distortion for Lunettas?

[DUBmatze]

hi,

its me again. is there a easy way to add some distortion to our Lunettas (somthing like a Transister, a pot and a hand full diodes)....
All distortion circuits i try sounds well wehn the pitch of the sound changes but not with a simple squarewave sound (cant made squareisher...)

[commathe]

Try a "Tubescreamer" style distortion. Basically diodes in the feedback loop of an op-amp.



If you make it two or three diodes in a row then you get more headroom. You can also try using a "diode connected mosfet" in the loop instead. You tie the gate and drain together and then add a diode facing "inwards" towards the source if it is an n-channel or facing "outwards" if it is a p-channel. Mosfets have a built in slow switching, high-capacitance body diode. It would definitely round off those corners!

Alternatives might be some JFET or inverter based guitar distortions like this or this respectively.

Maybe check out the "triple fuzz" or the"simple wave folder". You'd want to run a low-pass before you feed square waves in though, as the fun all happens on the rising and falling edges of the wave. You can also hook up a pot, like in this schematic for a little extra fun and control. It lets you "move" where the wave-fold occurs

[DUBmatze]

thanks commathe,

they all looks like guitar distortions (i was already testing a big muff green russian - and wasnt happy with it - but the pedal is broken now so no more testin at the moment).

But i think i start with the distortion of the Tubescreamer arround a LM741 without the input and output stage. The finish line looks also pretty simple (hope bc547, 557 will work here, think i can give it a try without the 386 the Lunetta shut be "hot" enouth..)

what irritates me is when i put diodes like 1n4148 from signal to ground and from ground to signal (like in the tripple fuzz) this will limit my Output signal to 0 - 0,5 V (instead 0-5v) or ?

will breadboard this sunday evening.

[commathe]

Don't knock guitar distortions, some of the designs are very good for synths. Some are really bad though. You'll want to avoid ones that tend to make a signal turn into a square wave - which is what the big muff basically does. I would say get rid off the in/out stages on the tube screamer but also get rid of the cap+resistor to ground that come of the feedback loop, or at least play with the two. They add a midrange hump and scoop a lot of bass off. Probably not the best for a synth. Maybe use a much bigger cap or no cap at all.

I generally don't use diodes to ground and I remove them from pretty much every design, even when it's for guitars. In guitar circuits they are used for "hard-clipping" which is totally pointless when dealing with square waves as it just makes a signal more square-y. Also, like you, I don't like cutting the output down that low, even with guitars.

The JFET based distortion I linked might actually be one of the more interesting ones for feeding square waves through. JFETs tend to lend themselves to very asymmetrical wave shaping as they have a changing transconductance that makes it less effective at the lower negative peaks of a waveform. The only inconvenience is having to use trim pots.

Any NPN and PNP is fine I think. I used a 2N4401 and the 2N4403 in the one I built. The emitter resistors are large enough that pretty much any pair of transistors will work so long as they are the right "flavour". The purpose of the amp at the beginning of the design is not about voltage amplification though, it's current amplification. You will probably get the best results from actually attenuating the signal down to something like 2V peak-to-peak and then amping it back up after.

Other advice I'd give is try low-passing signals before they go through distortions of any kind. Distortions are "wave-shapers" most of the time and their nicest characteristics won't come out if the signal going in has edges that are too steep.

Also, if you really want to stay away from guitar distortions (which would be missing out on a lot of gold in my opinion), then maybe try the cat girl synth modules, there are a few waveshapers and folders and what not in there. Again though, I'd probably add at least a little low-passing before feeding square waves through any of it.

[Cynosure]

The most common lunetta distortion I use is an overdriven linear 4069, like in the Wasp synth. However, that is just clipping the top of the wave with a bit of rounding from the lowpass filter that is added.

Another option for distorting a squarewave is by creating some sort of modulation to the sound. Here are some simple options:

Some of my early waveshaping exeriements (good audio demo in the second post) - http://electro-music.com/forum/topic-49963.html

4011 Ringmod - http://electro-music.com/forum/topic-48145.html

Ian Fritz 5Pulser - http://electro-music.com/forum/topic-19084.html

They aren't typical 'distortion', but they chop up the sound, add harmonics and make it more interesting.


Another trick I like to do is to send two or more notes into a 40106 schmitt trigger and then actively mix the output with the original notes. They cancel each other out in interesting ways.

[commathe]

I really like the idea of actually feeding a waveform to the power pin of a chip! I played with a similar thing to make "dying-battery" percussion sounds a while back which was fun but I never ended up fleshing it out. I just used large-ish caps to ground that would get charged up when you switched the power on and run a few 40106 oscillators, then when you cut the power the cap would discharge and make that spluttery sound unplugging a synth sometimes does. I never manged to get enough control over the attack and decay though.

[DUBmatze]

Thanks commathe, Thanks Cynosure,

i tested the circuit from the tubescreamer. and the recomended thing with the npn/pnp Transistor. with pure squarewaves they have only a minimal effect. a LP filter in front of it will help a bit. I testet the circuits with a pure Squarewave out of my Lunetta and sound from my mobile. the normal audio is easy to destroy .... but this squarewaves a realy hard ...

your 40106 "distortion" is exact what i use in this video: http://www.youtube.com/watch?v=kPm48FfLiy0
it makes the high sounds from the clicks. but in my case it dont makes distortion its more like some duck sound.

when "ringmoding" with a XOR i also got the mod sound when no source sound is there.

the best results i had so far is when iam using a oscilator with a very high freq. and AND Gate this to the source square (digital). after that i mix the source sound and the AND signal together (analog) and filter it with a LP rc filter below the mod freq.. (very unhandy, i need a pot to adjust the mod. freq. and a pot to adjust the filter freq., (in different values, so no stereo pot...))

[commathe]

For the NPN/PNP circuit, try making the signal about 2V peak-to-peak going in and then re-amplify it coming back out. The folding that it causes only happens between something like +0.6V to -0.6V so on a 9V peak to peak wave it's going to get lost.

Part of the problem is also going to be that most "distortions" rather than wave-shapers/wave-folders/wave-manglers/modulators are really just pushing a waveform closer and closer to a square wave anyway. Squarewaves won't get anything "added" when they go through a distortion generally.

You could maybe make a square wave sound as if it was more distorted by high-passing it and then mixing the high passed signal with the original? I'd imagine that might work quite well. That'd add more high-frequency content and could maybe make it seem distorted, as distortions add more high-frequency content as a result of the clipping.

Maybe trying out some sort of "cross-over distortion" could work too. This is speculation though, I haven't tried it with square waves.
[img]http://electro-music.com/forum/phpbb-files/xover_847.bmp[/img]

---

xover.bmp
Description:	cross-over distortion	Download
Filename:	xover.bmp
Filesize:	69.66 KB
Downloaded:	423 Time(s)

---

[LFLab]

Blackstone appliances Mosfet overdrive

4049's where mentioned before, but this is a specific application. Haven't tried it myself, but plan to.

[Cynosure]

[DUBmatze]

its working. thanks for the support, i LIKE the sound!



here is a short demo with a simple Lunetta patch:

[commathe]

Glad it's working out for you now!

I hadn't tired highpassing going into it before so I'll definitely give it a go at some point. Seeing your little waveform sketches has made me realise that it'd have the same effect as what I was using lowpassing for: makes the rising and falling edges less steep and emphasizes the folding.

[acidblue]

I bread-boarded you square wave destroyer and used it with this vco:
http://electro-music.com/forum/topic-28799.html

Granted it's not a square wave but it does work and sounds kinda gritty.
However it sounds better and has a bit more of a range if I remove the 47k
resistor going to ground after the 4n7 cap.

So I have to ask: why does it sound better after removing the resistor and what exactly is the resistor doing??

[commathe]

The cap and resistor to ground form a high-pass filter. When sending square waves through to the wave-folder high-passing them helps emphasize the distortion because it makes the "falling-edge" of the square wave smoother and that helps emphasize the effect. This isn't necessary with other waveforms though, as they are smooth enough that the folding is still audible.

However, it also serves another purpose in this circuit. DUBmatze's lunetta is running at 5V, but the circuit works best when the waveform is only around 0-2V max. The two 47k resistors form a voltage-divider which effectively halves the volume of the wave going in. The cap inbetween makes the high-pass with the resistor to ground, and also blocks DC from coming in to the circuit and adding offset.

[acidblue]

Ahhhh I see. thanks for the reply.
Think I'll add a switch to include/exclude the 47k R.

Definitely gives a different character to the sound when excluding it on
the falling ramp wave I'm using atm.