| Author |
Topic: Pedal steel guitar stepper motor experiment |
Karlis Abolins
From:
(near) Seattle, WA, USA
|
Posted 10 Nov 2020 3:41 pm |
|
In another thread, we discussed using alternate mechanisms to change the pitch of a pedal steel guitar string. I decided to experiment with using stepper motors. It has been a fun experiment. Here is a link to a video of my results https://www.youtube.com/watch?v=WL-ByqMsCVs.
Karlis |
|
|
 |
Jim Palenscar
From:
Oceanside, Calif, USA
|
Posted 10 Nov 2020 5:17 pm
|
|
| Hat's off to you Karlis! Were you able to hear the stepper motor via the pickup? |
|
|
|
Paul Sutherland
From:
Placerville, California
|
Posted 10 Nov 2020 5:20 pm
|
|
It looks complicated, but it clearly works. Now expand to a full 10 string set-up and take it for a test ride.
_________________
It don't mean a thing if it ain't got that swing. |
|
|
|
Karlis Abolins
From:
(near) Seattle, WA, USA
|
Posted 10 Nov 2020 6:44 pm
|
|
There was no discernible stepper motor noise through the pickup. The physical noise is also indiscernible. The stepper motor driver I used, TMC2208, is called silent stick driver and it lives up to its name. I tried an older stepper motor drive, A4988, but it was too noisy physically. Its sounds were like my 3D printer stepper motors - that is noisy.
Scaling this up to 10 or 12 strings is a challenge that I am going to defer. However, a six string pedal steel design has crept itself into my imagination.
Karlis |
|
|
|
b0b
From:
Cloverdale, CA, USA
|
|
|
|
Doug Palmer
From:
Greensboro, North Carolina, USA
|
Posted 11 Nov 2020 8:17 am Future steel
|
|
This is great. I had only dreamed of this. This could be the future of steel.
_________________
Emmons D-10, ST-10,LD-10 III, NV-112,Fender Deluxe Reverb. Authorized wholesale dealer musicorp.com! |
|
|
|
Jim Pitman
From:
Waterbury Ctr. VT 05677 USA
|
Posted 11 Nov 2020 9:17 am
|
|
That's great Karlis.
Like a mechanically linked pedal steel it would be necessary to insure the changer finger angle is proportional to the pedal/knee lever angle.
My two cents as an automation guy:
The best candidate for a pedal input device IMO would be to use a brushless transmitter resolver. A resolver is a rotary position indicator. It can be small. It uses an electro magnet that is energized by a stationary reference coil. The magnet is on the shaft so rotates within two other coils which are also stationary and are positioned at 90degs to each other. As the shaft rotates the amplitude of the waves on the two other coils change. The reference waveform, (usually a 400hz sinewave) appears on both output coils, however, the amplitudes of the waves of the two coils will change, the amounts being related in one case by the sin of the angle of the shaft and in the other by the cosine of the angle of the shaft.
The advantage of a resolver over a potentiometer is, one, there is no mechanical contact to wear out, and two, infinite resolution is easily accomplished and is absolutely repeatable.
The advantage of a resolver over an incremental optical encoder is that it is absolute rather than relative so therefore doesn't require initialization.
All pedals and knee lever resolver mechanisms can share the same reference signal. It can be bussed. There are analog ICs available designed to convert the sin/cos output amplitudes to a linear voltage. This voltage could be read using an Analog to Digital converter device (A/D). So the resolution is only limited by the number of bits for the A/D device. The 12 or so linear signals from the levers and pedals could be multiplexed into one A/D to be read by the microprocessor.
Only if I had the energy I had when I was younger I'd get right on making a prototype. Nonetheless I'm happy to consult for free. |
|
|
|
Jon Light
From:
Saugerties, NY
|
Posted 11 Nov 2020 9:30 am
|
|
Not only is this really interesting; I extra-appreciate it because as a tinkerer, I've wondered about some of this stuff and now I know that most if it is way beyond me. Saves me from the situation of buying stuff, thinking that it'll become obvious once I start mucking around. Nope. I'd need to do some serious remedial studying before I was ready for any practical application. That's valuable info!
Nice job. |
|
|
|
Jim Pitman
From:
Waterbury Ctr. VT 05677 USA
|
Posted 11 Nov 2020 5:16 pm
|
|
I'm covid board.
I've decided to make a pedal input device for such a stepper motor controlled PSG.
All I need is a mcehancal mechanism that has a hard stop for the start and the stop of the pedal/knee lever travel. A spring will return it to the the start position. Perhaps some sort of feature whereby you can adjust the rotary friction would be desirable.
Karlis - Pm me if you are interested in collaborating. No pressure of course. |
|
|
|
Bill Terry
From:
Bastrop, TX
|
Posted 12 Nov 2020 5:22 am
|
|
This is pretty cool... Great job Karlis! I hope you and Jim keep this thread active with your work, it's very interesting.
One time I was sitting in a bar changing a broken 3rd string and had the thought that no doubt it broke because it gets a lot more 'exercise' than others.
I'm an EE, slanted more toward the embedded software side, but I got to thinking how easy it would be (for somebody more mechanically inclined than me) to install sensor or position indicators of some type, and actually log every pedal and lever move for an entire night, or by song, or however you wanted to collect the data.
I finally decided it was an achievement of no merit, because who cares right? Besides, I didn't want hard evidence of how limited my playing really is.. BUT.. this project and Covid have me thinking about it again. With that data, you could in theory generate tab automatically maybe, or with some mechanism to actuate strings, even make a 'player piano' style steel, using the ideas Jim and Karlis have come up with.. Load up the performance data, hit 'play',
Ok, too much coffee already today. Haha.. |
|
|
|
Ross Shafer
From:
Petaluma, California
|
Posted 12 Nov 2020 8:14 am
|
|
| way cool, good job Karlis! |
|
|
|
Jacek Jakubek
From:
Mississauga, Ontario, Canada
|
Posted 12 Nov 2020 8:23 am
|
|
Interesting. The only limitation I can see with this is that it might not allow you "squeeze" the pedals, or press them slowly or faster for expression, and also to achieve half-pedaled notes.
I was very skeptical of this kind of motorized string bending until I saw a guy on you tube that played a 6-string with these motorized benders. He had them mounted on the body of the guitar just behind/above the bridge and it sounded great. I forgot the title of the video and can't find it now, otherwise I'd leave a link to it. |
|
|
|
Bill Terry
From:
Bastrop, TX
|
Posted 12 Nov 2020 8:35 am
|
|
| Quote: |
The only limitation I can see with this is that it might not allow you "squeeze" the pedals, or press them slowly or faster for expression, and also to achieve half-pedaled notes.
|
I had that thought initially as well, and that's certainly true with Karlis' 'button' approach for proof of concept. But if I understand correctly, the resolver Jim mentions up above allows the finger/changer 'amount' or 'position' to be controlled by the position of the resolver. i.e. a given degree of resolver of rotation has a corresponding 'position' of the changer finger. So squeezes, half-pedals etc. should be possible.
Jim, did I understand correctly? |
|
|
|
Earnest Bovine
From:
Los Angeles CA USA
|
Posted 12 Nov 2020 8:59 am
|
|
| Bill Terry wrote: |
| finger/changer 'amount' or 'position' to be controlled by the position of the resolver. i.e. a given degree of resolver of rotation has a corresponding 'position' of the changer finger. |
I still maintain that ideally a sensor would report string tension, not changer position. |
|
|
|
Bill Terry
From:
Bastrop, TX
|
Posted 12 Nov 2020 9:06 am
|
|
| Quote: |
| I still maintain that ideally a sensor would report string tension, not changer position. |
Hmm.. is the tension to pitch correlation precise enough for that to work? Maybe it would require some sort of initial 'calibration' step?? For that matter, either method would I guess? |
|
|
|
b0b
From:
Cloverdale, CA, USA
|
|
|
|
Earnest Bovine
From:
Los Angeles CA USA
|
Posted 12 Nov 2020 9:34 am
|
|
| Bill Terry wrote: |
| Quote: |
| I still maintain that ideally a sensor would report string tension, not changer position. |
Hmm.. is the tension to pitch correlation precise enough for that to work? Maybe it would require some sort of initial 'calibration' step?? For that matter, either method would I guess? |
Yes, but I think calibration would be less frequent than it would be with a position sensor. BTW, calibration could be done automatically; all you would have to do is pick the string, and I suppose you could even imagine a system with something like an E-bow to do that for you (not worth the trouble). |
|
|
|
Bill Terry
From:
Bastrop, TX
|
Posted 12 Nov 2020 9:39 am
|
|
| Quote: |
| calibration could be done automatically; all you would have to do is pick the string |
Yeah, open and raise/lowers I suppose. This could put a lot of changer wrench vendors out of business.  |
|
|
|
Jim Pitman
From:
Waterbury Ctr. VT 05677 USA
|
Posted 12 Nov 2020 2:40 pm
|
|
| Yes, tlresolvers atlre absolute and don't require calibration nor are they susceptible to temperature drift etc. |
|
|
|
Bob Russell
From:
Virginia, USA
|
Posted 15 Nov 2020 8:45 pm
|
|
Karlis,
Kudos to you and Jim for exploring this new possibility!
_________________
Lots of stringy things, many of them slidey. |
|
|
|
Jim Pitman
From:
Waterbury Ctr. VT 05677 USA
|
Posted 24 Nov 2020 2:47 pm
|
|
So you go to a PSG show and you sit down at a brand new (insert brand) PSG and try it out. Immediately you are confounded because the Eb lever isn't where you are are used to it being for example. This clouds your first impression. And you simply can't get by it. "This steel sucks" you think. Which of course is not right. It's just different.
This is the primary application for a stepper motor PSG. The ability to change the copedent quickly is a great marketing tool for the PSG builders. |
|
|
|
