'Theory' question: Can changer string travel be calculated?

[RD Bennett]

Hey all. This may get interesting, or at least I hope it will. A friend of mine who's into robotics has been getting interested in ways to move changers and B-bender-type devices electronically. As a first step to selecting some parts to experiment with, we need to have some idea what kind of torque is required.

I've been able to figure out how much tension is on a given string of gauge X at pitch Y, but so far I have not been able to turn that into useful push-pull distances, which I think I would need as a first step to figuring out the *theoretical* torque applied to the changer. (I know it will be a lot more complicated to figure this out than just considering what must be applied to the string... but it's a start.)

Is anyone aware of some kind of equation / formula that would be able to spit out the distance a string needs to stretch or compress to execute a given pitch change?

Heck, if you have info on how to figure out the torque necessary to make it happen, that's even better.

[Lane Gray]

It'd be a semi-dynamic number, varying with string tension and string gauge.
And also remember that the tighter a string gets, the slower it responds to additional changer travel: to raise a B to C# is MORE than twice as far than from B to C and the reverse is true, the G#-F# is LESS than twice as far than G#-G.
The torque could probably be easily figured: you know the distance from finger crown to point of force, and most strings sit open around 22# of tension.
I suspect your easiest route to success would come from stepper motors, possibly with the pedals driving mouse sensors.

[Donny Hinson]

I've been able to figure out how much tension is on a given string of gauge X at pitch Y, but so far I have not been able to turn that into useful push-pull distances, which I think I would need as a first step to figuring out the *theoretical* torque applied to the changer.

Speaking practically, the system must be variable, by nature, so why worry about exact distances and torques? Movement distances and torque specs would depend on scale, string guage, and the note required, but I'd guess that an actuator that could move and hold 25 plus or minus 5 pounds, with a travel of .125" would be more than adequate. Of course, these figures would vary considerably with any leverage available.

But...

Not to be a wet blanket, RD, but I think your friend is on a fool's errand. Some things are best kept as simple mechanical systems, and the desire to automate them is frought with problems. Not only must you have speed and accuracy in a tone changer, you must have reliability, "feel", and also a reasonable cost, along with a design that will remain supported for decades. That's a tall order. Electronics technology moves at a very fast pace, and unplanned obsolescence is always lurking just around the corner. Mechanical systems can always be fixed or replicated, while electronic systems must be realized to be "disposable", at some point. Peavey stopped production of one of their ground-breaking amps (the Session 2000) because a costly IC went out of production, and numbers of old Standels were similarly rendered useless when potted modules went out of production. Home craftsmen can make levers, rods, and pedals...but they can't make IC's, linear actuators, and pressure sensors.


So, let's say someone could develop such a system. What would be it's lifespan? A lot of us here have seen one technology supplanted by another, and yet another, until what was once thought of as technologically superior (8-tracks or audio cassettes, for example) is left wholly unsupported. Witness the IVL Steelrider, whose midi pickup is no longer made, and thereby the system was made impractical because a key component is no longer produced. You can design a perfect system of electronic components...only to discover that some parts likely won't be available only a decade or two down the road. Then, your expensive "whatever" becomes a paperweight when 10, 20, or 30 years from now, the only replacement parts available are used, decades-old parts.

While there are some similar systems used now ("autotune" guitars, for instance), I wonder what the chances are that they will be working or fixable 20-30 years from now? Technology does get better and cheaper with time, but it would be a big culture change to have musicians willing to buy a new guitar every few years, and trash the old one, like we do now with cellphones.

[Jim Palenscar]

Contact Ed Packard.

[Greg Cutshaw]

Forces are really easy to measure using miniature loadd cells. even dynamics peaks and static versus dynamic coefficients of friction can be had with a load cell, amp and lap top. Sensotec, Lebow, Interface, Force all make such devices. Also advanced CAD systems such as Pro Engineer and to a lesser extent Solid Edge/Works can calculate the forces at any point in the system. I do a lot of servo automation work for a living. Servo systems are getting lighter smaller, cheaper and more responsive. The cost would still be high for a pedal steel and the space required would still be pretty large. Response lag is also a concern. More torque requires larger motors which respond more slowly. To maintain the feel of the pedals, you would want a proportional system that would respond to partial pedal depressions and different pedal speeds.

Building a prototype is now feasible and it could have programmable custom copedants and many other advanced features. I wouldn't mess with it, however, as the existing mechanical systems are excellent and probably the right system based on cost and performance needs.


Greg

[Lane Gray]

The biggest plus I'd envision would be allowing for combining two pedals or levers that you don't use together to do something different and unrelated

[Joseph Meditz]

To maintain the feel of the pedals, you would want a proportional system that would respond to partial pedal depressions and different pedal speeds.

I wonder how one would make the pedal controller. Each pedal has to send a variable voltage to control the servo motor. What that be, a voltage divider pot?

Also, how much noise would these motors make? Would there be a whirring sound?

But, in theory having servos opens up some interesting possibilities: No pedal bar or rods. The pedals could even be wireless. The legs of the guitar could be lighter. The end bells much lighter since they would take no lateral force. The levers could be made of plastic since they would require virtually no force to actuate the raise/lower. Of course there is the issue of feel, etc.

The modern pedal steel is like an old car from the '60s. It can sit in the garage for 20 years, and, if properly stored, there is a fair chance it will start right up.

[RD Bennett]

Thanks, everyone, for the input. For the time being this is not something that would ever be intended for production, but the advantages-- hot-swappable pedal changes in particular-- are pretty cool to contemplate.

I think right now we're trying to see if it's something that can be applied in a much simpler situation, such as a B-bender unit on a Tele, to get over some of the disadvantages inherent in applying bending systems to an instrument you're meant to strap on and rock out with.

Another advantage of this kind of system would be the possibility for entirely automated changes-- the system could be programmed to move a string up and down on a timed cycle, or to move it up within x number of seconds post-attack, with the possibility of such automated changes occurring polyphonically and/or polyrhythmically. This unlocks some interesting and new creative possibilities for experimental musicians.

Personally I think the pedal / lap steel of the future could be something that is purely synthesized, with the output being generated through softsynths and input coming in from a faked "pedal steel interface" that would let the user pluck strings and move the bar but would actually produce no amplified sounds of its own. Such a device could theoretically be "retuned" or have entirely new pedal changes at the click of a button. No need to have a doubleneck for E9th and C6th!

A few years back I saw this Roger Linn prototype interface:
http://www.youtube.com/watch?v=AoAOx97G8ew
...and more recently, there are programs for iOS like ThumbJam that have similar facility for executing carefully controlled polyphonic bends post-attack. (ThumbJam even "ships" with a pedal steel patch that is not too shabby at all, considering that it's probably a real afterthought on the part of the programmers.)

Anyway, we're just having some fun right now; obsolescence and long-term serviceability aren't really concerns. Thanks for all the useful input so far and I'll of course post here if we get anywhere musically useful with this experiment.

[RD Bennett]

To maintain the feel of the pedals, you would want a proportional system that would respond to partial pedal depressions and different pedal speeds.

I wonder how one would make the pedal controller. Each pedal has to send a variable voltage to control the servo motor. What that be, a voltage divider pot?

Also, how much noise would these motors make? Would there be a whirring sound?

These are things we're hoping to find out. I'm thinking that the feel of a conventional pedal steel pedal could be be reasonably emulated mechanically-- or at least close enough for the same feeling of control once you got used to it.

Getting a useful control voltage out of it would be another thing. I am thinking some kind of magnetic or optical sensor would be the best thing if we can figure out how to swing it, I don't like the idea of controlling things via a mechanically linked pot or rotary encoder. I think it would have adverse effects on the physical feel and probably wouldn't be able to generate consistent or musically smooth output.

Whirring / vibration may be unavoidable, but this again may be device-dependent. As long as it doesn't end up in the output and isn't so loud that you could hear it in an amplified-band situation, I think I'd *personally* be fine with it.

My pal and I are thinking something like an Arduino would be just fine to serve as mediator between the output of the pedals and the mechanical device. That remains to be seen way down the road of course.

[Howard Smith]

Forces are really easy to measure using miniature loadd cells. even dynamics peaks and static versus dynamic coefficients of friction can be had with a load cell, amp and lap top. Sensotec, Lebow, Interface, Force all make such devices. Also advanced CAD systems such as Pro Engineer and to a lesser extent Solid Edge/Works can calculate the forces at any point in the system. I do a lot of servo automation work for a living. Servo systems are getting lighter smaller, cheaper and more responsive. The cost would still be high for a pedal steel and the space required would still be pretty large. Response lag is also a concern. More torque requires larger motors which respond more slowly. To maintain the feel of the pedals, you would want a proportional system that would respond to partial pedal depressions and different pedal speeds.

Building a prototype is now feasible and it could have programmable custom copedants and many other advanced features. I wouldn't mess with it, however, as the existing mechanical systems are excellent and probably the right system based on cost and performance needs.

Gregg that's an exciting idea...I've thought of it, and was talking to my brother about it the other day while looking at the underside of my Carter. We built an experimental airplane together and several R/C airplanes, and he can machine about anything. But I don't have the access to the technology like you do. The servo's I've used are lighting fast, and really powerful for R/C stuff. Do you know of anyone attempting this. I sure like this idea a lot better than modeling.
Greg

[Greg Cutshaw]

The pedals would just have an encoder or resolver attached to them to input their position to a drive controller. The actuators would likely be linear motors like the Extar linear cylinders that consist of a motor driving a mini ball screw to which is attached a slide that would link to a changer. A system would be needed to replace the tuning keys to adjust the open tuning and compensate for it and be used for feedback. Doing all the servo work and having manual tuning would not be too exciting. All hysteresis could be eliminated and effects like cabinet drop could be dealt with intelligently.


Greg

[Dickie Whitley]

Good luck on your project, but I'm just thinking why are we trying to complicate what is now simple? Are we getting too lazy to push the darn pedal now?

[Jim Cohen]

Dickie, re-read the last sentence in Greg's post just above yours. I believe that answers your question. Nothing to do with laziness.

[Dickie Whitley]

Yes, but Jim, it still seems to me we're trying to overly complicate something that doesn't need to be. Isn't hysteresis already pretty much done away with in keyless guitars? Don't we already know how to deal with cabinet drop tuning-wise? All I'm saying is the all-pull system we have today is about as "simple" as it gets. The system being talked about is complicated and maintenance-wise is headed in the wrong direction. Why do you want to add more things that could go bad to the mix? I'm sorry, doesn't make logical sense to me to try to fix what isn't broken.

Steels now run $3500 - $4500 for single necks, and $5500 up for double-necks, so are we trying to push this even further with all this added new-fangeled gagetry? Just push the darn pedal.

[Bud Angelotti]

I used to know the guys that built this thing >> http://www.gizmag.com/go/4951/picture/18002/
It's a monster. Never caught on.

[RD Bennett]

I'm personally not as interested in a guitar that tunes itself, although it would be nice for some folks and I guess it could be something added on.

Simultaneously, I'm personally *very* interested in a "virtual PSG / lap steel" like what I describe above, especially for lap steel applications-- I would LOVE to be able to change my tuning to whatever I wanted without having to swap strings out, save all my tunings as instantly-recallable presets, maybe even smoothly morph between them as I play.

I could also see a future where such a virtual instrument could be made MUCH more cheaply (at least in larger quantities) than a current-production "real" pedal steel, and could also be set up in other ways that would be much more welcoming to steelin' newbies than a present-day starter E9th. It could be an instrument with all the possibilities of the steel guitar and many more. But that's not quite what we're discussing here (although I someday hope to see such a thing!).

As for "being too lazy to press the da[r]n pedal"-- I see automated changes like what I suggest in my post above as being something totally different in sound and expression from contemporary styles of play.

Even if all this could be automated like I suggest, I would *still* very much want to be able to play the pedals manually-- that would be the way I'd expect it to be played most of the time. But there *are* some pretty interesting advantages and very cool new possibilities in an electronically controlled system, just as there are potential drawbacks.

[Dickie Whitley]

"being too lazy to press the damn pedal"--

RD, put your glasses on and go back and re-read, I said "d-a-r-n". (both times)

And I'm not disrespecting what you say, I just don't think we need added gadgetry that adds more possibilities for failure in it's operations. I worked in plants for years, and it's not quality but quantity and how cheap can we get it out the door.

I do indeed wish you well in what you seek, but being the old fuddy dud I am, I just don't see how complicating the process helps.

My apologies if it appears otherwise.

[RD Bennett]

Dickie, no disrespect perceived! And sorry for the misread, the prescription in these glasses is a year or two out of date.

[Greg Cutshaw]

The virtual lap steel makes a lot more sense! Great idea! But then think how easy it would be to add a pedal rack to a virtual lap steel. Proportional pedals all done in software. The whole rig would be super light. Even 10 or 12 strings and swap tunings and pedal setups in a flash even during a song.


Greg

[Bud Angelotti]

RD - That crazy guitar I pointed you to also has programmable tunings. You can be in standard tuning and then go to a different tuning right in the middle of a song, and then back again. The limitations are the string gauges. But believe me, I'm not trying to sell you one.

[Jim Cohen]

As for being "lazy", well that's why I play pedals in the first place: I'm too da[r]n lazy to slant the bar! LOL

[Howard Smith]

Here's an interesting link. Antares which was the first real vocal correction company has something new for guitar. Its just a module that fits inside the guitar, and uses common midi foot switches to control it. I'm sure the Module will be available to the public before long. So we'll be seeing them in lap steels and all other kinds of string instruments.

HS

http://guitar.auto-tune.com/

[RD Bennett]

RD - That crazy guitar I pointed you to also has programmable tunings. You can be in standard tuning and then go to a different tuning right in the middle of a song, and then back again. The limitations are the string gauges. But believe me, I'm not trying to sell you one.

Ah, so, Bud, this kind of thing is already here-- except being executed from the tuner end, and certainly not very "playable" as it were. Didn't Gibson also experiment with so-called "robot guitars" along these lines a few years ago, with rather unhappy results?

My biggest problem with lap steel is always the darned string gauges; I'm always wanting to try out some tuning that is completely irresponsible with the strings I already have on the axe.

Also, Howard, I don't know if you've seen the Peavey AT-200 demo or not. That guitar should be available shortly and it uses the Autotune for Guitar card. It would be nice if Antares would make this available to the public but I don't think it's going to happen, at least not for a couple years.

The other disadvantage of that system is that a) it seems to need to "hear the open strings first" before each "retune," b) the probable latency and lack of sonic transparency. Antares claims neither is going to be a problem, but I don't know how this could possibly be the case. I can't play through a basic modeling amp without feeling the latency to some degree.

I think you'd be better off completely synthesizing the entire set of strings with a physical modeling synth and a fast CPU, rather than trying to continually process and re-assemble six digital audio streams at all times.

[Will Cowell]

We had a thread on here a couple or three years back - search "solenoid pedal steel". Some contributors here miss the point entirely. We don't want to avoid pressing the pedal. We want something different to happen when we do.

I did a lot of work on this myself, and have what I would want to push ahead with as a feasible design if I had the time. I don't - I'm too busy trying to improve as a player. B0b and Georg Soltun were interested too.

Measuring things like pedal position could be done by many means. Optical sounds like it would be "steppy". If so, no good. Resistive wears out too soon. Magnetic (e.g. Hall effect devices) is iffy as it is highly nonlinear. Contactless LVDT devices or similar seem pretty promising, did some work on them. They would work as positional feedback for the changer elements, and as position pickoffs for the pedals and levers.

To have a guitar where several different copedents were available on demand, touch a button and there you are, is very attractive. Obviously, no buttons these days, LCD screen and a touch sensitive overlay would work very nicely.

The actuator motors for the screwjacks to change the string tension (only way of getting enough torque and speed of response) would generate magnetic noise that the pickup could sense. So the "changer" would have to be at the opposite end of the guitar to put enough distance between them and the pickup.

The connection between pedal board and the "top deck" should of course be wireless, but that's trivial with the advent of XBee and Bluetooth etc wireless chip-sets.

The point really is, mechanical parts are expensive to produce. Electronics is dirt cheap. Electronic devices can evolve. There will always be programmable controllers. There will always be miniature electric motors too. Look inside your mobile phone handset. What do you think vibrates?

Software revisions would take the place of new changer designs. It's do-able, and economically too, it just takes enough input to get the prototype off the ground.

[ed packard]

Calculating string stretch can be replaced with measurement. Below is a photo of the tuning system that I use…it allows measurement of motion.

http://bb.steelguitarforum.com/viewtopi ... sc&start=0

I don’t like mechanical systems attached to the body…they will add noise. Using a contact mic (or accelerometer) against say the front apron and listening to the result while using the pedals will show squeaks from metal on metal etc…it would be nice to eliminate that. The link above is a thread on vibrations (on going) in the SG/PSG.

A system exists for using a contactless sensor for ped/lev position. The output voltage is used via an ADC to a uC. The other input(s) to the uC is the individual string output(s) from the “semiconductor” pickup. “pitch shift” is used as an electronic changer. Whole tunings can be changed, as well as individual, or groups of strings without changing string tension or length. The ped/lev sensors are single channels from the semiconductor pickup: More info here:

http://s75.photobucket.com/albums/i287/ ... ?start=all

Presently, retuning an E9/B6 universal to C6 works fine…as does raising up to one octave,, and lowering up to two octaves. The lowering allows replacing wound strings with plain strings to eliminate string noise and also the quick falloff of harmonics on wound strings. Much more.

The electronic changer also removes the issues re the mechanical changer such as hyst’, cab drop, etc. from tension changes.

The electronic changer allows for computer/tablet control hence is more flexible than the mechanical changer.

The contactless ped/lev sensors can be used to drive any motors/solenoids if one wants to use mechanics.