curtisa

You have to use the resistor to limit the amount of current passing through the LED. The LED will typically only survive 20mA of current at the most, and the resistor is there to prevent the current from rising beyond what the LED can handle. No resistor = no current limiting = burnt out LED. You can make all sorts of current regulators out of those voltage regs, but it's overkill in this application - why spend extra time and money buying/wiring up a regulator when you can hook the LED to a 5 cent resistor instead? 9V or 18V, probably doesn't really matter. The current draw isn't excessive like you say. My only preference for using the full 18V is that the load from the LED (or LED's if there's more than one) is shared equally across both batteries - there may be other sonic ramifications to running with one battery gradually running flatter than the other that I don't know about with that preamp.

Both in parallel will do the trick.

What's wrong with running the preamp and LED off the same 18V supply instead of 1 9V battery each? Just double the value of the resistor feeding the LED and you're good to go. Cheers, Curtis.

The DSP on offer at Wavefront Semiconductor looks like pretty good value. Check that link I posted a lttle further up. Not at all, although if it were me just starting out on PIC coding I'd probably start with the 16F or 18F range - there are plenty of kits and free software available to burn the chips. If you consult the datasheet for the Wavefront DSP you should be able to work out exactly how fast you need to provide it with instructions, and from there you can work out your required processor speed. You may find that the dsPIC is overkill and you can do it with something much lower-specced.

Sorry, I don't have any schematics for battery charger systems. Just do a Google search and see what you can uncover This is still your weakest link - you still have to multiplex to obtain the highest bit depths for the number of channels you're talking about. In between each conversion of each channel, you have to ensure you have sufficient processing time to send the previous sample serially to the transmitter, service any calls to the encoders, apply any digital volume control the the A/D data, service push buttons and switches, drive the LCD display, and run the MIDI. All without introducing any glitches in the audio. Even at such high processing rates, I still think you're trying to squeeze too much into the PIC. Microchip expressly state on the website and datasheet that the dsPIC is primarily intended as a "signal controller" ideally suited to motor control. They make very little or no mention of high quality audio performance, I think that's still the domain of specialised A/D's and DSP's, and for very good reason. They may share the requirement of an A/D, but there is world of difference between the output of an analog speed sensor and an audio signal. Another reason why I'd prefer to see it done with the proper A/D's and DSP's, or indeed without any digital transmission at all!

Nah, forget it. If you want multichannel audio at the sample rates and bit depths you're talking about, you're going to have to consider dedicated multiple ADC's and a DSP, with a CPU controlling the whole lot. The all-in-one solutions offered by Microchip and AVR simply can't do it, the processing power required to do it all seamlessly is too much for 1 chip. If that AVR you've highlighted doesn't have intergrated A/D, you'll have to use an external A/D chip to do it. You could do "pseudo A/D" in software with the AVR, but the processing overhead would be too great for it to work properly over 10 channels. Effectively you're talking about installing a 10-channel soundcard (with a digital transmitter) into a guitar, and even the cheapest ones on the market only do it using dedicated chips. Wavefront Semiconductor seem to have low cost A/D's and 8-channel DSP's, check out their range: http://www.wavefrontsemi.com/index.php?products My feeling is that this is the ONLY way to do it - one chip per function. Well if you consider using a DSP, the processing overhead in the uC drops dramatically. Potentially you could use a 40mHz PIC to "run the show", and leave the DSP to look after the really difficult stuff, ie the audio. This will be difficult too, yes. I suspect that you won't be able to use regular 9V batteries, they simply wont have the capacity to run it all for very long. You really may have to consider something like a laptop battery pack, though I don't really see that as being such a drawback - when you consider the size, weight and capacity of modern laptop battery packs they're pretty good in comparison to a bank of "C" cells or 9V batteries.

Well, here's another thing to consider - because you're running multiple audio outputs at once, and doing an A/D conversion on each one simultaneously, you're going to have to multiplex the A/D conversion process, which means that your net sample rate will drop by a factor of the number of signals being converted. The dsPIC may have 16 analog channels, but it only has one A/D on board, and it can only do one conversion at a time. For example if you were doing 8 channels at 200ksps, each channel's sample rate would then be 200/8 = 25ksps. Nyquist's theorem then dictates that the highest reproducible frequency would be half the sample rate, or 12.5Khz. Obviously performance gets worse as you add more inputs, especially as the PIC cannot run at 200ksps with more than 8 channels. I think a 25kHz 12 bit digital system is going to sound pretty bad - I reckon you need to reconsider your options with this one. The PIC also has to handle all the MIDI and controls aswell, and it really has to do it without causing glitches in the digital audio side of things. If you have the PIC handling the digital audio and you're also adjusting MIDI paramaters with a control knob, the last thing you want is for clicks and pops appearing in the audio output as the PIC code starts branching off to service routines to handle the MIDI and peripheral interfaces. The digital audio has to take priority above anything else, and I doubt the PIC has the horsepower to handle 8-10 channels AND digital interfacing aswell. Is there any reason why you can't consider using a PIC purely as a controller? Does it have to be digital audio? You can still do all your switching and volume control in the analog domain, eg CMOS4066 analog switches and digital potentiometer chips, leaving the PIC to handle the MIDI and interfacing quite easily.

OK, this is starting to make a little more sense to me now. So you could conceivebly have the "look" of a 5-way switch/volume/tone, but the actual function of those controls is a lot deeper than that. You'll probably find you'll need some kind of display built into the guitar so that you can see exactly what parameters you're adjusting with the encoders at any given time. 2x16 backlit LCD would be my preference. Hmmmm...dunno. Do you have a specific DSPic part in mind? Datasheet? Oh, I'm sure it's fully realisable, but I'm just questioning if it's cost-effective, and worth your time as a DIY project, especially the wireless side of things Wired ethernet sounds like an interesting idea, but the cables invovled are a little on the "light duty" side of things for my tastes - I wouldn't want to take an ethernet cable with me on stage! Maybe ethernet with a different cable, something more rugged? Another thing to consider - are you sure you want to do all this customisation to your PRS? With all the extra control circuitry you're wanting to install, there's no way you're going to fit it all inside the guitar without routing more cavities into the body. Have you got a cheap, crappy guitar that you could experiment on first without worrying too much about ruining the body and/or finish while you develop all these systems?

Why 10? Even assuming a separate output for 3 pickups and a piezo that's only four volume controls. If you were to implement hexaphonic output in the piezo aswell, you'd go nuts trying to alter 6 more volume controls for each string output, irrespective whether you had 6 separate volume knobs or one multi-function volume encoder controlled by a uC. You're doubling up the amount of research and development if you incorporate AD/DA into the audio signal - by doing so you'll have to develop some kind of DAC at the receiving end of the guitar output. The AD in those PIC's won't be the highest quality in the world, but it may be usable. I haven't heard of anyone doing audio DSP with a PIC - I think you'll find a dedicated DSP chip is preferred for that sort of thing. Using only two encoders will make controlling all your pickups and outputs quite tricky, esepcially if you're planning on using the guitar in a performance. Imagine playing the instrument and at the same time scrolling through a bunch of menus to find one volume parameter for the 3rd string piezo output, and you accidentally scroll past it, and lose your place in the performance because of it If it were me I'd still prefer to have 1 knob per pickup (4), rather than 1 knob per output (10), or two multi-function knobs. Sounds to me like you're starting to talk yourself out of the idea now! I've got no idea how you'd implement multichannel audio over wireless ethernet, but I know that it's a lot of work to start from scratch - it's not the sort of thing I'd want to DIY, waaaaay too complicated Honestly, I admire your passion and tenacity for wanting to do such an adventurous project, but I'd take a couple of steps back from the ideal and focus more on what is practical. I've done similar things to you aswell in the past - I very nearly completed a DIY version of the Mesa Boogie Triaxis preamp many years ago, even to the point of having all the CPU code developed. But it eventually got to a point where I was aiming way too high and achieving very little in return, so I stripped it all down and made a simple valve preamp with a bunch of knobs and no CPU control whatsoever. What about thinking a little laterally here? What about having a guitar with 3 pickups (1 volume knob for each one) and a hexaphonic piezo (with one volume knob for all 6 outputs), and then doing all the fancy blending and combining away from the guitar, with a foot controller or something, like what Roland do with their VG systems? If you're not wanting to use such a customised guitar in a performance, what about running all the multiple outputs into a small mixer and doing all your processing that way?

No, you won't gain anything by changing the vibrato system - the scale length of your instrument has to stay the same. If anything, you may find you'll lose some space due to the positions of the pivot points on the floyd being reasonably far forward compared to a stop tail or Strat-type bridge Maybe a rotary encoder (or multiple rotary encoders) with an integrated push-switch in the shaft? You'll need to implement it with uC though, say a PIC micro or an ATMEL micro. Doing anything like this with descrete logic chips is not practical. I think whichever way you look at it, the only way to do everything you're hoping to achieve with the instrument, you'll have to do it with a uC - there's no way you'll be able to fit it all inside the guitar and power it otherwise. Omeg offer dual concentric pots and dual concentric switch-pots, although I think they will only manufacture such parts if you place a minimum order of 100 units. http://www.omeg.com. Google searching reveals a number of places where you can get dual concentric pots if you only want certain pot values. OK, but wireless power isn't available yet I would have done the phantom powering with an XLR at both ends of the lead - you still have to plug the lead in to the power supply, and that carries the same risks as plugging into the instrument with the TRS plug. If you were to go for XLR's at both ends, you can just use a regular microphone cable which shouldn't be any harder to find than a regular guitar lead. OK, so you want a regular guitar signal, MIDI, and possibly phantom power. You'll need 2 poles for the guitar signal, 1 pole for the phantom power, and a minimum of 3 poles for the MIDI (MIDI does use a 5 pin plug as standard, but only 3 of the pins are actually wired up). You can get 6-pin XLR plugs, which would cover all your signal and power supply requirements in a single connector. Even if you want to transmit all the signal wirelessly, you'll still need on-board power for the MIDI, piezo pickups, and the MIDI/audio radio transmitters - lots of batteries! I wouldn't let the problem of implementing "non standard" components into the guitar bother you - based on the amount of customisation you're talking about, the fitting of an XLR plug in the guitar is the least of your worries!

True, but then the guitar becomes limited by the power supply box you have to plug it in to. If you wanted to use this guitar in a performance via a wireless system it's out of the question. If you're quite happy to use the instrument in private or onstage with a lead all the time, then that's less of an issue, but you will have to make sure you have the power supply box with you all the time. Personally I would never use a TRS jack for phantom powering as there is the risk that you'll plug the lead in with the power supply switched on, and short out the power supply as the lead is inserted into the socket. I'd be using something like an XLR connector where you cannot accidentally short two or more pins together as you plug the lead in. Yes, you can get dual concentric pots (never seen triple concentric pots before), but they are mondo expensive, and typically only a custom production item. Other alternatives are pot on top/rotary switch on bottom, but again they're expensive and hard to come by. That said, you may be able to scavenge some parts from dead car radios or television sets. Which pickup you leave out is entirely up to you, and also dictated by exactly how much room on the guitar there is for all your pickups. Personally I think you're being too adventurous in trying to extract so many different sounds out of one instrument, however I'm more than happy to be proved wrong! I've got no idea how much room there is between the bridge and neck of your PRS for the pickups, but from what I remember of that model line there's only room to fit 1 bridge humbucker, a single coil middle pickup, and a neck humbucker. That leaves pretty much no room for a sustainer (which has to go in the neck position). The piezo floyd system will look after itself, but you've still got to retrofit the bridge into the PRS which wouldn't be an easy or cheap task. I've got no idea what you can expect in the way of sounds that can be extracted from such a switching system, but it certainly looks like what you're after. It certainly gives you just about any permutation of humbucker/split/phase/series/parallel combination you could ever want, although I notice that it still relies on the existing selector switch to select the pickup combinations - I'm guessing you need to program all your sounds into "banks", each bank containing 5 pickup combinations (if you have a 5-way switch), and if you need further permutations you change to the next bank of 5 combinations. If you can program your own code and CPU you've got the scope to include or omit anything you want Cheers, Curtis.

So you're going to retrofit a floyd to a PRS? Guess you could do it, but it wouldn't be cheap. I certainly wouldn't want to attempt it myself, make sure you locate a good, reputable luthier to do the work! And that's 3 pickups, a hex piezo AND a sustainer - I reckon you'd struggle to fit so many pickups on the guitar. One volume pot per pickup, 3x pickups plus the piezo makes 4 volume pots all up. Do you have room on the guitar for all of them? It could be done, but with so many volume controls for each pickup you'd need to do it actively, ie with a battery-powered on-board mixing system, which may not be such a stretch as you'll need the battery for the piezo anyway. Now this is getting really complicated! Yes, this could be done too, but again you're looking at active control of the pickup selectors and volume controls. And an "endless pot" is raising the bar even further because now you're talking on-board digital control. That's a lot of work in there! Nah, too much. You'll never fit it all in the guitar. You'll never get enough batteries in the guitar to power everything for any reasonable amount of time, let alone all the control systems you want aswell. In my experience, having a super complicated control system in a guitar is more of a hinderance than a help. When I did a similar thing with one of my old Yamahas I found myself spending more time fiddling with the controls than actually using the instrument. I'd start with the separate volume controls idea myself and see how far you can take it before moving on to something more adventurous.

Hi Jack, How exactly are you measuring the pots? 0 ohms across which pins? Does this resistance vary when you turn the pot? You need to remove them from the circuit before you measure them with the multimeter, otherwise your measurements will be affected by any surrounding circuitry connected to the pots. You should be able to get a static full-value resistance reading of the pot if you connect the two multimeter probes to the two outside legs of the pot. By connecting to the middle and either of the outside legs, you'll be able to measure the wiper resistance of the pot, and you'll be able to see the resistance change as you rotate the shaft. The total change in resistance from one extreme to the other will be 0 ohms to whatever the full resistance value of the pot is. If either of these tests come up bad you've probably got a dud pot. CHeers, Curtis.

Still here, just not as often at the moment! You might find the NE570 or NE571 are easier to find and cheaper too. They're essentially the same thing with slightly lower specs.

I tried using the sustainer unit with my acoustic which has a piezo pickup and preamp in the bridge some time back when I was first experimenting with this project. It worked perfectly as I recall, and you could position the driver as close as you liked to the piezo without any feedback or fizz whatsoever. The only drawback was that having the driver too close to the bridge reduced the amount of sustain you could get out of the strings (harder to "move" the strings closer to the bridge), and generally speaking the only strings that worked particularly well were the plain unwound high E and B - the bronze-wound G, D, A and E probably don't have enough magnetic material in them to sustain as well as the plain steel strings. *sigh* I really want to get back into experimenting with the DIY sustainer, but I've just got too many other DIY projects on the go at the moment to fit it all in! After my initial success with the neck position sustainer I kinda lost momentum on the project! Cheers, Curtis.

But surely we're only looking to shield excess radiation of EMI from the sides (away from the pickups either side) and rear (away from the control wiring) of the driver, channelling the driver magnetic force forward towards the strings? That would simply require a magnetic shield that cups the driver leaving the top exposed to drive the strings. The patent mentioned it as a cost cutting exercise - cheaper to add an extra winding than manufacture special shields. What if the shield winding was driven by a DC voltage, so that the outer winding is magnetically "biased" towards one pole? Any ideas if that'd act as an effective magnetic shield? An easy way to get laminations to experiment with would be to take a small power transformer, pull it apart, separate all the lams, and cut and shape them to the new profile to suit the driver. The only trick is to make sure all the lams are electrically isolated from each other (varnish perhaps?), otherwise the advantages of a laminated core are lost. Farnell and RS Components sell transformer kits. Where did that photo essay on dissecting a sustainer go? I looked for it the other day but couldn't find it in this giant thread!

I think so...I hate patent speak...I think I caught some of it...grrr Not sure. Probably. Like I said, I was only skimming through the patent. It could be something as simple as that, but I was thinking (hoping!) that there was something different happening there. I'll have to re-read the patent, and check through the schematics again. I'm certainly not discounting the slimline driver It certainly does work, but I do wonder if there is a fundamental reason why the commercial manufacturers of sustainers do not use it - they have all seemed to have gone with much more complicated drivers (dual rail, dual winding, laminated cores, sliding magnetic shields etc), or at least drivers that are expressly used for that purpose (no stacking of pickup/driver). Hmmmm...any idea how magnetically shielded speakers are made? You know when you place your hifi speaker too close to a regular cathode ray tube TV and get all those crazy rainbow colours and distorted pictures? Magnetically shielded speakers are specifically designed to avoid this - there might be some design clues in there for us aswell. Also, the Sustainiac patent talked about having a driver with a dual winding - one that is wound directly around the core for the driving, and a second one wound around the driver winding to act as a magnetic shield. Not sure what you'd do with this second winding though - earth it out at both ends? Earth it at one end? Drive it with an antiphase driver signal? It's probably in the patent aswell... I'll have to find a cheap and nasty single coil dual rail pickup that I can trash for the purposes of experimenting... Noise signals can be induced into adjacent circuits in several ways - one is magnetic coupling, caused by the radiation of magnetic forces directly proportional to the signal (sitting too close to the TV while playing guitar, holding a microphone too close to a power transformer), another is capacitive coupling - any two conductors running near each other exhibit capacitance between them. So if everything is "coupled" to everything else by invisible capacitors, then it's possible for signals to "leak" from one point to another. 99.99999999% of the time it isn't a problem because the capacitances involved are sooooooo tiny, but occasionally it can be a problem, particularly if we're using really high frequencies, really high circuit impedances, or lots of parallel conductors. Electric guitars don't use hi freq's, but will use high-ish impedances (500K volume pots, 16K pickups etc). That's why circuit layouts can be critical in RF circuits or guitar tube amps - poor layout can result in excess capacitive coupling that can cause oscillation, poor response and other problems. Cheers, Curtis.

Hehe...no, the guitar that I built to run the sustainer is the weakest link. The sustainer itself is cool! BTW, the sustainer performed very well at last night's gig That has me worried a little too, but on leafing through the Sustainiac patent yesterday, it seems that their system is perfectly happy running mid position with either the neck or bridge pickup selected, although their system is for a fully active guitar. Also interesting to read yesterday in that patent was their comments about feedback - they talk about induced interference (magnetic coupling - the fizzies) and electrostatic capacitive coupling, something I hadn't considered before. They mention that the way to get around the capacitive coupling is to invert the driver signal so that the the driver and pickup signals cancel each other out. Normal and Harmonic modes are then handled by clever manipulation of the driver signal rather than simply flipping the phase of the driver leads like we've all been doing. Their patent also seems to do away with the need for a dedicated AGC amp, and the driver is driven from a pulse-width-modulated signal from a current sourcing amp instead of the venerable LM386. Interesting stuff in there. Judging from what I read yesterday in the Sustainiac patent, I'd say the bilateral driver is the ONLY way to go for a mid position sustainer. The current slimline system (like the one you, me and others have built), while fully working, is just too "dirty" - it radiates a lot of garbage into the highly sensitive guitar electrics and is a major source of instability.

Hehe...still here, just not as often as before. Been far too busy, and have a lot of new DIY projects on the go at the moment. The Jaycar "Champ" circuit is the one I used with the gain pins shorted for max gain, with the addition of a N-channel JFET source-follower circuit in the front end for buffering purposes. The Champ by itself has bags of gain to drive the sustainer, probably even too much, which is one of the reasons why Col and myself were pushing for the AGC circuit to automatically remove gain where it wasn't required and boost it back up when the drive was too weak to excite the strings. I recieved my free samples of the THAT2181 VCA IC the other day. I wouldn't mind trying to see if I can create a self-contained VCA solution using this chip and seeing if I can get better performance from my current Champ/JFET sustainer - the experiments I did with the feed-forward limiter were certainly promising, and Col's experiences with his new VCA are encouraging too. Will try to devote as much time as I can to this, but the Christmas period is fast approaching and work gets busy this time of the year. I'd also like to explore at some point getting the mid-driver option working for one of my main guitars - a HSH Ibanez S470QM (I never use the mid single coil position). The DIY sustainer guitar I put together for the original project does work, but it's not the greatest instrument in the world Obviously this is going to mean developing some kind of new driver that has very little EMI radiation that will fit in a single coil position, so I'll have to go back through the last 50-odd pages of this thread to get all the info for doing so The cool news is that I'll be using the sustainer guitar at our album launch gig tonight, which will be recorded to mobile 24-track for possible release later on. I've managed to incorporate the sustainer into some of our live interpretations of our stuff, to substitute for some of the keyboard parts, and it works a treat. Cheers, Curtis.

I think this whole phase distortion stuff is a bit of a red herring. Slight phase distortion (particularly in fundamental mode) would manifest itself in some notes sustaining better than others. Gross phase distortion (circa 180 degrees) does exactly what we want it to - harmonic mode. I think the fizz is nothing more than our high gain preamp/drivers being run at their limits, clipping away merrily, driving a very inductive load, in an unshielded (or at least poorly shielded) environment, inside an instrument that has a penchant for picking up all sorts of EMI garbage (ever played an amplified guitar while sitting in front of the TV?) I'll give you an observation from my sustainer guitar. When I'm running clean I get the fizzies. When I damp all the strings and then just take my hands away to let the sustain build naturally, as the sound gets louder the fizzies only start to appear at a certain volume level. If phase distortion was the cause of this the fizzies would be present no matter what level the string was being sustained at and they wouldn't get worse as the string vibrated more. Because the driver is in a fixed position, and the note I'm playing is also of a fixed position (for want of a better term) the phase has to be constant too, even if it is shifted by some degree. The fact that the fizz appeared at a certain output level and gradually got worse as the string vibrated more indicates to me that the fizz is due to clipping as the preamp/driver runs out of headroom and starts to "square off" the signal, causing lots of upper harmonics to be passed to the driver and being induced all through the guitar electronics. When I was doing the experiments with the compressor/limiter I could get much less fizz than with just the LM386 - I would think that this is because I'm simply running the system more efficiently with less clipping, and less drive when it wasn't needed. The fact that Col has built a similar system and got a similar result would indicate that it's "cleanliness" of the signal that's one of the major keys to reduce the fizz. I'm starting to wonder if we actually need a further level of improvement to the system - not only do we need to run the preamp more efficiently, but we may also need a more efficient driver - something that takes less drive to move the string just as much. Cheers, Curtis.

Are you using the stacked single coil driver/pickup that Pete and myself used? I had high pitched oscillation from mine too. Don't know what frequency exactly, but it would've been around the 4k area. I think Pete found that he had to disconnect ALL the pickup wires except the source pickup to get his to settle down. I found I had to short the pickup wires together on my driver/pickup stack to get rid of the squeal. I'm personally not particularly interested in extra outboard doodads just to get the sustainer working. I'd much rather have a system where I just plug the guitar into an amp (or permanent rack) using a plain old guitar lead. The outboard control option also becomes a problem if you want to use the guitar in a live situation with a wireless. That said, I can see that many other people might want to have an outboard controlled sustainer as it opens up a lot more tonal trickery with the effect. Just not my cup 'o tea I guess Hehe, thanks Spazzy Interesting side note - I tried out the sustainer guitar this afternoon while playing with a slide. Sounds quite bizarre, almost Theremin-like. Harmonic mode works on the slide too. Cheers, Curtis.

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 A preamp of some sort is mandatory. Plugging the sustainer and guitar in parallel into your computer mike input is loading both down too much. Increased loading will reduce the amount of signal getting through to the sustainer and also affect the frequency response of your guitar pickups, changing the effectiveness of the sustainer to work on different strings and positions. Interested to see what circuit you've come up with, Col I'm a little bit concerned about the battery drain figures you've quoted, but it may be unfounded. Cheers, Curtis.

Thanks Pete and Spazzy. Yeah, I'm a bit of a Dave Gilmour fan, although as the saying goes, I like his old stuff more than his new stuff Cheers, Curtis.

Well, it is annoying, yes. But it does sound funny Either way, yes, it'd be better if it didn't exist. The software is free from the manufacturers website - http://www.microchip.com It's called (or at least used to be - haven't used it for some time) MPLAB. Here you go then : Linky Please 'scuse the quality - it's just a pair of mikes in the room. Just after the fade-in, you can hear me switch the sustainer in on fundamental mode, and then at about the 15sec mark I switch to harmonic mode for the rest of the recording. Cheers, Curtis.

Welllllll, I wouldn't call it the same symptoms. You say you have an annoying "pop" whereas I have a mildly amusing "thbpbpbpbpbp" What's interesting to note is that I have a volume control (sits in the position in which the guitars' tone control used to be) between the JFET buffer and the LM386, so I can dial in the amount of drive being passed through to the sustainer. Even with this all the way to zero, ie no signal being passed to the sustainer, input shorted to ground, I still get the "thbpbpbpbp" when I turn off the sustainer. I'm starting to think that the noise I'm hearing is the LM386 in it's death throes as the battery supply disappears, not a backlash pop from the inductive load of the sustainer. If you think about it, it's very unusual to get a pop from most hifi amps when turning them off (at least in my experience), and they're connected full time to an inductive load too, ie the speaker. When I was testing the LM386 circuit in the bench before connecting it to the guitar, I had it driving a little 2" speaker just to confirm it was working. It never made much noise when I turned it on and off. At the moment my switching arrangement is toggling the battery +ve and -ve, whereas yours toggles the +ve only. Could be a clue there? You're also using a multi-stage transistor preamp in front of your LM386. I've only got a single JFET buffer connected as a source-follower, ie no extra gain. I'm also curious about some of these class-D chips I've seen mentioned here. Some of them have a standby mode which is claimed to be completely pop-free. All it does is place the driver chip to "sleep", but still powered up, with a current drain of a few nA (yes, nano not milli, that's not meant to be a typo). Cheers, Curtis.