Stepper Motor Pedal Steel

[Ian Worley]

Would such a guitar make robot-like noises as you pushed the pedals?

these are not the musicians you are looking for...

[Eric Moon]

The use of midi with pedal steel is complicated by the lack of midi support for instruments that have more than 6 strings. The industry leader in analog to midi implementation, Roland, has excellent products that convert analog pitch to midi via segregated pickups but their interface to the midi control box has only 6 pins for input.
As a side issue, midi pedal steel using the Roland products lends it self to use on a 6 string pedal steel such as the steels made by Jim Flynn, Jimmie Hudson, and others.

Karlis

With the MPE spec, there is support for 16 strings, with a total of 4 octaves pitch bend, as used in Roli seaboards, and the Haken Continuum. So actually making a midi controller for PSG players has never been better supported. For that matter, it is quite possible that steel players would be better than keyboardists at picking up the continuum, although maybe being a cellist would be best.

[Eric Moon]

A screw drive – as I have mentioned earlier, works same as a screw jack for cars but in miniature. Basically it is a screw/nut-in-a-tube arrangement that takes the place of a pull-rod, and it does not need power to stay in place in neutral position or anywhere else along the travel.

- A screw drive can be made very narrow, and mounted as tight or tighter than distance between strings on a PSG.

- Miniature DC motors for screw drives can be staggered, so size of those isn't much of an issue. Just make any other screw drive longer than its neighbor(s).

- Pulling-power for a screw drive can be made plenty high enough even with the smallest DC motors. The "gearing" is in the screw/nut arrangement.
Whether or not to add "counterforce springs" to lessen the load on the screw/nut is a matter of choice. May make them last longer.

- Speed isn't an issue with screw drives either, as they can react as fast as anyone's mechanical pedal-push, or as slow as anyone would like.

- Whether to use screw drives in a direct pull-release setup (10 drives for 10 strings), or via all-pull scissors (20 drives for 10 strings), is a matter of choice surrounding complexity of the mechanics and of programming the controller.

Think I have included all the basics…

Do you have any sources for these kinds of screw drives? Sounds like this is similar tech to what is in linear actuators, but would offer more options for power/size... I assume you still have the issue of lost steps? It would certainly simplify things if you didn't have to figure out how to 'reset' periodically with a pull-release setup. And going to 2 motors per string is a lot more complication and space.

[Edit: read further about 'flagging', which seems like it would be fairly accurate. I guessing you'd need to create your own endstop mechanism to stop the motor at the end of travel. Although if you didn't use ALL the travel available, that might be just an emergency measure. If you have the flags to count rotations, then wherever the steel returns to with no pedals or levers pressed could be 'zero', which would make sense, as that's where you do your manual tuning from.]


Coming from lap steel, I have been experimenting with copedents ongoingly for as long as I've played, and I find the harmonic possibilities really intriquing, so unlike you I do have a real use for a servo steel.

Barring another PS210 showing up, it's the most promising upgrade from my multi-kord. And I'm probably barely tech-savvy enough to figure it out.

[Ian Worley]

I believe what Georg is referring to is the sort of mechanism used in CNC and 3d printing. Also called Acme screws, check robotic supply places.

I am skeptical about this sort of approach, or any approach that relies on the electronic actuating device to control the absolute center "neutral" position like in your diagram in the other thread. I understand the allure of a Jetsons-style instrument, self-tuning and able change copedents at the press of a button, it might be a fun project but it seems like the complexities would tend make it very impractical and expensive to build without a lot of trial and error to refine and make it reliable. A rotary servo or stepper with a crank arm pulling a more conventional mechanical all-pull changer would be much more reliable, but maybe that's missing the point.

[Eric Moon]

Well, if I could buy the steel I want, I would. Otherwise, I have to make something. I made my own keyboard, for instance. A long process that's taken 15 years or so to get to where it is, and still evolving.

It's not particularly marketable. But it does what no other keyboard does, and it keeps me in studio work because it gives me very quick access to thousands of top quality sounds and it sets up in five minutes. (It does help that I know how to play it.) I programmed my own midi interfaces for it, and retrofitted the yamaha keybed for aftertouch. I built my own breath controller, and I run the whole thing from custom touchscreen software on a computer I put together myself.



Was it worth the hundreds (thousands?) of hours I put into it? Probably not. But it does what I want it to do, and nothing else does. When I first started performing live with laptops in 2006, everybody thought I was nuts. I heard all the comments I hear in this forum. But I stuck with it, and it basically feels worth it at this point, although I'll be the first to admit it was a rocky road.

The steel is actually easier from the standpoint I don't get hired to play it, so the pressure is off

But does it really seem so much simpler for folks to buiild a mechanical steel guitar from scratch? (I'd probably steal 90% of the design from the ps210.)

It certainly has fewer unknowns, I'll give you that.

[Benjamin Poilve]

Hi,

Just to chime in on the subject, I've had my own thoughts, and the plan I had in my head was something along those line :
- A cheap and simple pedal steel
- Pulling strings only
- In the changer, each finger is replaced by a semi-proportional rotary solenoid (https://www.youtube.com/watch?v=YpIwH7tgnjc). Those are cheap and small, so could be integrated inside the changer itself (one issue would probably to find one that supports 100% duty cycle without overheating, which could affect the strings)
- Tuning still achieved manually by setting endstop. A PID could be used to control the solenoid and avoid that
- Hall sensors for the pedals and knee levers
A little less technically advanced, but probably easier.

On a related subject, I am also working on a digital steel-like instrument (I would mostly like it to be copedant compatible, but I am a little more flexible on input methods).
I did a "neck sensor" (https://www.benjaminpoilve.com/sources/ ... _neck.webm), but I'd like to try something with a string pull pot (something like the Onde Martenot). Still have to assemble a hall pedal rack.

[Eric Moon]

Acme screws would be right, thanks Ian.

According to the web, Acme screws are just a particular subset of lead screws... I found this for a lead screw.

https://www.amazon.com/ReliaBot-100mm-T ... ef=sr_1_29

Most cheap chinese ones on Amazon are t8, which is going to be large for this application I think, but this one is T5. That's around #10, IIRC. That's the diameter I replaced the cables on my multi-kord with, and seems sufficiently robust. Diameter of the nut is 13mm, so should fit in two staggered rows. In any case, this looks like a reasonable starting place for the one-string test.

Mcmaster has a much larger selection, and higher precision screws, if I needed them... at about 8x the price.

Flag-wheels can be mounted directly on the screws, or on a geared-up wheel for higher precision. Using more than one optocoupler/detector is another way to increase precision, and prevent "missed steps". That there are no steps on regular mini-motors - which is what I would use, means precision depends entirely on the flag/detector setup.

I'm searching for flag wheels, but it wants to send me to south dakota to ride go-karts. Not turning anything up. I believe I'm familiar with the concept, from taking apart an old mouse. Are these something you could purchase? Any clues where to look?

Also, do you have an example of the kind of motor you would suggest in this application?

I have put all my plans for such an instrument on the shelf, simply because I have no need for it and am too old to waste time on such projects. Someone will build a working one some day, but it won't be me.

I'm certainly not asking you to build one, but I do very much appreciate your expertise!!

[Ian Worley]

...I'm searching for flag wheels, but it wants to send me to south dakota to ride go-karts. Not turning anything up. I believe I'm familiar with the concept, from taking apart an old mouse. Are these something you could purchase? Any clues where to look?...

I'm not familiar with the term flag wheel, but do a search for "rotary encoder", they perform the same function, they may be the same thing that Georg described, just a different name. Used to measure 4th axis position in CNC applications and lots of other stuff

[Eric Moon]

Thanks for the link!

I had another thought. Wondering about using a hall effect sensor to measure the absolute position of the changer finger.

Thoughts? I'm going to look into the precision of them, but the idea of an absolute rather than relative measurement is appealing.

[Eric Moon]

well at $490 each, the encoders linked are going to add up fast

Wondering if something like this would work:

https://www.amazon.com/Willwin-Measurin ... 37-9752005

My first thought is whether it would have trouble knowing which direction the motor is moving. But I guess you could get that information from the control input.

[Eric Moon]

Looks like you could also use a hall effect sensor instead of an optical sensor:

https://www.instructables.com/Hall-Effe ... n-Control/

Don't know if the precision is sufficient. Still trying to figure that out.

There's also linear inductions sensors. Still looking into those.

And then there's LVDT:
https://www.youtube.com/watch?v=XlZvoFeN6PI

these are surprisingly expensive to buy. Wondering if there's a cheap source I haven't discovered. Looks like it might be precise enough for this application.

That said, I don't necessarily need to figure out position sensing right away. Plenty of other variables to test in the meantime.

[Earnest Bovine]

The sensors should report tension, not position.

[Ian Worley]

well at $490 each, the encoders linked are going to add up fast ;...

There are other options for less $. Check places like Digi-Key or Sparkfun. There are cheapo encoders under $10 but they are low res, 24-36 pulses per rev. You would likely need a much higher resolution, but using a stepper motor it seems like an encoder with the same resolution as the motor would do it (or perhaps double the res of the motor to avoid calibration issues). I believe that beyond that would just be overkill. So if you were using 1.8 deg stepper, a 200 pulse encoder would probably do the trick. Here is one on Sparkfun that is 200 step res for $30: https://www.sparkfun.com/products/10790. I'll bet if you search around you will find something cheaper.

The sensors should report tension, not position.

Tension is only relevant when tuning the initial pitch of a given change. On a regular mechanical pedal steel, open notes, pedals and levers are all tuned to fixed, repeatable positions. Once the position is adjusted to achieve the desired pitch, the mechanical functions of the guitar are only concerned with repeating those tuned position consistently thereafter. Tension sensors would not be necessary for self-tuning either as tension and pitch vary based on string diameter and composition.

[Ian Worley]

That's an interesting idea Georg, like an old Morley volume pedal.

[Eric Moon]

Okay, there's some more acronyms, and another technology to go read up on!

Found this kind of stuff on Amazon, eventually:

https://www.amazon.com/Metal-Reduction- ... ef=sr_1_49

[George Frachiseur]

Okay, question? Who's going to haul that beast around for me? It's going to weigh a ton and a half! BTW we might as well hire someone to play that dang thing for us also LOL....

[Eric Moon]

Well, if you want to follow my progress, (and chime in!) come on over to:

https://forum.makerforums.info/t/buildi ... r/84597/33

And watch the drama play out in excruciating detail. There's about six yards of discussion already.

I've made a super-basic test string for measuring torque requirements, and ordered a test motor and lead screw, due to arrive in January... the motor seems tiny, but it does have a huge mechanical advantage. And with a spring to balance string tension, I got it down to about 3 lbs pressure to make a change of a minor third, on a string with 20 lbs of tension.

It's become obvious that changing copedents is not going to be a 'press a button' affair, unless the open tuning is the same, and the string gauges can handle it. Changing the open tuning will require adjusting the return spring, and may require a different string entirely.

That said, you won't have to disassemble anything to do it.... and you could easily have an AUX pedal or two that you assign differently from song to song.

My dual reasons for doing this are, yes, to be able to experiment more with copedents, but also to make it small enough to sit on top of my keyboard, since that's the instrument people pay me to play

[Eric Moon]

got a prototype 1-string setup working...




This is not the most powerful motor I can buy in this size, but it is (barely) capable of bringing the string up to 23 lbs of tension, which is the tension of the high string on an E9. The spring is stretched far past its working range, but it's one I had laying around, and it works for now.

The mechanism is pretty C3pio-sounding right now, but I've ordered a lead-screw nut tap, and I'll try making a nut out of delrin, which will hopefully be quieter. Also, I'll have a cover over the mechanism.

This is a drawing of my current proposed changer design, in situ:



I'm planning to have the mechanism sit aft of the nut, multi-kord style. I'm also planning to make the frame shown above out of carbon fiber covered plywood. Then the tone wood (probably maple) will rest on top. The bridge and nut will sit on the wood, which will be supported only on the edges, to hopefully give some wood tone, but the frame will take the tension from the strings, to hopefully combat cabinet drop...

The motors are 10mm wide, so I'm going with a 10.5mm string spacing, which is what I'm used to with my lap steels....

The instrument should weigh no more than a typical lap steel, and could even be played like one, if you didn't need knee levers.

[b0b]

I'd like to order an 8-string, please. 4 pedals, 4 knees.