col

Making the driver very narrow will compromise it's performance. Having a driver with its own permanent magnetic field so close to a pickup will change the sound of the pickup noticeably because it will significantly alter the shape of it's magnetic field. Moving the pickup further from the neck to allow space for the driver will also alter the sound of the pickup.

None of the info on the sustainer ideas thread will tell you how to have neck and bridge pickup that can both work with a sustainer driver. Its pretty much pickup at bridge with sustainer driver at the neck. Pete's 'piggy back' idea is neat if you want to keep the neck pickup, although you won't be able to have sustain on when using the neck pickup. It works by having the driver and neck pickup share a permanent magnetic field. I think that this will have minimal impact on the sound of the neck pickup, assuming you can modify it to make room for a thin coil to stack on top (not easy with some pickup designs).

You didn't say what style, or whether you can pay or need free, so: Toby Walker does live skype lessons (I havn't used this service, but he is a good player - check him on youtube). Stefan Grossman has free youtube samples (whole song lessons) from guitar workshop dvd's. Guitar workshop tutorial videos can be bought as digital downloads from their website. Matthieu Brandt has a great website with lots of great stuff, he also has some great tutorial vids on This is a good jazz guitar site with lots of stuff to learn. For rock and pop type stuff, youtube has millions of useful clips that will keep you going for years. Best thing to start is get a guitar and some guitar music you like, then try to work out what the notes are. Try to meet guitar players and ask them to show you stuff. Get a good book (one that you can understand) that has some music theory for gutarists in it. Cheers Col

No problem. If you are getting some sustain on the lower strings, then a buffer is likely to help - it will prevent 'tone sucking' ie. roll off of higher frequencies due to the input of the circuit loading(whatever that means) the guitars pickups.

Gotta disagree there. The reason Pete's design works is because through an extensive process of trial and error, he homed in on a set of specifications that provide suitable electrical performance: The wire Pete specifies when wound to 8ohms around a suitable(!) core will produce a coil with an inductance value within a suitable range. The combination of this Inductance and DC resistance mean that the Impedance of the coil is fairly linear over the guitars frequency spectrum - only rising considerably at the point where you are playing higher up the neck so sustain is easier to generate (strings closer to driver, driver closer to belly of notes standing wave) The combination of output cap coil inductance and resistance generate a low Q resonant peak at or close to the frequency of the open high e string - the most difficult string to energize. the low inductance of the coil combined with the value of the output cap provide a phase relationship linear enough that phase lock can be achieved between driver and pickup without wasting too much power (assuming that you drive circuits phase response is compatible - big assumption). As far as build quality of the coil - yes that is crucial, but has nothing to do with the reason the design is successful. Any system of any design for any purpose needs to be manufactured to a reasonable standard or it won't work. Assuming the coil is reasonably well made, the really important factors in getting Pete's design to work are - the combination of core material and winding wire give you an inductance close to the inductance Pete gets (whatever that is). The circuit you use has similar phase response to the circuit Pete uses (whatever... you get it). So there are some unknowns that cannot be solved other than by trial and error. try using different core materials, different magnets etc. Try a different circuit - it may have a more suitable phase response... mustn't forget that the guitars setup - action, strings, pickup - can have a big impact on success. The main thing is though, that if you're in the ballpark, you will get some sustain on some strings - even with a slightly shoddy coil, a fetzer/ruby, and a non-ideal core (The difficulty is getting from there to a nice system with a strong even response and reasonable battery consumption.) So I'm thinking that there is something more fundamental wrong with the Sabudum's project. cheers Col

8.8 ohms is a touch high, but you should still get some sustain on some of the strings, so I think the main problem is elsewhere. more questions: Are you sure you have the magnet positioned correctly - ie. with the poles vertically in line with the driver core? If you have accidentally put it at 90degrees, it will not work. Are you testing the driver by holding it in your hand as close to the strings as possible without touching ? Try it near the 12th fret - it will sustain more easily there. Have you tried reversing the polarity of the driver? Either reverse the connections, or turn the magnet over Once you have checked these out, the next thing is to suspect that there is a problem with your drive circuit, or the connection between it and the driver coil - e.g. are you sure all the caps - in particular the output coupling cap - are the right way round (correct polarity)? Double check that all the components have the correct value. swap out the LM386 for a new one (you did use a socket?).. etc. If you have a breadboard/plugboard, build a new version of the circuit on the breadboard to see if you get the same results. cheers Col

You need to measure the actual resistance of the actual coil with a multimeter. Theory is great for designing stuff, it can tell you what should happen, but when the thing doesn't work, you need actual real world measurements to tell you what did happen. Measuring the coil's resistance will tell you if there is a short somewhere, or maybe if there is damage causing a much higher resistance.

there's 113 turns to the coil, in order to have 8 ohms impedance The resistance is 1.378 ohms and resistance per meter is 0.549 ohms/meter Which resistance is 1.378 ohms? that should be fine. I'm concerned about the resistance. with Pete's design of drive coil, the impedance at low frequencies is dominated by DC resistance, so DC resistance needs to be about 8ohms. if its a lot lower, it could cause the LM386 to fry, or go into shutdown protection (if it has that?). Where is the 1.378 ohms figure from? cheers Col

how many turns on the coil? what is the dc resistance of the coil? how is the magnet attached to the core?

You say you used a single coil dual blade pickup - this doesn't make sense. Do you mean a single coil 'sized' humbucker that actually has two coils? or a single coil pickup that has a strange dual bladed core? If it's a single coil driver, then winding to Pete's specs should have given you some definite sustain action - even if not ideal. There should be something to work with. If it was a 'rails' humbucker style with two coils, then it is very important how you connect the coils. Assuming the pickup used the standard humbucker approach of a single bar magnet, at the bottom between the two rails, then the coils need to be wired out of phase with each other, so that they push and pull at the same time. If you wire them in phase, one will push while the other pulls == no sustain ! Whether you wire them in series or parallel is also important - if you would each coil to 4 ohm nominal impedence, you should wire them in series, if you would them to 16ohm each, then wire them in parallel (If you would them to 8 ohms, then oops!... I recommend wiring them in parallel and wiring a 4 ohm half watt resistor in front of them to bring the overall impedance up to around 8ohm). As has been suggested by everyone else, you need to post some photos and details of the specs of your driver. The wire gauge you used for the coil(s), number of turns, DC resistance of coils, how they are wired, the more detail the better. good luck Col

Recently got my hands on a scope. My circuit as is has a problem with parasitic oscillation on the virtual ground. To fix this I swapped out the LM324 quad op-amp with a TL074cn. The LM324 cannot sink enough current at it's output to provide a stable virtual ground, the TL074cn seems ok. TL074cn might not be the best chip for the job - it was just what I had in my box - I guess any that can handle 40mA at the output will do. FWIW, this fix didn't have any noticable effect on the functionality! It worked fine with the oscillations (they were at about 1Mhz). Might improve the efficiency a little, but I doubt it as most of the juice is used by the LM386 which doesn't use the virtual ground. cheers Col

LOL :-D awesome clip forget AGC and bi-lateral drivers, all we need are two hydrocoptic marzel veins and a drawn reciprocation dingle arm !

that extra 4 ohms will be significant! DC resistance is only part of the story, a more important parameter is impedance, which in the case of a sustainer is a combination of DC resistance, capacative reactance and inductive reactance. The problem is that most folks don't have the kit to measure the inductance (from which impedance can be calculated), so we go with a DC resistance reading. The capacitance is easy to adjust - just tweak the output cap. The reason why the baseline driver spec works is because the coil inductance, resistance and the output capacitor together form an LRC circuit in which the resonant frequency is near that of the open high E string - the most difficult to drive. Additionally, the bandwidth of this circuit is in the region of 1000kHz - exactly what we need to cover most of the guitars range. look here http://www.calctool.org/CALC/eng/electronics/RLC_circuit here's another really good calculator: http://pr.erau.edu/~newmana/imped.html plug in a resistance, capacitance and inductance, and this one gives you overall impedance, and phase shift ! If you change any of these parameters, you will have to find other ways of tailoring the drive to match the response of the guitar to the sustainer. Ok here's the point to all this babble your extra DC resistance equates to about 50% more turns than the required spec (either for Pete's coil or the twin coil I use). Inductance increases with the square of the number of turns, so if ~126 turns would have given you 8ohms and an inductance of around 1.2mH (assuming the correct wire guage and a suitable core), 190 turns would give around 2.7mH, Plugging 12 ohms and 2.7 mH into the first calculator, the resonant frequency is now about 220, and more importantly, the bandwidth is reduced to 700Hz (If you reduce the DC resistance to 8ohm and keep the inductance at 2.7, the bandwidth is reduced further) The resonant frequency can be corrected by switching from a 220u cap to a 100u, but the bandwidth is not so easily sorted! To demonstrate the effect of the reduced bandwidtch, lets use the second calculator I linked to to look at the impedance at the extremes of the range we are interested in ie. 82Hz and 1000Hz: with a 220u cap, 8 ohm and 1.2mH: the impedance at 82 Hz is 11.5ohms and the phase shift is -45 degrees the impedance at 1000 Hz is 10.5ohms and the phase shift is 40.5 degrees with a 100u cap (better match for 2.7 mH), 12 ohm and 2.7mH: the impedance at 82 Hz is 21.65ohms and the phase shift is -56 degrees the impedance at 1000 Hz is 19.5ohms and the phase shift is 52 degrees (at ~300 Hz, both have an impedance of close to their DC resistance, and phase shift close to zero) So you can see that in addition to your extra 4 DC ohms, the extra inductance, pushes the impedance up dramatically at the edges of the required frequency range - beyond what an LM386 can effectively drive. The phase distortion is also worse meaning that less of the energy getting to the strings is actually helping drive them, due to the phase shift. Anyway, enough. My point is that the difference in number of coil turns can make a significant difference to the success of the device. hmm, it almost sounds as though something is broken - a bad solder joint, or some fault in the coil ? It shouldn't be that sensitive to 'fiddling', and it shouldn't jump between not working at all and working quite well... anyhow, keep at it do you have a breadboard ? if not get one, they are not expensive, and make circuit tweaking MUCH more practical. Col

thats great. good luck. do you have a breadboard to build it on first ?

The specs are pretty much: .23mm wire wound to ~4ohm on each coil. steel/iron cores. cores 2mm thick, 10mm deep 56mm long (could have been a bit shorter... but that would alter the inductance for that DC resistance, so stick with 56mm for now) coils were wound to give as near to a square cross section to the wire bundles as possible to maximize magnetic coupling, so they are ~3mm thick and wide (4mm would work as well for sure. there is 1mm of core above the top level of the coils. the wiring as I said is series. And, although I'm not sure if they are in or out of phase, that is something you could test after construction - keep your options open. here's the driver up close and personal: http://i100.photobucket.com/albums/m15/col_012/angleddrivercloseup2.jpg and here's the circuit with a pick for scale http://i100.photobucket.com/albums/m15/col_012/toviewwithplectrum.jpg If you just want a working sustainer, you would be better buying a commercial unit. cheers Col

You need a voltage reference to attach to Vr I used an op amp stage to create the voltage reference - maybe a simple potential divider would work here ? here are links to info: tangentsoft.net/elec/vgrounds.html http://www.swarthmore.edu/NatSci/echeeve1/Ref/SingleSupply/SingleSupply.html you should get slightly better phase response at low frequencies by increasing the input cap to 220n.

?? no circuit? Ok, Col - do you have a confirmed driver design spec that marrys with your circuit as posted further above? (it looks a little complicated for my first attempt, but I've only a limited amount of time I can spend on this and want to go with a full confirmed design (ie driver coil *and* correctly matched circuit). No, I don't! And have never claimed to, although I will give you specs of the driver I used. To my knowledge, there has never been a DIY sustainer system complete with publicly available driver and circuit schematic that has been proven to be consistently successful and repeatable. That's one reason why I (and others) get annoyed with Pete. My most successful setup used the circuit posted, combined with a dual coil driver. The driver has two coils of a bit under 4ohm each wound with 0.23mm wire around steel(flat iron according to B&Q). These were mounted using an alnico magnet from a humbucker. They are wired in series to give a total DC resistance of just under 8 ohms. I think that they were wired out of phase (it was a long time ago), so that they pull together magnetically. This setup works for me better or as well as any others I've tried - significantly better than when using the driver I built to Pete's specs. I get good sustain on all strings, open and up the neck. There is very little fizz (you might get a very small amount on the lowest notes if you set the AGC up to be as responsive as possible by tweaking the trimmers. Unfortunately, the one other person I know of who tried to build it seemed to have some problems with the circuit - there were language difficulties, but he was clearly not happy with the results he was getting. The biggest issue with it and with other systems I've tried is battery usage - there's no way I could recommend this as something to use on your regular axe because it eats through batteries (the simple circuits without AGC are worse in this respect) Personally, I do not promote my system as something that everyone should make - there are no guarantees. I just put up all the details of my full system as info for others working on their own sustainers. Don't build it unless you are ready to debug it and make changes to get it working. You also need a pretty big cavity in the guitar to put it in as its a big circuit. Potential problems: One problem is that some of the component parts are not easily repeatable in a DIY sustainer. The core material is important, and that is usually pretty random - everyone uses what they can find handy or local. This can cause a difference in inductance, hysteresis, and eddy current losses which will affect the functionality. The pickup on the guitar is another major variable - it's sensitivity to EMI, and the frequency response will have a large impact on the success. My sustainer setup uses a Seymour Duncan JB which is a high output humbucker. I don't know if this is a good choice because I don't have lots of different pickups lying around to compare with. (For a long time Pete claimed great success with a single coil bridge pickup feeding the driver. It became apparent later though that for a lot of his successful tests, he was using a stacked humbucker! I think now that his system works well with all pickups?) The setup of the guitar is crucial - I like to play with a high action which is bad for sustainers. The magnetic pull of the driver falls off with the square of the distance - so if the string is twice as far from the driver due to high action, the pull from the magnet will be a quarter !!! My circuit uses a fet based compression stage. Fets are notoriusly variable, so getting that to work could be tricky. It worked for me, but that might have been pure luck. Anyway, personally, I don't like the way these magnetic sustainers effect the guitar tone. My initial motivation for building one was to try to get a loud guitar feel when playing quietly. Unfortunately, due to the phase gap between driver and pickup, a sustainer actively damps the guitars natural sound. It reverts to a sinusoid much more quickly than it would naturally. This is just the nature of the beast, some folks don't mind it, others love it. Another example would be pinch harmonics and similar techniques - magnetic sustainers kill them. When I discovered this, and then through research understood why it happens, I pretty much gave up developing my version of the sustainer (I can go into a more detailed explanation of this effect if you are interested). So I'm still interested from an electronics hobbyist point of view, but not as a guitar player. You are welcome to build a copy of my design in the knowledge that it might not work for you. If you try and it doesn't work, I won't have a go at you and claim you didn't follow the specs closely enough Just to reiterate, There are no publicly available designs for sustainers that have been proven to be repeatable and successful in a variety of guitars. The people who do claim to have repeatable successful systems, have refused to make them available. I'm sure they all have their reasons. So you will have to get your hands dirty, and try things out. Good luck and enjoy. Col

indeed!! That doesn't wash - there is no such thing as a 'generic' sustainer drive circuit. You should publish specs and docs for a complete sustainer system, or you should stop saturating these discussions with huge multi-page slabs of pointless dogma (I didn't read the last WOT, life's too short - I'll read posts if they are concise and relevant. Why not put all the other stuff in a web-page and link to it rather than repeating it all endlessly ?). cheers Col

The ruby is presented as an audio amp - not as a sustainer amp. The Fetzer is presented as an impedance buffer for audio applications - not as the input for a sustainer system. Someone other than ROG cobbled the two together and presented it as a sustainer circut. I know it wasn't you Pete, and I'm absolutely clear that you don't endorse it!. However, you do often attribute the fetzer/ruby to ROG and then criticise it - implying that ROG have created a poor design. I imagine that if ROG did design a sustainer circuit, it would be a whole lot better than yours or mine. Fact is: on the one hand, you claim - at very great length - to have presented a simple working sustainer system and criticize folks for deviating from your design. on the other hand, you have only publicly presented one component of your system - the driver. Your circuit remains private. You seem to be misleading people, so you need to make it clear that your 'DIY sustainer' is not - it's a 'DIY sustainer driver' As such, it's not a ready to go project for folks to make - it never was. It seems that most folks (not me) don't want to 'do their own thing'. They want do do your thing - that you promote endlessly. Unfortunately they can't because - contrary to the implications of all your endless posts on the subject - your thing has never been published! Amazing really. You can't expect folks to stick to your design unless you give them the design to stick to - they can't read your mind !!! cheers Col

The LM386 power stage is part of various amp circuits - including ones I have used. It is not useful by itself. You need at least a buffer. Unfortunately however, it's not as simple as sticking a buffer stage on the LM386. An important issue for a sustainer when trying to find and use a pre-existing amp design is phase response. When amplifying audio for listening with a speaker, it doesn't matter a whole lot if the phase at the output is significantly different from the phase at the input. So, you won't see phase response graphs for most amp schematics - it's not an issue for most applications. Unfortunately, for a sustainer, it really does matter. This is a reason why the fetzer-ruby is not particularly good in this project, it's phase response is not linear enough over the required frequency range. The component values for buffer stages, gain stages etc. need to be chosen to minimize phase distortion. Another important point is the gain. 200x is way too much. This will cause heavy clipping which in turn will create lots of unwanted noise (aka fizz, grunge). If Pete would publish his existing, working, simple LM386 based circuit that is already tweaked to match his driver specs, everyone could just build it and not have to worry about trying to 'design' their own. They wouldn't have to worry about deviating from his design! cheers Col

Unlikely, that's just a variation of the schematic from the app note in the LM386 datasheet. It has no input buffer for starters, so connecting a guitar pickup to the input would not work well. cheers Col

ROG didn't design the 'fetzer ruby', and should not be held responsible for it - their designs are top notch IMO, and it's very unfair to criticise them for something they didn't design or endorse. Sorry Pete, going to have to pull you up on that one. Your 'design' is not (to my knowledge? correct me if I'm wrong) publicly available, so to suggest that it is and imply that other peoples problems are a result of not following that design is unfair. A functional sustainer is a system. All parts of that system are equally important to the success of the project. The driver is important, but the circuit is equally so and must be matched to the driver. According to you, your driver specs work well with your circuit - which is great. unfortunately, your circuit has to my knowledge not been published, and as you so often repeat, you do not endorse the fetzer-ruby, so why do you continually tell folk to follow your design ? They can't, it's not available! A wall wart based circuit will 'hard-limit' voltage peaks similarly to a battery. If your circuit is current limiting due to the battery, then that is a problem, not a feature - battery life will be very poor. Interesting, I'll have to take a look at the gallery cheers Col

The most likely reason you are having a problem is the inductance of your driver. here are the basics: There are two really important aspects to getting this to work (there are other important things as well, but these are paramount) Impedance and phase ==================== Impedance --------- Impedance is like resistance, but more difficult - go do some research Impedance is a combination of resistance, capacative reactance and inductive reactance The main point here is that impedance due to inductive reactance increases with frequency. A higher inductance gives you a more powerful magnet..... great. A higher inductance means higher impedance..... not great. While the resistance is the main part of the impedance, things are ok When the inductance is big enough that the reactance becomes the main part of the impedance, things go pear shaped for us. To cut a long story short, if you are using a 8 ohm coil and driving it with an LM386, you want the inductance to be about 1 - 1.2mH. If it's much lower, you won't get a strong enough magnet. If its much higher, the impedance at higher frequencies will act as a low pass filter cutting off the drive to the higher strings. Phase ----- if you're trying to drive the string up when it's going down, and trying to drive the string down when it's going up, it's not going to work well. You want to be driving the string with it's natural vibration, not against - ie. 'in phase' with the string. Unfortunately many circuits designed for audio - particularly low end cheaper or simpler ones - don't have an ideal phase response. This means that they will be in phase at some frequencies, but a bit or even a lot out of phase at some other frequencies. Additionally, the driver coil and the coupling capacitor that connects it to the drive circuit will have an effect on the phase. The physical gap between the pickup and driver on a guitar also creates a phase difference - ~90 degrees when fretting at the 14th, around 43degrees on an open string. The most difficult note to drive on a guitar is the open high e string (and the first few frets depending on action), so ideally, we would have the phase shift of our whole system at about 90 degrees for this frequency ~300Hz... Unfortunately, unless you have a pretty good grasp of electronics and maths, or you have a scope, or means to simulate your circuit, you won't be able to calculate or measure the phase of your system, so your only option is to try lots of circuits, drivers and coupling caps (220u or higher should be good for the cap), until you get something that works. Good luck and don't give up Col

One idea I've mentioned in the past that could help here is to create a sandwich stack. (yep, more brainstorms I'm afraid) You'd have 3 coils: one main 'pickup' coil - a low powered pickup coil, but still many more turns than a driver coil two small driver coils, each identical, one on top of the pickup coil, and one on the bottom. The idea is that you feed an opposing drive current to the lower coil. The symetry of the two driver coils with respect to the pickup coil should cancel much of the drive signal current that would otherwise be induced in the pickup. At the same time, the asymetry with respect to the strings should allow the top driver coil to still drive them and provide sustain. (there was something tricky going on with the core structure as well, but I don't remember the details exactly) The main problem with this idea was that the manufacturing tolerances required make it impractical. I had wondered if a microcontroller could be used to tweak the balance to the driver coils in real time in order to prevent oscillation, but I doubt it would be possible (I have pages of notes somewhere). However, if the drive signal was being alternated with the pickup signal, this stack format might help - it may allow the system to function at higher frequencies due to the cancellation of the drive signal in the pickup. ----------------------------- For whole idea to work, you'd need to sample at 2.5KHz or more. As much of the clock cycle as possible would be spent driving. As small a 'frame' as possible would be used to take a pickup 'sample' from each cycle. As the sampling frequency rises and the sample frame size is reduced, the inductance of the driver coils must be reduced in order to allow the driver coils to react fast enough to prevent bleed-through - the cancellation of the stack helps here. The lower the inductance, the more current you need to provide the same magnetic field strength. The bit I'm least sure about is: Does the cancellation property of this sandwich stack idea mean that the relatively high inductance of the pickup coil doesn't slow the response of the magnetic circuit down? cheers Col

hmm... thinking out loud time... does the signal in the moog go straight from pickup to amp, or does it go through some dsp that 'reconstructs' it ? if the pickup signal is only sampled while there is no sustainer drive present, then the switching frequency can be removed as it is a known quantity. The 'holes' in the output could then be removed either with filtering or some sort of dsp using interpolation or some such... I suppose the crux is how high the switching frequency can be. I guess this depends on how quickly the field of the driver/pickup can be collapsed and reconstructed. For the system to work with an analog output - like a traditional guitar - the frequency would have to be high so that any switching noise can be filtered out without impacting the guitar tone....