Strange sustainer distortion and squeal/beep

[kr2p]

Hi everyone,

I'm new to these forums, so I hope it's okay to jump straight in and ask for some help! I've lurked for a while and built a few simple pedals from the knowledge I've gained from here, but I'm now attempting to build a sustainer, and there are a couple of problems I can't quite fix without help.

I've built a standard 8ohm driver connected to a fetzer ruby circuit, and I've put them into my guitar and I get sustain, so far so good. However, there are three problems:

1. When I turn on the sustainer system the guitar starts sounding slightly distorted. I would understand if just the sustained notes became distorted because maybe the feedback loop distorted it or something but no, it's just normal picking and strumming and I have absolutely no idea why this is.

2. I have a killswitch in the guitar as well, and when I press it with sustainer off it's fine it cuts the sound. But when the sustainer's on it cuts the sound except for a really high pitched squeal/beeping feedback noise. This changes pitch and volume as I change the volume in the fetzer ruby but I don't know how to stop it. My best guess is cutting the signal out forces more signal through the fetzer circuit which increases the feedback? I don't know if this is it and if it is I have no idea how to fix it. Putting the killswitch before the signal splits to the output and the driver would do it but it means I couldn't use the killswitch on sustained notes without them dying out, and that would be a shame.

3. When I turn down the volume I get similar behaviour to the killswitch problem. The volume starts going down until a point when it starts feedbacking and squealing. Again, I could put the volume before the signal split, but it would mean I couldn't swell the sustained notes, and that's a feature I would quite like to have.

I'm not sure how to attach a schematic, but I'll try and describe the signal path. The bridge pickup out splits, one side goes to the killswitch then to the guitar out, the other goes to the input of the fetzer ruby which is connected to the driver. The fetzer ruby is always on, and power is supplied using a stereo guitar cable with +9V through the ring if that makes any difference. To turn the sustainer on and off I just lift the positive side of the driver when I don't need it, the reason for doing it this way is so I don't need batteries in the guitar which is a pain, and it also means it doesn't pop when I turn it on and off which some people reported problems with.

When the neck pickup is off both side of it are lifted so it's completely disconnected from the circuit as I read that is what is needed, so I don't think the problem is with feedback between the neck and driver.

This is so close to working, and with the addition of the power supplied through the guitar cable, a lot of the problems I've read about with these DIY sustainer builds seem to be fixed so I'm quite keen to complete it to share the build!

Thanks for reading all of this and thanks in advance for your help! Any ideas?!

[ashcat_lt]

I suspect the bulk of both of the problems has to do with capacitive coupling inside the control cavity.  That's pretty high gain circuit running in proximity to wires carrying very small signals which are terminated at both ends by very large impedances.  I have a surface transducer that I run off a headphone amp to physically shake the guitar (much more like actual amp feedback), and if its (shielded) cable runs too close to the actual guitar cable, I can hear some bleed, and if I put the thing itself near any of the wiring, pickups, or even the bridge of the guitar I get feedback.

How is the kill switch wired?  Is it actually shorting the output jack like it should?

[kr2p]

That surface transducer idea is pretty cool, I've never heard of anyone doing that before!

At the moment the killswitch is push to break and just breaks the line out, so no it doesn't send the tip to ground. However the volume pot also does the same squeal when its all the way down and that is shorting the out so I doubt it will fix it, but it may prevent future problems so I should probably do it anyway, thanks for pointing that out.

I'm afraid I have very shaky knowledge when it comes to impedances and shielding so I apologise if I say something dumb. I was thinking of instead of just splitting the cable in the naive way to do a buffered split seeing as I have power to the guitar anyway for the amp, could this help with the impedance problem?

I should also mention that I've installed this into a Jaguar guitar, and the amp circuit is in the top control cavity, with everything else kept in the lower cavities. The only wires that go anywhere near it are the driver wires, +9v, ground, and the input. The wires carrying signal are all quite far away, at least 10cm, so I can't see capacitative coupling being an issue. I've also shielded all of the insides with conductive tape and I've checked that it's all connected so I don't think there are any issues there either.

[Quackzed]

its possibly oscillation, due to the very high gain of the fetzer ruby. if you have a gain control between pins 1 and 8 of the ruby, try turning the gain down and see if this removes the squeel... increasing the resistance between these two pins MAY lower gain enough to get rid of the oscillation and still leave enough to drive the driver, but it should tell you if your dealing with oscillation due to high gain...

[kr2p]

I've actually left those pins disconnected so I have the minimum gain you can get from this circuit. I found it was still possible to drive the driver without making it any higher!

[kr2p]

Okay, so I've added in a buffer for the output signal, in particular before the killswitch, and as expected the squealing when I press the killswitch or turn down the volume has stopped. After putting this in I had to reduce the volume in the fetzer ruby otherwise it squealed all the time, presumably because of impedances or something (I don't really know what that means, I will do some googling though). I like to think of it like this; before the buffer was in, adjusting the volume adjusted how much signal was split to the driver and how much to the output. Reducing volume or hitting the killswitch sent all the signal to the driver pushing it into the nasty squealing point you get whenever you push a sustainer system too far. Putting in the buffer meant that no matter what I did to the signal after it, the signal to the driver became consistent. Does this seem reasonable?

Anyway, that's that bit sorted, I still have the problem that when the driver is on it's like theres an overdriven signal mixed in with the clean signal. The volume of this overdriven sound seems to be independent of the volume of the clean signal, i.e. when I turn down the volume or hit the killswitch the clean signal goes down how I want it to, but the overdriven sound remains.

I find this really weird because, when the killswitch is on, the line to the output is cut so there should be nothing at all going to the tip of the output jack? The only thing physically connected to the output is the volume pot and tone pot which are only connected to the rest of the guitar through a ground connection with a capacitor and the volume which acts as a resistor. So unless the signal is somehow going through ground or something I'm completely stumped? This is way beyond my electrical knowledge!

Another strange thing is it's only when the driver is connected. So in my setup when the power to the guitar is on the fetzer ruby is always on, I turn the sustainer on and off by lifting one of the ends of the driver. There's no overdriven sound when the amp is on but with the driver disconnected, only when the driver is connected.

[ashcat_lt]

Your idea of why the buffer helped is as good as anything I can come up with. 

I wonder if that distorted signal isn't actually on the power supply somewhere.  Just like in the LED thread I just posted in, this thing is pulling a pretty good pile of current, and is quite likely shaking your rails pretty hard.  This will add noise to any other active circuit sharing those rails, but worse yet you've got it running along a pretty long wire directly paralel to your guitar's output signal via that TRS phantom power setup you've got.  BAC across the 386 and see what happens.

Now let me ask you:  Every time I've ever talked about this TRS phantom thing, everybody has always complained about the possibility of shorting something something as you plug/unplug something something.  Do you see any problems with that?  Did you build in current limiting on the supply to avoid this?  I've always thought that any likely short condition would be extremely brief, but you've actually done it, so I'm curious about your experience.

[kr2p]

Hi ashcat_it, thanks for your response. I've done some fiddling and maybe have some more info that might be helpful, here are the results!

1. The distorted signal is very quiet when the driver is disconnected but the amp is still on, like a very faint background signal, and gets significantly louder when the driver is connected, to the point where the signal sounds like a subtle overdrive pedal (it sounds quite good, but not what I'm looking for!).

2. The level of the overdriven sound is completely independent of the guitar volume pot.

3. The level of the overdriven sound is entirely dependent on the fetzer volume pot, when that's turned down it reduces to zero. I think this doesn't support the current draw argument, because I don't think the volume pot affects the current draw does it? I could be wrong about this.

4. If I remove the lm386 but otherwise leave the circuit on and connected the sound completely goes, this supports your current draw causing noise along the TRS cable argument.

Sorry I'm fairly new to electronics, what do you mean by BAC across the 386? If it is the fact its running parallel to the signal is there anything I can do about it without adding a battery inside the guitar? I'd rather not do that if possible for various reasons. Also writing this I just realised I put in a dc jack and socket inside the guitar to connect the fetzer ruby while I was testing it without the TRS setup, so I'll test it later with power supplied separately and see what that does.

As for you question about shorting the current when you plug/unplug, I've done that a few times and as you say it's very brief and it hasn't caused me any problems so far. The way I've set it up is I built an enclosure that goes on my pedalboard that has power going into it, and a 3pdt switch that lets me stop power going to the ring of the TRS, with bright green and red LED's indicating whether power is on or off. I do my best to only plug or unplug when power is off to avoid this kind of issue. The only problem would be if you plugged in a normal cable while the power is on, and in that case because I supply power using a voodoo lab pp2+ the short protection should kick in, leaving me with a fully functional standard guitar setup. I've also chosen my TRS cable to be bright blue and put blue tape on the enclosure so I can't get confused and make the whole thing fairly idiot proof. I hope this helps! If I get it all fully working I'll do some post about how it's set up because someone might find it useful.

EDIT: I just found the LED post, BAC is a big ass cap haha. You mean from pins 4 to 6 don't you? I will give that a go in a bit.

[ashcat_lt]

3. The level of the overdriven sound is entirely dependent on the fetzer volume pot, when that's turned down it reduces to zero. I think this doesn't support the current draw argument, because I don't think the volume pot affects the current draw does it? I could be wrong about this.

IDK for sure, but I'd imagine if you want it to pull an 8ohm load to a higher voltage, it needs more current to do it.  How much the current draw changes from idle to actual full drive is dependent on the internal circuit in the chip, but I'd have to imagine it gets bigger when you drive more signal through it.  Part of the issue might be the bottom of the coil shaking the ground point to which it's connected, if that makes any sense.

As for you question about shorting the current when you plug/unplug, I've done that a few times and as you say it's very brief and it hasn't caused me any problems so far. The way I've set it up is I built an enclosure that goes on my pedalboard that has power going into it, and a 3pdt switch that lets me stop power going to the ring of the TRS, with bright green and red LED's indicating whether power is on or off. I do my best to only plug or unplug when power is off to avoid this kind of issue. The only problem would be if you plugged in a normal cable while the power is on, and in that case because I supply power using a voodoo lab pp2+ the short protection should kick in, leaving me with a fully functional standard guitar setup. I've also chosen my TRS cable to be bright blue and put blue tape on the enclosure so I can't get confused and make the whole thing fairly idiot proof. I hope this helps! If I get it all fully working I'll do some post about how it's set up because someone might find it useful.

Good idea with the switch, but even better idea to power it from a short-protected wall wart.

EDIT: I just found the LED post, BAC is a big ass cap haha. You mean from pins 4 to 6 don't you? I will give that a go in a bit.

Yes and yes.  Basically, you need better power supply filtering in general.  It definitely would not hurt to filters on both ends of that power run, possibly plus the local decoupling (=filter, just a fancy term for it) at the 386 itself.

[kr2p]

I tried moving the power to go directly into the amp, and the distorted sound completely vanished when turning the volume down, however it's still there when I turn the volume back up, at a similar level to before. My conclusion is that the power running next to the signal, and by the pickups, inside the guitar is distorting the signal, by whatever electronic magic makes that happen, and when the power is running through the guitar cable it happens all along there regardless of whether the volume pot is up or down. Things I don't understand are; what is actually causing this and why does it sound like the guitar distorted instead of just a hum or buzz or something, and why does connecting the driver make it so much louder? Also more importantly how do I stop this?!

[kr2p]

Just saw you posted just before me. I tried putting a cap from pin 4 to 6 but as far as I could tell it didn't do anything. I also tried it at the jack going out of the guitar, and on the input of the fetzer but alas no improvement, however the biggest one I have on me is 220uf so maybe that's not enough? Do you suggest I put one before the long cable run, one on the guitar input jack, and one on the fetzer ruby? That already has a 100uf but again, maybe that's not enough?

I agree about the short protection, although I think the voodoo lab has that built into anyway. I might try and learn how to make a short protection myself as an extra precaution though, plus it's another useful thing to learn!

EDIT:

Basically, you need better power supply filtering in general.  It definitely would not hurt to filters on both ends of that power run, possibly plus the local decoupling (=filter, just a fancy term for it) at the 386 itself.

Success! Sort of.. I put 220uf caps both sides of the guitar cable and one at the fetzer, and theres no audible distortion no matter how I set any of the controls. The sort of is because now the damn thing won't sustain.. I haven't just buggered the circuit because if I turn the volume of the fetzer up it still squeals like it did before, I just can't get it at a setting where it sustains but doesn't squeal. It may be that i need more gain from the lm386, and that before the distorted signal helped with the sustain so I didn't need to worry about a gain pot, but now that it's clean it won't work (everyone knows a distorted guitar feedbacks easier). I'll add a gain pot and see if that helps.

[kr2p]

I spoke too soon. When I add more gain to make the guitar sustain the distortion comes right back, although the guitar is now silent when the driver is off and the volume is down, whereas before there was a slight background distortion, so there's some improvement. The volume of the distortion is now pretty much proportional to the level of the driver. I'm thinking that because the driver is just an electromagnet, it's moving the electrons in the copper shielding tape that's surrounding it in the pickup cavity? This is connected to ground so I don't exactly know how that affects the output but it's an idea; is this what you meant by

Part of the issue might be the bottom of the coil shaking the ground point to which it's connected, if that makes any sense.

?

The driver could be doing the same thing to the +9v and the the line for the fetzer input that also run by it, so I'm thinking for those two i'll tape them to the other side of the pickup cavity behind the neck pickup with copper shielding tape and that should reduce the problem if it's due to that. But if it is shaking the ground how could I fix that problem? The whole cavity is covered in shielding tape and obviously I can't shield it! I could remove some of the tape but then when the guitar is in normal operating mode it won't be shielded.

EDIT: I opened up the guitar and separated the ground, live, and signal wires to different sides of the cavity, securing them with copper tape which should act as shielding. I then put a layer of insulating tape inside the cavity, a layer of copper shielding tape on top of that as an extra shielding layer (I have no idea whether that works or not but it can't do any harm) separated from ground, then another layer of insulating tape so the driver doesn't touch the copper. This is as much shielding as I can think of right now. I also did the separating with copper tape thing along the side cavity away from the driver where the power runs next to the signal. I didn't bother in the main control cavity or the pickup selection cavities because the wire runs are really short in there and not next to each other. As a result there may be slight improvement but nothing significant. There are 220uf caps before and after the guitar cable, a 220uf cap at the fetzer, and a 100uf cap at the buffer, I don't know what more can be done! I've made a recording of the sound it's making but I'm not sure how to attach files?

EDIT 2: I took the driver out of the guitar, but still had it connected to the circuit, and the distortion was exactly the same. So it's nothing to do with the driver directly interacting with anything in the circuit. It seems to be just the high current interacting with signal, but I don't understand how, because it's still there if I move the power to go directly into the guitar rather than through the TRS, and all the live wires in the guitar are separated and shielded from the signal wires. Sorry for all the posts, I like to tinker to find the problem, but I've run out of things to try out now. Any ideas?

[Quackzed]

shielding is a good idea. either shield everything thats 'not' the driver circuit, or everything that 'is' the driver circuit... or both... but the sensitive wires from the pickups before any active buffer/splitter/ driver circuit will be especially sensitive to picking up electromagnetic sh..tuff... and the wire to the driver coil and coil itself will be especially good at spouting em 'shtuff'... 

[PRR]

> I'm thinking that because the driver is just an electromagnet, it's moving the electrons in the copper shielding

I do not know what you are talking about (never did it).

But if you are sending signal to an electromagnet on the guitar, your _pickups_ are sure to catch some of that field. They pretty good at that.

I imagine this has been tried before, and you have looked at some implementations. But seems to me the driver has to be quite far from the pickups. But not at the bridge because the strings don't shake there. Not up the fretboard because your hand is there. So whatever point you try is a compromise.

You want the driver-pickup path to have more loss than the driver-strings-pickup path. Otherwise it is just instant howl.

This assume standard electromagnetic pickups. If piezo pickups, then magnetic signal field won't pick-up (and electric field can be shielded). But piezos are not generally rock-n-roll. A piezo driver is possible but would mean a massive attack on bridge design (now the bridge is a better drive point), and probably extraordinary piezos and high-voltage amplifier.

[kr2p]

The implementation I'm using seems to be fairly standard, with the driver at the neck position, the neck pickup just behind it but completely disconnected from the circuit when the sustainer is being used.

I don't think it has anything to do with the driver interacting with the pickups or the wires in the guitar, because the level of distortion is independent of the distance the driver is from the guitar. If that was what was causing the problem I would expect either the distortion to decrease as you moved it further away, or there to be a constant squeal as it interacts directly with the pickups. What actually happens is it's like the guitar is being run through an overdrive pedal with the gain being dependent on the volume the fetzer ruby amp is set at. I can make it interact with the pickups by putting it nearer them and the squeal you get is a very different sound.

I realised when I tried moving the power to directly into the guitar to test the trs cable I stupidly forgot to disconnect the 9v going into the ring, so when I thought I had bypassed the cable I still had it at 9v! Unfortunately I'm away this weekend so can't test it properly until tomorrow, but everything is pointing at that being the issue, which brings its own set of problems haha.

shielding is a good idea. either shield everything thats 'not' the driver circuit, or everything that 'is' the driver circuit... or both... but the sensitive wires from the pickups before any active buffer/splitter/ driver circuit will be especially sensitive to picking up electromagnetic sh..tuff... and the wire to the driver coil and coil itself will be especially good at spouting em 'shtuff'... 

I've tried to shield everything I can, so wires carrying power, and wires carrying signals have been separated to opposites sides of cavities when they can be and covered in layers of copper shielding connected to ground. The only time they aren't separated is in the holes drilled to join up cavities that they all have to pass through, but these are about 1cm long and there's two of them, so total of at most say 3cm which I think is fairly insignificant? Especially compared to the 8m cable with the signal running parallel to the power. What can I do to increase the shielding, would doubling the thickness of the copper help?

If it is interference causing the distortion of the signal, do you know why I would still hear the distorted guitar signal when the volume is down so the tip is connected to ground? This is the bit that's really bugging me, because distortion along the wires carrying signal from the pickups doesn't explain this.

[Quackzed]

the ruby circuit will start to squeel when the battery starts getting low. this might be it. the lower the battery voltage gets , the sooner the squeel starts as you turn up the volume on the ruby... could test this with a slightly drained battery vs. a new one.

[kr2p]

I've managed to fix the squeal using a buffer on the output, the issue that's left is that there's a distorted signal mixed with the clean signal, and that even when I turn down the volume and reduce the clean signal, the distorted signal is still there. Also it's being powered from a power supply which I've measured to be at 9.14 volts, and it doesn't matter what power supply I use the issue remains so I doubt it's a problem with that.