Stepper Motor Pedal Steel

[Karlis Abolins]

The recent post by b0b of the ElectrAdair pedal steel got my gray cells churning. Those of you who are 3D printing enthusiasts will probable understand the inspiration for my concept.



The idea is to use a sensor on the pedal which senses its current status (rotation). The current status is converted to a digital signal and sent to an Arduino control circuit which controls the motion and location of the stepper motor output shaft. The motor has a pulley which is connected to a step-down pulley which then pulls a cable that controls the position of the changer finger.
In action, the up and down motion of the pedal (side-to-side for a knee lever) is accurately and smoothly echoed by the appropriate motion of the stepper motor resulting in a raise or lower of the string tension. Each string would require only one stepper motor regardless of the number of pedals that control it.

This is only a concept but could be prototyped on a single string model with off-the-shelf components.

Karlis

[Ian Rae]

You could change the copedant at will - no more moving cranks around

[Pat Chong]

Interesting concept. One would wonder, though, that through the step down process, if the motor could move fast enough for "bouncy" licks. A more powerful motor may do the job...... And as Ian mentioned, copedents could be changed (programmed??) quickly. Or even maybe tuning, say E9 to C6!.......Pat.

[Georg Sørtun]

Fine for a single-string demo.

For multiple strings a central "brain" is needed to cope with splits, compensations etc. and all that one can think of for multiple pedals/levers affecting a system of raise/lowers at any speed. Note that most stepper-motor systems are not very speed-variable.

Also note that stepper-motor setups need some form for "reset" from time to time - to counteract tendencies to loose track of steps, where all PSG-changes have to be "zeroed out" one way or another in order to stay synced.
Easy enough to program in such resets in most applications where stepper-motors are used, but those are "pre-programmed" and not under "musical and emotional human control".

Once up and running it can cope with all kinds of tunings without missing a beat, but lots of theory and programming needed to make it compete with the all-mechanical systems we have today.

[Karlis Abolins]

Good points, Georg. The unit would require a micro-computer to control the interactions and to ensure centering.

Pat, there is probably a trade-off between size and torque. The larger high torque motors may not fit 10 or 12 under the hood of the guitar. The smaller ones may not have enough torque. There are extremely fast stepper motors like the ones in your computer hard drive.

Clearly, this is a concept and not likely to ever become reality but I believe that it can be done with today's technology.

Karlis

[David Ball]

Some kind of linear servo might work--having some kind of feedback loop so that the controller always knows the exact position of the servo would take care of the lost steps that can happen with steppers. They can be very fast, very accurate and small.

But all of this sort of reminds me of the Gibson G force autotuners that were in the news a few years back--it's obviously possible, but...

Still it would probably be fun to tinker with.

Dave

[Joe Naylor]

We looked at this some years ago. Another problem is cost. Besides making something simple complicated. We decided that it was and interesting idea but not a better mouse trap. Looked at servos too.

Think is you had a E9th and C6th how many you would need. More cost.

Then there is the issue EVERYONE worries about - WEIGHT.

Not for me.

[Paul Sutherland]

Push ahead full steam. It's about time for some real innovation of the PSG.

[Ian Worley]

Karlis, there have been several discussions about this and similar ideas here on the forum in the past. There is a guy who developed a working system on a six string guitar with servos and an Arduino controller: viewtopic.php?p=2363624#2363624.

The concept would easily translate to psg if someone were willing to invest the time and $ to develop and refine it. You could even start with his Arduino code for the controller. I think Joe Naylor's point is probably the main reason no one seems to have done it though -- it's not really a better mouse trap, just a different one. More possibilities for different tunings and tweaks, but also way more potential points of failure.

[Larry Bressington]

Brilliant Karlis.

[Pat Chong]

"more potential points of failure", I can see that.

I NEED HELP. Does someone here repair steels? I have a gig coming and my pedal steel crashed and won't boot up!

But, it is an interesting idea, and could have possibilities. .......Pat.

[b0b]

Brilliant, Karlis!

Some years ago, I envisioned a hardware interface to an electronically controlled pedal steel changer. The pedals and knee levers on the guitar would be wired to a computer instead of pulling the changer directly. The computer controller would allow the player to instantly change tunings and copedents. I couldn’t find a manufacturer who was willing to buy into the idea, but a guy can dream. Here’s the “user guide” that I wrote for the fictional controller:

Pedal-Steel-Changer-Controller.pdf

[Paul Strojan]

I think a better solution would be to bend the strings electrically. Joni Mitchell has a guitar that has a pickup on each string so she can change tunings on the fly.

You could put a piezo on each string and wire them to a computer that will "auto tune" the input according to the desired copedent.

[b0b]

I think a better solution would be to bend the strings electrically. Joni Mitchell has a guitar that has a pickup on each string so she can change tunings on the fly.

You could put a piezo on each string and wire them to a computer that will "auto tune" the input according to the desired copedent.

Would it still sound like a steel guitar? I sort of doubt it.

[Benjamin Davidson]

This is an interesting topic, but to develop and build such a unit wouldn't be cost effective by comparison. I have some experience electronic control systems that would be where I would start an undertaking like this. However there is a lot of things to consider, well we have the time on our hands. Instead of using servo motors though, I'm going to recommend a precision solenoid, transducers on each pedal and lever, and a central processor. Something like the signal generated from the mechanism inside a Hilton or Teleonics volume pedal for pedal position would be the most logical starting point for the cross shaft transducers. I'm also going to assume we keep either style (Keyed or Keyless) head stock - that won't be manipulated by the system we are discussing.

-- First, the changer design itself. Are we attempting to maintain a triple raise, triple lower with 6 potential pulls each - or simplify it to single raise, single lower, and a split tuning stop? Both processes could be employed, and there would be advantages to both. With the triple raise/lower changer, we can keep the familiar tuning processes we have come to expect, however we would lose the ability to change things on the fly quite as much, and changing things would require the individual tuning nuts to be adjusted. With a single raise/lower changer and the software handling tuning, we reduce the equipment, and allow for tuning presets without the need for any mechanical adjustments to be made between Copedent Setups. In the end I feel that once the software development is dialed in one could achieve more, and have more reliability out of a unit that is single raise/lower. Either way we go about it there will be a different feel.

-- Floor Pedals. Going to this system could reduce the need of floor pedals on a Double 10, in reality we could have 5 pedals, and shifting of tuning presets (just like shifting necks) would reassign the Pedals from A B C, to C6th 4-8 for example. A 12 string universal Copedent would still require 7 pedals as I see it. However since most of us have become accustomed to the layout we will assume that we'll maintain the typical 8 floor pedals. I wouldn't have the sensor located on the pedal rack though. First that would require an interface cable to run the bar and back. Along with needing the analog to digital converter mounted down there as well. I would mount the transducer in the undercarriage of the guitar for both the Pedals and Knees. This will allow the traditional look of the guitar, and maintain cable lengths to a minimum. With a spring to replace the tension of the cross rods we are used to now, it should be possible to maintain the same feel on our floor pedals. A bell crank like plate would be used against the transducer(s) to create a precision input signal to the processor in so that our feel of the pedals and emotion can remain part of the music.

-- Knee Levers. Similar to the spring tension and transducer setup of our floor pedals, again we have the chance here to reduce the number of knee levers we have under the guitar. Most players would keep the 5 most common locations, however changing what those levers do between presets could be very creative. For example our LKV could be a B to Bb one song, and shift to G# to F# the next.

-- Latency. Processing Latency is the largest factor in why I don’t feel this would catch on for the industry. Even with all the technology coming through with modeling amplifiers and effects units, there is a latency as our analog pickup signal is converted to digital, processed, and returned to an analog output to drive a speaker. Right now, we move the change - the changer moves – and the note bends. With any form of electronic steel, we move the change – transducer feeds the analog to digital converter – signal passes to the central processor – change in interpreted – signal is passed to a digital to analog converter – the solenoid pulls the changer – and the note bends – a feedback signal returns via an analog to digital converter and tells the processor the changer is in the correct position for that note.

-- Pickup Design. We have an opportunity to advance here as well on this new process. Should we maintain the pickup design of old (single coil or humbucking doesn’t really matter here) or shift our thought process and develop a pickup of 10 individual coils and independently process each string signal? With the processing power of some effects units (and studios now able to micro tune individual notes of an cord) it may be possible to capitalize on individual pole coils and process each string individually. This would also allow for application of that signal to be part of the feedback loop, and be checked at the processor for proper tuning and EQ. Piezo pickups were mentioned, however I think if one was to use piezo they would have a lot of mechanical artifacts coming through the pickup.

-- Power Requirements and Hardware. We all would agree this whole thing is going to be digital, with a decent processor as a central hub. We have a lot going on here before we start talking about transmitting a note to front of house, DAW, or amplifier. We would need something powerful enough to do the following: 1) strobe style tuner, with sweetened tunings, that can be displayed and used by the processor for feedback purposes. 2) have the ability to set and recall various Copedent presets easily, this includes the ability to reconfigure the pedals and knee levers digitally. 3) process the pedal and knee lever signals to bend the strings at the changer with a similar emotional feel that is really the soul of the pedal steel guitar. 4) Process feedback data from the changer transducers and the pickup signal to ensure that the sound is in tune (which becomes an extra variable when a player’s bar hand is off. 5) Provide the output signal either as a 1/4in plug or perhaps line level MIDI output. 6) Interface with the Volume Pedal easily.

-- Tuning and String Timbre. Logically we can retune the headstock, shift presets and transition between tunings and Copedents. The downside to changing this way is the individual loss of timbre a specific string gauge imparts on the tone of the instrument. A double neck guitar may still remain desired for players looking to routinely shift between two tunings. On the other hand the concept of shift on the fly tunings may be preferred by other players just as easily.

-- Tone, I really don't foresee this device sounding like a steel guitar as we know it... but could get there with the right investment (if that would be profitable is another story).

[Earnest Bovine]

some kind of feedback loop so that the controller always knows the exact position of the servo

Much better if the feedback is tension, rather than position.

[Earnest Bovine]

I think a better solution would be to bend the strings electrically. Joni Mitchell has a guitar that has a pickup on each string so she can change tunings on the fly.

You could put a piezo on each string and wire them to a computer that will "auto tune" the input according to the desired copedent.

Would it still sound like a steel guitar? I sort of doubt it.

No, it was surely just something like Gibson G-Force or Min-ETune.

[Joe Naylor]

These are very interesting ideas. I was a partner in a computer control company for many years. Some facts remain: (1) COST of research and development
(2) Cost machine after development
(3) Weight
(4) Very few could or would really be interested due to the complication factor.

I do not in anyway want to through a wet blanket on the idea but facts are facts. Much simpler but a good example after building steelseats for several years I have found that even on seats many people want certain options. I develop and offer it then the exact person wanting it never buys it. Then others can not believe it cost more. And of course the always concern that it is too heavy.

I would love to see this and can only imagine the buzz if someone came up with one and introduces at a show there would be lots of interest and more ideas.

I will stay tuned and hope someone says that they are going to undertake this. Just hope they have buckets full of money.

Once again, some great ideas, and hope that there are more to come.

Stay Safe

[Rich Upright]

OK but...Why? It would make a ridiculously complicated instrument even worse. And...something else to plug in onstage. I think I'll stick with the tried & true.
remember...Just because you CAN do something doesn't mean you SHOULD.

[Mark McCornack]

I don’t know. To me, whether it SHOULD be done or not doesn’t detract from the intrigue of thinking of the possibilities. Many legitimate pragmatic aspects of this have been raised, but it still seems like a fascinating notion. Having ten steppers chooching away (maybe each as big as NEMA11 or NEMA17 to do the job!) and the power, weight, and cost associated Is pretty over the top, BUT, consider these points

ANY and ALL strings could be pitch compensated if needed for ANY and ALL pedal/lever combos.
Any and all cabinet drop issues could be eliminated (see compensation above).
ALL pedal and lever functions could be functionaly “reassigned” at will. Even mid song. Maybe a small bank of switches near the pickup?
Lever and pedal action could be made ridiculously light. You’re just turning pots. Less general thrashing around of the guitar when playing.
Closed loop open tuning could be achieved (at least to the degree of the current state of the art).
The actual processing power required may be overestimated here. My gut feeling is that would probably be one of the lesser challenges. Most of this would be based on table lookup, and at the rate things are changing from the player/listener perspective (milliseconds) versus machine cycle times in simple processors (nanoseconds). Again, just a gut feeling, but I’ve done a bit of embedded development over the years.

I totally agree with this threads previous comments regarding practicality, but still, the Franken-Steel is fun to think about. As a retired engineer, it sounds pretty neat. As a retired engineer who didn’t win the lotto, and who has other better things to do, meh. It’s fun to think about in short bursts though.

[Georg Sørtun]

I think a better solution would be to bend the strings electrically. Joni Mitchell has a guitar that has a pickup on each string so she can change tunings on the fly.

You could put a piezo on each string and wire them to a computer that will "auto tune" the input according to the desired copedent.

IIRC, Ed Packard did some work in that direction years ago. He used Hall-elements as single-string PUs though. Can not remember much more about his ideas, but I do not think "lack of natural sound" was a problem.


I went through a theoretical design with motors over several years, and if we ignore the fact that I designed the entire changer into a traditionally looking keyhead, it was an All-Pull design with one raise and one lower to cover all changes one might think of and the strings can tolerate. Maybe you'll spot why I ended up with two motors for each string, instead of one with zero-position in the middle, if you read further down in this short description.

To get the immediate reaction and high resolution known from regular, mechanical, PSGs, miniature DC motors was found ideal to drive short pull-rods via screws. Positions was counted up/down via flags on the screws.

Screws have the advantage that they stay put when no voltage is applied to the motor they are directly connected to, and high speed (for immediate reaction – follow the pedals/levers travel exactly) combined with well-defined stops when position-counters reach set values, are easy to achieve with very small motors and screws (for low weight and insignificant inertia).

While each miniature motor will have its own two-way driver/analog controller for course settings – +/- a few cents, the CPU will calculate and control that same driver to within acceptable tolerances – say: +/- 1/2 cent of whatever it gets programmed to. The CPU will then also activate the drivers for raises/lowers on other strings for optimal cross-string tuning, but only when main changes are within a few cents from programmed destinations – pre-programmed compensations if you like.

NOTE 1: Flag-counters were chosen because all other position-detection methods – string-tension, linear positions, etc. – will be affected by the player via the strings and not be reliable. Flag-counters are also easy and cheap to integrate with a screw mechanism, and with an All-Pull setup auto-reset at pedals/levers releases is a given.

NOTE 2: Miniature DC motors small enough to fit side-by-side in the space provided for regular mechanics in a standard-width PSG, are on the market. Power is not much of an issue with linear screw-drives, so weight of the entire motorised mechanics can be kept at level with, or lower than, traditional pull-rod/bellcrank mechanics.

NOTE 3: As pull-length for strings varies greatly – from 3d E9 string about 1/10 inch for a halfnote raise, to the much smaller pull for the extended low on an S12 E9 or C6 – I chose to level out some of that difference in the mechanical All-Pull changer instead of leaving all that to the electronics controlling the screws/motors. Easier to achieve the wanted precision that way.

NOTE 4: Although my "design" allows for jumping between pedal/lever actions and compensations in any order of setups imaginable, open is tuned as we are used to for traditional PSGs. No built-in auto-tuning either.


My reasons for not having built one already, is that I am no longer interested in all the hassle related to marketing such a PSG, and I have no need for one myself. Thus, figuring out how I would go about it if I should ever bother to build one, has given me plenty enough fun over the years.

[Jim Pitman]

Being a motion control electrical engineer and pedal steel player for many years, i have only one thing to contribute.
A synchro device would be an ideal feedback sensor for pedal postion because:
1. They are absolute over a single revolution - need no power up reset to determine zero positon
2. They are analog and have infinite resolution so only the A/D device limits the resolution
3. They are reasonably inexpensive.
4. They require a sinewave source that could be shared by all pedals via a two wire bus.
5. Chips to decode sine to linear are available.

[b0b]

Lots of good ideas. Maybe we should have a contest to see who can build the best 1-string electro-mechanical pedal steel changer prototype.

[Karlis Abolins]

That's a great idea - to have a "competition". One result would be the ability to use the best aspects of the various designs to take this concept to the next stage, which is a prototype 12 string universal electro-mechanical pedal steel guitar. The members of the forum have an unplumbed depth of knowledge and experience that can make this happen.

Karlis

[Ross Shafer]

3-2-1-GO!