Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Questions about electormagnetic driver for guitar sustainer

  1. Aug 6, 2006 #1
    Hi, I've had a good search around here and cannot find the answers to my questions elsewhere on the forum so here goes...

    I'm getting ready to build a 'sustainer' for electric guitar based on developments described in a monster thread on the project guitar forum:

    The folks there seem to have done a lot of work, and have had some great success.

    My problem is that I like to understand why things work, and noone over there seems to have any convincing theories as to why certain things work and why others don't.

    The system is conceptually simple. The guitar bridge pickup is fed into an preamp/amp which powers an electromagnetic coil driver in the neck pickup position.

    So far the only really successful design uses a shallow (3mm) coil about 60mm long and 12 mm wide with a 3x3mm 'blade' core
    with either some little magnets under it, or the magnets from an existing pickup. The impedence is ~8ohms. the winding wire is 0.2 - 0.25mm

    The designer believes that this thin coil allows a faster transition as the signal changes the polarity of the magnet, giving it a much better trebble response, and fewer issues with phase delay.
    He also 'thinks' that the narrow core and 0.2mm wire are crucial.

    What I want is a better understanding of the relationship between the dimentions of the coil, the diameter of the wire, number of turns, core material etc. and how these variables effect the speed that the coil can respond to changes in a signal.
    I'd also like to know how they effect the phase of the signal accross the frequency spectrum

    Another major issue in this project is feedback caused by EMR from the driver being detected by the pickup. If i can start to understand the answer to my question above, I should get a better idea if it would be possible to create some sort of stacked 'humbucker' that can cancel much of the EMR feedback without having a significant effect on the frequency response and +- transition response time.
    I've been looking at the latest developments in stacked Humbucker pickups - they use two quite different coils, one to sense the string, and the other to cancel the noise - look down this page:

    Thanks, and apologies for my limited ability to describe this issue :)

  2. jcsd
  3. Aug 6, 2006 #2


    User Avatar
    Science Advisor
    Homework Helper

    First the difference between sustain and feedback is just a matter degree. You might think of feedback as just sustain that never quits.

    Here are a couple of issues here that need to be accounted for.
    One is phasing. Different amps, patch cords, pickups and driver magnet construction and placement will all have an effect on this. You would need an oscilloscope to adjust every thing properly.
    If guitar pickups are still made the way I remember then it seems likely that you could make it do the driving as well, without a separate driver magnet.
    This would require some circuitry something like what a karaoke machine uses to remove vocals.
    The point being that it would help eliminate a lot of phasing problems, if it can be made to work.

    The other issue is that you need to control the amp gain for each string. Like a frequency equalizer.

    What do you get out of this that you wouldn't get from going to the music store and buying a sustain pedal?
  4. Aug 7, 2006 #3
    however there is an enourmous difference between the desired sustain/feedback in which the pickup detects the strings, the signal is split and sent to a standard guitar amp and to an amp for the sustainer driver which in turn re-energises the vibration of the string... and the bad kind in which a squeeling feedback is caused by the pickup directly sensing the driver coil due to EMI (not EMR as i initially mis-remembered)

    Yes qhase is a key issue. It is possible to calculate the phase changes caused by the amp section, I would like to understand better how the physical design of the driver impacts the phase, and how this can be manipulated via altering variables such as core material/dimentions, coil shape etc. Unfortunately I don't have access to a scope, although with this all happening at audio frequencies, it may be possible to use a sound card and some software to get similar results ? has anyone got experience of this approach?

    Most guitar pickups have a very high impedence, a LOT of turns in the winding and VERY fine wire. My understanding is that many people have tried, and the results barely work have terrible frequency response, are extremely inefficient and burn out after a short time.
    I thought karaoke just switched the phase of one stereo channel by 180º simple by flipping the + and -, thereby muting everything in the centre... although i guess they're more sophisticated with filters to preserve bass etc.
    I'm thinking phase correction in a sustainer would require all pass filters. The thing is that the current best design works really well and has no fancy phase correction - just a simple amp and a driver. I want to understand how this works :)
    There are not seperate pickup/drivers for each string. 1 pickup for all strings, 1 driver for all strings. Good pickups will have the core of the coil shaped to be closer to some strings and further from others - thats as good as it gets in terms of equalising the response per string.
    A sustainer pedal is a different beast entirely - just some distortion and lots of compression afaict. They drastically alter the tone of the instrument.

    The sustainer is a way of immitating the kind of sustain that can be achieved when you turn up a guitar until its loud enough so that the sound waves in the air are loud enough to re-energise the strings (all the classic rock guitar sounds were achieved in this manner :)). The sustainer can also give sustain to clean sounds - even when the guitar is not plugged into an amp. Crucially it shouldn't have a negative impact on the natural tone of the instrument - just give it infinite sustain.

    One other thing I forgot to mention is that the whole caboodle needs to run reasonably efficiently from a pp3 9 volt battery.

    I have done some more searching on the subject of designing coils, and Im wondering if the self-resonant frequency may be an issue in the working range of this device - roughly 80Hz - 5khz. Could it be that the more succesful shallow driver coil works better because a hand wound coil of that design will have less capacitance, therefor a higher self-resonant frequency? I know very little about this area, is this a sensible theory? what is a good(cheap) way to measure/calculate the self-resonant frequency?

    Also i really want to know why .2mm winding wire seems to be optimum? thicker or thinner don't work as well...


  5. Aug 7, 2006 #4


    User Avatar
    Science Advisor
    Homework Helper

    That's different then. Using an E core for the driver might help a bit here. It will help to constain the mag field and be somewhat more efficent than than an I core like you seem to describing.

    Two channel stero sound card could work. How's your programming?
    Haven't seen any software to do this, but that doesn't mean much.
    You need something here.

    So much for that idea. OTOH rewinding one might not be out of the question.

    AFAIK. The idea being that you subtract your drive input to the pickup from the signal going to the preamp the same way.

    That and phase inversion.

    Just lucky I'd say. It happens :smile:
    Doubt anybody has managed to duplicate it.

    It occurs to me that fingering would cause a problem with single channel equlization. So I don't know if I would rule out separate pickup preamp and driver per string. Stuff is small nowadays.
    Pure mechanical, as you mention, could work.
    Don't know.

    That sounds more like a fuzz pedal to me.
    I was thinking of a bucket brigade type device.
    Nowadays, I have to take your word for it.

    I understand this.
    Used to work with groups, but years ago.

    Not in this frequency range.
    No way to get enough distributed capacitance to matter.
    However, coil design will effect power transfer from the driver output.

    Also likely power transfer function with the driver.
  6. Aug 10, 2006 #5
    Excuse my ignorance - would an 'E' core involve a central bar connected via its base to two other bars one either side of the coil ?
    My programming is good - however, this type of thing would be a big project and probably also require some extra hardware - something I have very little experience with.
    Yep, I'm planning to use some of the parts from a dismantled pickup combined with some tweaked and custom made parts.
    I've been thinking about the whole phase thing, and I don't think its so simple in practice...

    The driver is positioned as far away from the pickup as is practical in order to minimise the EMI issues - about 12cm
    This means that the natural harmonics that characterise a good guitar tone can be cancelled or boosted depending on the note due to the phase difference between pickup and driver - no matter what attempts are made to 'correct' the phase, some notes will be better than others.
    Also, with one pickup/driver combo for all strings, any attempt to 'fix' the above issue will only work for one string - not all 6.

    My guess is that the best setup is one that favours the higher/thinner strings as these are more difficult to drive because of they much lower mass, and that the phase/harmonics response is a 'feature' that adds character :). Ideally this could be controlled - maybe with a switch or pot installed on the guitar.... maybe an allpass with a trim pot so the thing can be tweaked for the guitar its on (is this doable)... I'm no filter guru, but I'm sure its possible to tweak the phase/frequency response curve so that it gives a 'better' response (in practical playing rather than theory). would this require a 2nd order filter...

    Lots of ideas :) maybe its time to build the damn thing and try some stuff

    Hmm :) I think you may have something there - others have had very varied levels of success. Maybe the ones that work have accidentally hit a good compromise that favours the higher strings and has a 'better' phase response... Or maybe their guitars/pickups/playing styles are more suited to the existing design.
    Many issues await if you go down that route... bending strings becomes problematic. All the extra electronics for 6 buffers and poweramps, and the associated power required drive them make the goal of mounting the whole kaboodle in a guitar with one 9v battery less attainable.
    Not to mention the fact that it is possible to make a successful working unit the monophonic way.
    Bucket brigade was used for delay based effects echo/delay/phaser/flanger rather than for sustainer ( I think)

    That makes sense, I've been reading about how most of the 'character' of different passive pickups comes from their self-resonance characteristics, but they have thousands of windings rather than hundreds and very high impedence...
    It might be interesting to put a capacitor in parallel with a sustainer driver in order to bring the self resonance frequency down to within the devices range to see if that can be used to tweak the response.. how would this effect the impedence of the driver - if a cap big enough to bring the self resonance frequency down to less than say 5khz was used, would it have enough of an impact on the impedence to need a different number of windings on the coil ? I can't even take a wild guess at this.
    my ignorance again... How does the guage of the wire effect the power transfer function of a coil ?


  7. Aug 10, 2006 #6


    User Avatar
    Science Advisor
    Homework Helper

    Just like a 3d E. In this case the 3 bars of the E would point at the strings. Laying on its back if you will. Most of the mag field will stay between the tips and will spread very little. Ferite would be best.

    You could just record and view the results in Wav editor.
    Bad for realtime adjustment.
    Second hand scopes can be had cheap.

    This why I sugest trying a single diver/pickup.
    Closeness counts here too. So a tilt to the driver or split winding coils will effect power transfer. Also appiles for separate driver.

    There are lots of issues, but size and power probaly arn't on the list. If you want to get fancy the whole bit could hide under a quarter and you can have all the power you need by feeding it up the patch cord.

    Mostly, but feedback/sustain is a kind of echo.
    Don't know how this would work out in practice.
    The way I recall people playing, you most likely wouldn't want it running all the time. So perhaps a short decaying loopback would do the trick for the chords you wanted to hold.

    Even here it is mostly about how the spcefic coil interacts with the input impedance of the amp.
    This is pretty much standard acrost all amps.
    Parallel caps I would say are no good, but tweaking the output characteristics of the driver can have a big impact. Different wire size means different resistance, thus a different load to the driver. A driver ment for a 4ohm speaker just isn't going to do much with a high impeadence coil.
  8. Aug 20, 2006 #7
    Ah, ok - would that work if the outside 'fins' of the E were attached after winding somehow? hand winding inside such a small structure would be extremely tricky - particularly with the epoxy I'm using for potting the coil.
    I found a few different free soundcard -> scope apps, and also there are a couple of projects with some hardware mods described...
    I don't see how it would be possible to make a combined driver/pickup. The pickup works by sensing the small changes to its magnetic feild caused by the string... the driver works by making large changes to the magnetic feild around the string... if the driver and pickup are in the same space, the pickup will just sense the output of the driver causing horrible direct feedback. How can this be made to work ?
    You're right that power may not be an issue, but the size of the electronic circuitry would - I don't have the skill or patience to hand solder tiny surface mount components :). And the complexity of the system would be comparatively enormous.

    makes sense - so finer wire gives a greater resistance per winding ?
    How does the mass of the core effect the power transfer ? I have a selection of iron bars to use. Have tried 2mm x 6mm x 60mm with about 130 turns of 0.23mm wire with an overall resistance of about 7.5ohms. The coild was 2mm deep. It seems to work pretty well, however the balance accross the strings is not great - when the frequency response of the amp i've breadboarded is tweaked so it will drive the high E string, it struggles with the low E. The G string (most massive unwound string) is a little too responsive compared to the rest as well.... I assume I should be able to tweak the string to string response of the 'final' version by cutting v shabed notches out of the core beneath the more responsive strings, or stepping the top of the core.

    What difference do you think using a more massive core would make ? say 4mm thick instead of 2mm ?

    Other ideas i'm toying with are automatic gain control for the driver amp, and a variable all pass filter to give me more balancing options...


  9. Aug 20, 2006 #8


    User Avatar
    Science Advisor
    Homework Helper

    This may be of interest.
    Just read an article in Aug Nuts & Volts.
    Roland offers the aftermarket GK-2A Hex pickup.
    Meant to use with their synthesizer, but it has independent channels for each string.
    Also available factory on Fender Stratocaster.

    Wind on a bobbin then drop in.
    Or perhaps some small toroid cores with air gaps.
    Glue or screws are ok.
    Iron bars will have eddy current problems.
    What you are doing is building a solenoid.
    I'd look up the equations for that.
    Core sizing, material and wire size depends on how much power you need and saturation effects

    Circuitry to cancel out what you put into it.
    Same basic concept the Karaoke machine uses.
    The same type of circuit can be used to cancel out magnetic leakage between a separate pickup and driver.
    Phasing is the key here.

    Spacing between the core and strings can do the trick.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook