce a strong enough electromagnetic field. The steel core really makes a difference.
What sort of voltages are you running these at?
Wurlitzer, this is so cool. I hope you'll find it in your heart to favor us with a video concert at some point. I realize that it would be no match for the power and majesty of a live performance, but still...
-Phil
Also, not only will the NIB be attracted to the solenoid's iron core, but at very close range the NIB's powerful magnetic flux will flow into that iron core and convert the core into a temporary magnet. Thus, when the solenoid is not powered, the iron core will be (temporarily) magnetized and effect two magnetic poles in the iron. Under those circumstances, the solenoid pole facing the NIB will be of opposite polarity to the NIB's end that faces the solenoid. And then at the far end of the solenoid, the iron core will have its other pole. The irony here (no pun intended) is that the magnetic field produced in the iron core by the NIB might be more powerful than any field you can create with the electrified coil, and so your NIB might dominate the scene even when the solenoid is powered. In other words, maybe the coil will not be able to counteract the NIB's effects (or the effects of any other kind of permanent magnet you might use, if it's strong.).
Having said all of that, let me emphasize two things: first, maybe I'm wrong about how I understand your intended design with the like poles of the NIB and solenoid facing each other. And second, I'm just sorta guessing about how the NIB might effect the steel core. As I noted earlier, I'm still learning about magnetism and there's a lot about the field (no pun intended) that really confuses me, so I might be totally wrong. I'm hoping one of these EE geniuses here will chime in and comment on my concerns with your design. But when it comes to designing things from scratch, it's essential to look at how things have already been done and ask yourself why other approaches were not utilized, especially if they look so technically sweet.
I hope other forum members will check my reasoning on this. And maybe some other drawings will help here, too.
Depending upon results from prototyping I was intending to use a cylindrical magnet as the core to avoid the issue you correctly spotted when power was turned off. This discussion on the forum has already helped in determining that a cylinder / solenoid configuration will be superior to my original thought of a disk shaped magnet.
I will have to rely on the existing spring (a length of piano wire that Wurlitzer originally used) to push the pipe valve close and the air pressure differential will certainly aid in the closing of the valve. In fact I will only need the spring when the organ is off then turned on as the air pressure builds slowly and if those valves are not closed it will sound like a cat on a screen door.
The prototype will demonstrate how large a magnet I can use and still at bare minimum retain the same response time otherwise I would only be able to play lullabies and not the up beat fun stuff.
The last major change in pipe chests was the Wicks Direct Election magnet and that has been around for 20-30 years. As you can imagine this limited field of interest is not conducive to new innovations other than the electronics used to create MIDI messages and solid state drivers for the pipe magnets.
Even Wicks did not use these solenoid only solutions on the larger pipes or at higher wind pressures because of the low force when attempting to open and the useless higher force when the valve is fully opened. They had to oversize the solenoids just to get them to open.
That is why I firmly believe if I can change the dynamics to high force when I need it and lower force when I don't I can reduce the solenoid coil size and quite possibly handle the 16 foot long pipes in the organ without pneumatic assistance.
There is an old beautiful organ in the old Wannamakers store in Philly and I got a tour from the artist one day and while playing fast pieces because of the delay he was a full measure ahead of what he was hearing.
Regarding me and a live performance; well if I were a professional organist I would be much thinner because I would be starving to death.
All kidding aside, I am not too far away from being able to do that and when it is done I will post it to the forum.
Just be aware that the magnetic saturation of the material from which a permanent magnet is made might not be as high as, say, a low carbon steel. Whereas a low carbon steel might saturate at 2 to 2.3 Tesla, a NIB for example might be saturated at 1.6 Tesla, so you might lose something in the kick the solenoid can provide. Again, I'm stretching the limits of what I know about this, so there might be some clever workarounds or other magnetic materials that will hit the sweet spot and give you the best of both worlds. It's just something to keep in mind maybe.
http://en.wikipedia.org/wiki/Magnetic_saturation
That's cool. I'm not suggesting you stick with the old ways. Especially if you learn to use some of the new-fangled software that the old guys didn't have and some of these new materials, I'm sure you can develop something that will work like a charm.
-Phil