col

As requested by fookgub, here is a schematic of my latest circuit with readable component values. Col

Since you have already six op amps, another two would mean you could use two quads for the same price...just a thought. maybe the monetary price is the same, but in terms of board real estate its not. I think a switchable input cap will be adequate. The mixed mode does highlight a limitation in this new circuit that would not exist if I could do what I defined as my ideal circuit and have seperate independant low and high thresholds. In mixed mode, the low E and A strings have very poor response - I understand the cause and without a re-design, there's not much that can be done about it. I don't think it will be a problem when playing though. Yes, and one of the best things about my original circuit was that the straight through sound is so clean and strong. I decided to go the other route this time partly to save an op-amp, but also so that if the battery runs out, you can still play. In a phantom/remote powered version I would definately go for a buffered through sound. One downside here for me is that I don't see any standard USB cable coping with the punishment I deal out to my guitar cables - I like to jump around while playing Rock'n'Roll Thats an issue for any system that uses non-standard connectors and/or cables - can it survive a Rock'n'Roll lifestyle ? I think what many many players don't understand is that there's a lot more to technique than speed. Good tone comes from your fingers - its that side of technique that is most difficult to master, and it's there that you find your own voice - playing fast can disguise deficiency in this area up to a point, and thats what a lot of players do.... guys like EVH, Hendrix, Page, Angus young, SRV, Brian May, Malmsteen have/had great technique, their tone was expressive and unique, how fast they could or couldn't play is always going to be secondary.... The problem for many young players is they only realise this when they get to the point where they can play as fast as their Hero after many years of speed drills, only to discover that they still don't sound any good... Unfortunately lots of them think they do sound good cheers Col

OK, I got a better 'mixed' mode harmonic up and running - just switching the input cap gives a pretty good result - will try a few values to see which are best, and if it's worth having more then 1 'mixed' setting. I put in a proper on/off switch, and the little click I described is still there - will do some fiddling to try and remove it. Col

That's a real break through, at least for me, and you have the SPDT harmonic switch in there too Yeah, I would imagine the lack of a pop is due to the equally balanced dual coil driver. Oh, and its a SPST switch Which strings it is depends on the guitar, hardware and sting guages used - the original (crappy) pickups that came with my guitar actually had a notch cut out of their rails under the G string, so I guess its a feature of this guitar. I get multiple strings sustaining, but whatever note is being played by the G string - or D if G is damped - becomes the dominiant tone. This is pretty much what happens on my guitar when playing loud with real feedback, its not unpleasant, so probably not too big a deal. It could be partially be balanced by adjusting the poles on my pickup, but I figure the only real solution would be to have some sort of biasing built into the driver core. I hope so, but as they (used to) say: The proof of the pudding is in the eating ! I'm not very happy with the full harmonic mode sound in general - it's too extreme for me, more of a special FX type thing. As far as Mixed mode, the trick is to use filtering to boost the frequncies where the harmonics you want are located... I tried doing that by switching a cap in the AGC part of my circuit, but thats not a good option because the respone of the low strings drops right off. What I need to do is put a bandpass or highpass filter immediately before or after the initial pre-amp stage. This would mean that there would be more high frequency content in the tone, but the AGC would bring any low level signal up to a usable level solving the dead 5th and 6th string issue. Problem is that that adding a seperate filter stage would mean at least one more op-amp... So what I really want to do is incorporate some sort of switchable filter into the pre-amp stage - just havn't done it yet. Personally I believe that its not being marketed correctly - they should be aiming it at the home recording crowd and selling it as a way to achieve a cranked amp tone without annoying the neighbours... Theres one thing a pod can't do - bring the strings alive - this system fills the gap. most DI units and digital fx boxes take the alive raw unpredictable quality out of a cranked guitar tone - which is crazy because it is the most important thing about that kind of tone... Another thing, in rehearsal studios I've played in, the guitar often gets swamped by drums and bass, getting a nice feedback in those situations is not easy... enter the sustainer... One potential reason for lack of enthusiasm might be that shredding is coming back again. Shredders never let a note ring long enough to hear the guitars natural tone let alone long enough to hear a sustainer at work Col

Yep, still on the breadboard, I'm still fiddling with component values and trying to come up with a better way to get 'mixed mode' to work as I want it to. My understanding of the stack as a pickup is that the two coils cancel each other - e.g. they fight rather than help ! This works reasonably as a pickup because the lower coil gets much less string signal than the upper one because it's further away, so it doesn't cancel too much of the sound. Seems Kinman and to some level Dimarzio have both had success developing a custom lower coil that is designed to pickup as little of the string as possible while still getting an equal dose of noise... It may be that we could do something similar with two very different coils in order to try and make the lower coil use as little power as possible while still pulling in stray magnetic flux to control EMI. Its always going to be less efficient than either dual rail or bi-lateral EMIbucker designs though. Can't really comment on that because I don't know how they work or how ther are constructed side to side would be no more natural than up and down. Also, my experience is that you get a sustain tone that is plenty loud enough long before the string vibration has as much of an impact on action as other factors such as driver construction and mounting, magnetic pull etc... However, I would be interested in hearing if there was any tonal difference between up'n'down and side-to-side. Seems good - my setup is not good for testing clean tones as I've already explained, and I'm hesitant to post examples because there is so much bad clipping from my cab sim that it would be difficult to tell where fuzz and distortion was coming from. I'm pretty sure by A/B ing clean tone with sustainer on and off that the fizz is much lower than it was with my other circuit if not completely gone. Its at the stage now when the most annoying thing is string to string balance - this time in chords ! the AGC evens up the balance well for single note stuff and rock chords, but when playing clean chord and letting them ring, the 'better' strings (G and D on my setup) tend to overpower the others - this is more or less noticable depending on chord voicing, but I know through messing around with sustaining chords that if we could make a driver that had better string balance, the thing would sustain clean chords a treat (I guess its a good sign that this is one of the big issues for me now) I suppose options for balancing would be pole pieces with adjustable height, or adjustable thickness - ideally these could be adjusted or replaced after manufacture... maybe using a couple of Tims groovy coreless coils. Yes, my coils came out well, but the magnet I'm using is from a standard sized humbucker, and that is the limiting factor as far as overall width is concerned - there is a gap between the coils. I suppose that also good really as many folks should be able to lay their hands on a magnet from a standard sized humbucker - it may not be ideal, but at least we know it works. Yes, with the socket turned 'inside-out' to give even more room. Ah, yes, I did some more pulling out and pushing back in of the power lead (both while sustaining and not) on my breadboard, and I can't get any pop or thwop or pdfffdpfffpfpfpfffff type of sounds all I get is a small faint 'tick' that is most likely caused by a tiny spark because I'm not using a switch for the test. I'll say again - NO POP (also, no neck pickup and no single coil pickups - but I don't think thats an issue) cheers Col

Yeah, look at that circuit I posted. The op-amp nearest to the LM386 is the phase inverter. If you look closely, just to the left of it there is a SPST switch (closed) marked somthing like 'Key = SPACE' That was my initial reaction, but I explained way back, that when I tried going back to the single core driver I discovered that the dual rail DOES make a big difference to the level of EMI feedback and fizz. I am interested in trying a bi-lateral driver at some point. I'm not so optimistic about a stack though - it is much less efficient, there is quite a lot of self cancellation inherent in the design of a stack - maybe if it was used with an off-board mains supply it would be a good option (it that what Brooks uses?). Well, I would say that it works a lot better than the simpler design. I am actually pretty much satisfied with the results I've had in the last couple of days, and if it works as well or better after installation, I may spend some time developing playing techniques for it rather than product research The only downside is that it isn't a universal solution yet because of the size of the circuit. There is plenty of room for it in my guitar, but many others will not have the space. As most of the circuit is using very standard parts it would be easy to minaturise it. The only thing that could be hassle is the FET, it depends how close the specs of any available SMD FETs are to a J201. If they're not close enough, then there would probably be a little tweaking required to the circuit. Edit: oof, of course J201 is available as SMD Basically, apart from pots trimmers, switches and a few caps, everything in the circuit could be miniaturised - small enough to sit alongside the driver if required. Probably talking about 1.5 square inches... have to find out how much a small batch of manufactured circuits can be had for. Looking at the Schematic: The one at top left is to provide a stable voltage reference at half the supply voltage. The one just below that and slightly to the right is the input pre-amp this sets the threshold level. The pair at the bottom comprise the 'precision' rectifier - I tried a halfwave rectifier with just one amp, but it didn't provide nearly as even a frequency response particularly at the lower end. The op-amp just to the left of the LM386(rightmost IC) as explained above is the inverter. The last and possibly most important, right in the middle at the top is the AGC amp, its feedback loop has a tap off to the FET that changes the amps gain depending on the level of the signal produced by the rectifier.... so a loud input causes low gain while a soft input causes high gain... Thats what they do The only places where you could conceivably reduce the op-amp count are # The rectifier - although I think this will reduce performance noticably # The half supply reference - this could be done with a transistor, but it wouldn't work as well, and would require enough extra parts that it would probably not save board space. # The inverter, you could remove this if you don't like the extreme Harmonic mode. It may be possible to remove a few resistors here and there (e.g. the voltage divider that tweaks the dc offset of the rectifier), but without completely re-designing the circuit, all you would do is seriously reduce the effectiveness of it. I've already stripped as much as possible out of the 'Fast FET Limiter' circuit that I based the FET section on, and everything else is already pretty much bare bones. In the end, it should work out at about the same size physically as my old LM13700 circuit. cheers Col

Battery keeps things simple and safe, but I think that the main reason for resistance is that most guitartists like to keep things simple. Using non-standard leads would be a big issue for many to accept - not being able to just grab any old lead and plug your axe in... also, not being able to just go to the local shop and get a new lead when required, or borrow one from another guitarist.... If it was just the power that was offboard, then it a bit simpler, but still not "plug n play". There are already enough things stopping folks from considering a sustainer - installation, potential modifications to guitar, possible change in tone, issues with pickup selection etc. Adding non-standard connectors to off-board units, just adds ot that list. yep, you're talking about a full-on umbilical cord for that Yes, I've been thinking about this, its the only reason I've not already decided to have no knob on the front of my guitar - might need to adjust the output or threshold as the battery gets low. I'm sure that with small scale SMD and class-D amps, some sort of regulator could be included to keep the supply to the circuit stable until the battery dies, but I wouldn't want to try that using chunky old pin-through-hole components. I'll get an annotated breakdown of the circuit ready soon. As far as refinements, I don't think it will be possible to make any major simplifications without compromising the functionality, althout swapping two of the dual op-amps for a quad will help a little. The main thing about this design is that the parts are readily available - I used a J201 FET because they seem to be easy to get, and are used in most of the FET stomp boxes I've seen - hehe RunOffGroove should do a licencing deal with J201 manufacturers. Thanks, although I don't know Heart, so not sure this is a compliment . The main thing I wanted to show is that this circuit gives a good sustain volume, and that my extra wide dual rail provides fine harmonic performance, in fact it may be better because it doesn't get so exruciatingly high pitched as with the single rail driver. This demo highlights another 'feature' that I'd not previously considered. The AGC circuit causes a longer hold before the harmoinic takes over, this is because the driver only fully kicks in when the guitar signal drops below the threshold level - so if you play soft, near instant harmonic, if you play hard - particularly with chords, you get some fundamental before the harmonic takes over. btw, I have a simple mod to my circuit to give a basic mixed harmonic mode - It's pretty subtle, but it does make a difference from fundamental mode - will do another demo with some clean sounds and 'mixed' mode great news, looking ofrward to hearing the results. Yeah, there are so many extra design variables and decisions with an off board unit that it kinda makes sense to wait until the on-board system is more mature, then work from that. As far as batteries are concerned, with a good AGC in the driver circuit, battery drain is reduced dramatically. If we can also ultimately swap the LM386 for a class-D amp chip, we should be able to further reduce power requirements - to the point where they're not much worse than some 'normal' stomp boxes. Personally, I would rather spend time on this avenue than developing an external solution - but then, my guitar has a big cavity (ooer missus... fnar fnar) cheers Col

Are they lower or higher than you would expect? They are lower than the figures I had from the simulation, but I didn't believe those anyway. Also with this circuit, there are a range of figures depending on how hard you want the sustainer to drive the strings - the figures I suggested are when the unit is producing a good sustain, but it can go louder or quieter. I obtained these figures by putting my meter between the +ve batter terminal and the +ve input to the circuit. The meter was switched to DC mA. I suppose I could get a more accurate valuse using a combination of DC and AC readings ? Disclaimer: as usual, this is in development, it is not finished, do not build it unless you are ready to tweak it to your requirements, and don't blame me if your guitar goes up in smoke or anything else bad happens Sorry its hard to read the values, can't get a better pic without using ExpressSCH and re-drawing the whole damn thing.. or a bunch of other fiddling that I don't have time for right now As I noted, I've still to work out the best way to do 'mixed mode' harmonics

What you really should do is take the line to the sustainer directly from the pickup lead before the volume and tone. I would also recommend some sort of pre amp between the guitar and soundcard. Not sure that this would cause the problems you described, but who knows for sure Col

What are you plugging it in to? The standard way to set it up is to split the signal from the pickup passively - take two wires from the pickup, one goes to the sustainer circuit, the other goes to the output jack and from there to your guitar amp. If you have it set up this way and its not working, then, assuming that your sustainer circuit is OK and battery is ok and connected, the first thing in the chain from the guitar to amp is probably a 'tone sucking' demon - if you have any stomp boxes or other FX between guitar and amp, try removing them and testing again. Col

I added another sustainer demo to my soundclick page - this time distorted rock guitar in harmonic mode col

I uploaded another demo clip to my soundclick page this time its a crappy rock style jam with computer drums... using my runoffgroove distortion box and cab sim, sustainer is in harmonic mode throughout. the clip is called 'rawk sustainer demo'. cheers Col

Headroom in the driver circuit seems fine - I've actually taken advantage of the limited headroom provided by the 9v/opAmp situation, I've set it up so that any input signal above around 300mV will be clipped in the input stage (well above the top threshold), this is to put a cap on the maximum through signal. maximum attenuation of th FET is limited, and without this clipping, louder signals > 400mV were being amplified close to the level of the desired 'correct' signal. The Headroom issue I am having is not sustainer related. When I plug my guitar directly into my RunOffGroove Condor amp sim, it doesn't have enough headroom and louder playing causes clipping. This means that I can't be sure that the fuzz is being caused by the driver circuit or the amp sim. In my earlier circuit, it was definately the sustainer, because it happened more in some quieter passages. The phase issue you're talking about won't have an impact on the crosstalk noise 'masking' I mentioned. What it does seem to do though is lengthen the attack part of the sustainer action. Some notes fight more than others before relenting and allowing themselves to be sustained. Sometimes theres a little dip in output as the sustainer takes over from the natural string vibration - I think this is caused by the phase difference between the driver position and the pickup. If the signal at the driver gets close to 180º out of phase with the pickup, we get the harmonic mode effect, I find this is most noticable on the high E string - maybe because it is such a wimpy weakling and can't put up enough of a fight The only way to improve this is to have the driver closer to the pickup. The issue will still exist, but it will probably not have so much impact because it will effect higher frequencies. It may be that with a good enough 'crosstalk masking' circuit combined with a dual rail (or bilateral) driver and possibly some shielding, we could get close to the holy grail of a mid position driver - if so, this would go a long way towards reducing any problems caused by phase difference between driver and pickups. I remember in my handheld driver tests with the dual rail, as I moved closer to the bridge, the fizz increased, but there was no squealing feedback until just after the mid position... however, handheld above the strings is not the same as fixed beneath, it just gives a rough idea, and the crosstalk noise was MUCH worse nearer the mid position. I still think that a driver in mid position is going to intorduce more problems than it solves - oh, and I don't have a mid position on my guitar My take is that in an ideal world, we could use the neck and/or bridge pickup with the sustainer in order to get a better range of tones. In reality, I would rather have one very clean pure tone than a range of noisy fizzy ones (col attempts to convince himself not to buy a router and attack his guitar... and to be happy with what he's got) cheers Col

Yep, I remember setting up the fetzer/ruby and not being able to get the bias anywhere near the 4.5v I was trying for. I was assuming that 4.5 was the target because thats the usual figure in fet stomp boxes... it ended up at about 2.2v as you say.... congrats Now the real 'fun' starts set the vol and gain as low as you can while still getting acceptable results. As far as the 'singing' driver, I got that at first - others have noticed the same thing, I think its just a sign that the vol/gain are too high. The whistling is parasitic feedback caused by the pickup sensing the driver directly instead of via the string - something to avoid.... you have to keep the driver as far away as possible from any in-use pickups. In Other news... I finally got around to breadboarding my version of the Fet attenuator - results are very good. Fizz is almost completely gone, and I have a hunch as to why. The circuit I'm using is based on a combination of the Fast Peak Limiter circuit and the Improve FET Gain Control circuit. I used the Fet circuit from the Fast Limiter because there is very little distortion and it can be set up so that the response stays linear right through the guitars frequency band. Hunch #1 I have set this up so that the agc causes very little distortion above 82 hz. I noticed that when I set it for a faster attack response which causes distortion at lower frequencies, I get fizz at lower frequencies... hmm Also, I checked my LM13700 based circuit, and right enough, It has loads of asymetric distortion particularly (but not exclusively) at lower frequencies... (not hard clipping, more of a soft fuzzface type and funnily enough a fuzzface style fizz). There seems to be some slight fizz noticable during transitions... after you hit a chord hard and it starts to fade, there is a little fizz as the sustainer kicks in - furtunately when the circuit reaches an equilibrium, it cleans up again... This may just be my amp sim though - my pickup is too hot for it :-|... I really need to test with some other kit that can produce a predictable clean tone with plenty of headroom... So it _may_ be that if the driver signal is in phase and has little or no distortion, any bleedthrough/crosstalk is unnoticable because it is nearly identical to the guitar signal. Hunch #2 it just pure luck Anyway, I got my positive vibe back again The current drain is back up to about the same as my first agc circuit ranging roughly from 35 to 60 mA... 13mA at rest. It's back up a bit because I boosted the volume so that the effect isn't so subtle. I incorporated a phase reversing stage (uses a SPST) for harmonic mode - this causes some squeal when you damp the strings, but that should be fixable by getting everything set up just right... gain, sensitivity etc. There is no through buffer so if the battery goes dudd, you don't lose sound. The circuit uses 6 op-amps and an LM386. so either 3 dual op Amps or 1 dual and a quad. It might be possible to replace a couple of the op-amps with transistors, but it probably wouldn't make the board smaller, and would almost certainly make the thing work less well. Just for the record, I'm using my dual rail driver with no shield. it is connected via about 50cm of home-brew 'inter-8 weave' cable This new circuit does a superb job of fundamental mode sustain - chords are great and single notes and/or harmonics sustain very well accross most of the neck... The maximum output volume is controllable via the threshold knob, although I may just keep that as a trimmer. there are still one or two frets on the high E string that go to harmonics because the just don't want to play ball... I still need to work out a simple and good 'mixed mode' to get a good harmonic bloom without going into full harmonic mode. I guess this will just be some switchable filter caps in the right places, just not got around to that yet. When thats done, I'll post the schem. @pete I tried to get a power off 'thump' by just pulling the power supply lead from the plugboard. No thump, just a crackly click... probably the click would be gone if i used a proper switch - will try that soon cheers Col

Very interested to see how this turns out - I have some sketches of the same 'horseshoe' idea, but I couldn't come up with a practical way to construct it that I thought would work out.... how to setup the magnets was the biggest issue, but there is also the problem of winding and potting evenly on a curved bobbin. good luck, Col

If it came down to it, you could turn the gain right down, this would cause the sustain to be 'switched off' (effectively), and would reduce the power drain to something reasonable - 7mA for curtis' buffer/LM386, 11mA for my old LM13700 circuit... Still not done any conclusive tests, however, I've not noticed a pop or ffdphhhfffddddp sound, and I normally notice those kindsa things. Mind you, I'm just running a bridge humbucker and a dual rail driver. I think we can experiment with some smd stuff without too much hassle, I havn't yet because money is tight and I don't have any smd chips to play with I would be tempted to wire them 'dead bug' style to a SIP adapter.... or use the technique explained here. Then, once a circuit is finalised, find a company who will manufacture a small number of boards at reasonable cost - they do exist !!... the folks that make ExpressPCB might be a good place to start looking. Col

I've never done any PIC work. Only messing with ancient simple stomp box circuits. Seems that it would be fairly simple then as far as support circuitry - thats good. I'm not sure that a D/A section would be necessary, there are a number of low power class-D 'solutions' that have digital inputs - i guess for all the tiny media players that abound these days. Excellent Programming is something I CAN do. My first assembly coding was on a ZX Spectrum in the early 80s ! More recently I've done some game boy advance stuff (ARM asm and C/C++). I've also done a little simple dsp stuff, using the VST kit from steiberg, and also building gadgets with sync modular and Reaktor(the new low-level 'core' tech). The PICs seem to have a lot in common with the ARM chips... risc instruction set... there are emulators for development.... anyhow, the coding wouldn't be a problem I think... Unfortunately, I don't have a pic programmer, and although some of them are really cheap(£20), I cannot justify spending money on that right now. Maybe at christmas I am very interested in this avenue though, it goes under vactrols on the 'try this next' list. I would also like to see just how easy it would be to set up a simple programmable dsp processor in stomp box format. I'd never considered it before, but now it looks possible. It would be good fun doing lo-fi diy digital fx hehe - nice departure from all that quasi religious germanium and tube based stuff. Imagine a DSP stomp box with a usb cable that you could plug into your pc and download homebrew fx from the internet - that would be mucho fun... like a lo-fi cheap POD with no pretentions and no restrictions. Thats a nice low quiescent drain... will be somewhat different when sustaining I would imagine Looking forward to hearing that As far as the fizz, it one of those annoying 'matter of principle' things... I agree that the best use is with a nice overdrivven sound, however, I also want to play around with other possibilities, and that fizz just wipes out a whole area of 'tonal research' for me. cheers Col

Just been looking at some PIC datasheets - wow. I didn't realise that these chips had so many features - built in rom, ram, comparators, ADC.. most of what we would need. One thing I couldn't find out is what power drain to expect... also, do they need an external clock pulse, or other external support circuitry ? If not, we could just use a suitable pic and a suitable digital input class-d power amp, and that would be it apart from a few caps and resistors ? Col

I'm not sure why you would want to adjust the frequency of the driver pulses ? If you want to get harmonics, surely the easiest and most effective way is to use filtering to encourage resonance rather than some artificial signal generation..... As far as 'latching' the signal - assuming that I understand what you mean - that is something I have suggested as a possibility - using a comparator (or schmitt trigger) to generate a square wave from the guitar signal. Then using some sort of window detector to limit the amplitude above and below two user variable thresholds. There are a few potential issues here: #1 complexity of the circuit #2 it may be that using a square wave will have a negative impact on how the unit functions when chords are played #3 the window detector needs 'soft' corners... particularly for the upper level, it is importand that the threshold is not an abrupt on/off, because the systems will reach an equilibrium somewhere between the on and off points of this threshold. I have been considering a setup that involves using standard feedback compressor to generate a even amplitude signal, then applying a 'squelch' gate to didtch the low level signals, and a feedforward limiter to provide a 'soft' upper threshold. This looks like it should provide everything in terms of functionality. Unfortunately, the circuit is way too complex - too many op-amp stages and FETS - too much power to drive them etc. So I'm going to leave it for a while and then have another go using vactrol style resistive opto-couplers. Or, just use the input signal via an automatic gain control to feed the driver... hey wait a minute... I would suggest that you try it on a real guitar _before_ you start making pcbs... just use a plugboard, or stripboard... Remember, this is a deceptive project - it seems really simple, but in reality its tricky to get your head around all the variables and truly understand what is going on.... The best thing to do is make a driver coil and a simple fetzer/ruby type amp and try to get that working on a guitar... get your base camp set up before you try to tackle the north face I would guess that a minimum spec would be mono 8bit 22kHz. (better to use a higher sample rate than try to use 11khz then have to use bandlimited algorithms to deal with aliasing issues.). I'm not sure how much ram would be needed. Possibly very little - just enough for audio buffer, variables and a small stack. If we use a simple IIR averaging filter for the limiter control and don't try to introduce and fancy filters, it should be doable (depending on the instruction set) with somthing running a very low clock speed < 1Mhz... how fast are pics ? I had a quick google for ready to use dsp modules, but didn't find anything useful - mostly they seem geared for mobile phone handsets - no surprise there... cheers Col

There was no microphonic feedback, but the fizz was just the same, so for my purposes at the time, it was a failure. Of course, it's possible that the fizz was caused by other problems.... I'm guessing that the fizz has multiple causes anyway . But if you think it through, if the fizz is caused by transformer style magnetic feild coupling between driver and pickup, then having a seperate electrically isolated pickup for the sustainer isn't going to help - it only helps if the fizz is coming through electrical connections e.g. earth/power wires. Col

Hmm, I still get crosstalk fizz with the driver in the neck position - trying to put it in the middle position makes that issue much worse = totally unacceptable. It is probably related to the high power pickup i have in the bridge position, but that is how I (and MANY others) get the sound I like best. 'plugging away' is exactly how it feels . Maybe when I get my hands on some vactrols, I might make a little new progress As far as the patents, I've purposly avoided looking at them closely - don't want to get too heavily influenced until we have exhausted all our own ideas. cheers Col

*Note: I wrote this reply just after the Da_free_runner posted, unfortunately, there has been a problem with my forum account, so I couldn't post until now. I'm not trying to 'shoot you down' DFR, just want to help point out possible issues thay you may not have considered, and also to discover if I am misunderstanding the intention of your design. But it's a manually variable pulse that is independant of the frequency that the guitar is producing? Are you sure? Assuming that each pulse is of the same polarity, it will push/pull the string the same way each time... with the string at rest, this will always exite it. with the string already vibrating at some arbitrary frequency, your circuit is as likely to damp the string as exite it further ! To help see this a good analogy is a swing in a playground... when you puch a child on a swing, you only need to give small pushes, the important thing is that the pushes are exactly synchronised with the natural frequency of the swing... What you are doing would be like sometimes pushing with the existing motion and sometimes against. Thats kind of a simplistic description, you missed a key point - yes it is about keeping the strings vibrating.... but the need to be kept vibrating at their existing frequency. Unfortunately, for your circuit to do this, you would need to manually adjust the frequency to somehow match the frequency of the guitars note. e.g. if your circuit was set to 200hz, and the string was vibrating at 300hz, then the drive circuit will only be in phase (helping) the string every 3rd cycle - so it will spend 2 thirds of the time either interfering with or completely opposing the desired vibration of the string. I suppose one thing would be to have the driver output short pulses at a much lower frequency - like tremelo picking.... This could work theoretically, although intuition suggests that it would have audible side-effects, and would require a _lot_ of energy behind each pulse... have you tested your system on a guitar ? I would be very interested to hear it working - say, sustaining natural guitar tones on a scale. I would also be interested to see what you could do with the approach we are trying to develop. You say that you are good at digital electronics... would you be able to develop a simple circuit that could digitise the driver signal and apply more controllable conditioning and AGC in the digital realm before using a class-T amp (or similar) to feed the driver ? What would be the minimum processing requirements to provide DSP for a usable mono signal? microcontroller? PIC? what would the current drain be like? how big would the circuit be? If you can get a circuit designed, I'd be happy to do the coding - I have plenty assembly language experience, some embedded (kind of) and some audio dsp knowledge. maybe there are existing off the shelf boards that would make this easy? just load in our custom dsp code and plug it in ? Col

Very interesting, although its not what most of us are building... seems like some sort of synth idea - if you could midi it up so a keyboard could select the timer frequency, then you could play the guitar with a keyboard . Unfortunately, when the note you play isn't a note or harmonic of a note being fretted on the guiter, it is unlikely to work. What most of us are working on is a system where the frequency of the driver signal is the same as the frequency of the note the guitar strings are playing.... so when you pluck a string or chord, that frequency is used to drive the sustainer. We basically use the amplified pickup output to feed the driver. AFAIK the only way to get enough drive from a battery is to use positive feedback in this way - the driver signal frequency must match the note frequency so the level can build up via the feedback. Col

There are now. Ladies and gentlemen, I present to you "The Bastard": congrats - looking really good Yep, installing it makes a big difference. As we have discovered, some guitar setups work better than others with just the basic circuit, how is your high E string on the lower frets? Another question, what is the current drain like ?

There may be some misunderstanding about what I mean when I say the frequency response is not linear with some of these circuits. The frequency response of the amps is fine. The issues is that the compression level and threshold can be effected partly by frequency. It is possible to avoid this without using extra filter stages. As far as using multiple amps each aimed at a different frequency band, this may allow control over effects like 'harmonic mode', but is unlikely to give us noticable improvements to the basic circuit. It is impossible to seperate the problem areas of the neck from the over responsive ones through frequency (unless you use a fully hexaphonic approach). e.g. High E string first fret has a very poor response - so you would want to boost it... G string 10th fret and D string 15th fret are VERY responsive - so you want to attenuate them - but these are the all the same note So you can't use frequency band seperation to balance their amplitude... My suggestion (ages ago) of using two seperate frequency bands was purely to provide better driver response (not better string/fret balance) - i.e. a custom woofer driver and a custom tweeter driver... this could possibly allow for better efficiency and/or smaller drivers... but it probably wouldn't give enough of an advantage (if any) and would require a lot of R&D - not to mention more complex circuitry (crossover etc.) Col