Author |
Message |
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Wed Mar 25, 2009 11:06 am Post subject:
WaveThing Multiply/Divide Wave Shaping Circuits |
|
|
For the past number of weeks, I've been designing a series of VCO embiggeners that I call the "WaveThing" series. Here's an offshoot of one of the designs that might come in handy for someone. I figured I'd break this one off before it turned into one of my more complex designs, as they are all wont to be. This one is fairly simple yet very effective.
It's a design that derives octave-down sawtooth, triangle, and pulse width modulated waveforms from a single downward ramping sawtooth input. I didn't take the extra step to put in a triangle-to-sine converter on the triangle, because I personally prefer a triangle wave as the "unobtrusive" sub-oscillator (the smooth add-meat-but-not-buzz signal). Sawtooth is a great sub-oscillator waveform, and a nice break from the standard square wave sub-oscillator. The pulse output can put out square waves, or fixed pulses of any width, or (really meaty-beaty-big-and-bouncy) a pulse-width modulated waveform.
The circuit should run on either +/-12V or +/-15V. I'm using 15V; 12V may require different values for R14/R15 to get full 10Vp-p output. It *must* be fed with a downward ramping sawtooth. The sawtooth output is also downward ramping. If you don't have a downward ramping sawtooth, an inverter or a switchable inverter for the input can be easily put in. You would want to keep the sawtooth output the same phase as the sawtooth input that you are using (mixing an upward sawtooth with its downward one octave down counterpart can dull the sound). An inverter for the output and input can easily be accomplished with a SPST switch, I believe. Or, if you never intend to put in a downward sawtooth, just put an inverter on the front end and connect the Q output of the flipflop to Q2 (instead of /Q), which will cause the circuit to generate an upward ramping sawtooth.
Note the power connections on the two CMOS parts. Half the flipflop is not used and two thirds of the CD40106 is not used, either. One could probably put in an extra octave down with those parts. I haven't, because that's not the direction I'm going with this series.
The one trim on it is easy to do - you just tweak it until the sawtooth output has a smooth, unbroken ramp. Very easy - just connect the dots.
Cheerios,
Scott
Description: |
|
Filesize: |
98.82 KB |
Viewed: |
2723 Time(s) |
This image has been reduced to fit the page. Click on it to enlarge. |
|
_________________ My Site Last edited by Scott Stites on Sun Mar 29, 2009 4:06 pm; edited 1 time in total |
|
Back to top
|
|
|
Sebo
Joined: Apr 27, 2007 Posts: 564 Location: Argentina
|
Posted: Wed Mar 25, 2009 11:42 am Post subject:
|
|
|
That's great!!!
I was looking for a sub-osc like this for some time.
I will be building one soon. _________________ Sebo
---------------------------------------
My Music:
https://www.facebook.com/cosaquitos/ |
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
Posted: Wed Mar 25, 2009 2:04 pm Post subject:
|
|
|
Very cool little circuit Scott,
I wish I had time to knock up a quick stripboard layout.
Maybe, later. _________________ What makes a space ours, is what we put there, and what we do there. |
|
Back to top
|
|
|
synthmonger
Joined: Nov 16, 2006 Posts: 578 Location: flada
Audio files: 3
|
Posted: Wed Mar 25, 2009 6:57 pm Post subject:
|
|
|
I've been doing similar stuff using a 40106 as the PWM source for multiple sub octaves. One thing I noticed in my design was getting sloppier symmetry on the saw and triangle suboctave outputs.
How clean does the saw/tri look in yours? |
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
Posted: Wed Mar 25, 2009 9:51 pm Post subject:
|
|
|
I've seen it, and it looks very clean! _________________ What makes a space ours, is what we put there, and what we do there. |
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Wed Mar 25, 2009 10:14 pm Post subject:
|
|
|
Dang - you're fast Andy.
I used the CD40106 in this circuit because the main concern with this method is the connection points at the splice. The splice on the triangle is at the peak; splice at the sawtooth is mid-slope. Results using just an op amp comparator alone, even with hysteresis, were unsatisfactory. Biasing the CMOS from the minus supply made it handy for controlling the FETs.
You can see the merest wisp of thin transients at the splice points; filtering can eliminate them, but I didn't bother - I can't hear them, regardless of the frequency.
The PWM circuit, BTW, is taken straight from Thomas Henry's VCO1. One could use the sine shaper from that design to crank out a sine wave if one wanted to (it's a hella good sine shaper).
Cheerios,
Scott _________________ My Site |
|
Back to top
|
|
|
synthmonger
Joined: Nov 16, 2006 Posts: 578 Location: flada
Audio files: 3
|
Posted: Thu Mar 26, 2009 12:07 am Post subject:
|
|
|
P'dang! That's a clean looking triangle. Have you tried to use the other half of the 4013 to knock it down another sub octave? Using another op-amp for a TH sine shaper could be a good idear. |
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Thu Mar 26, 2009 12:39 pm Post subject:
|
|
|
I haven't run a second sub-octave off of it, no room at the inn (and isn't the intent of what I'm actually doing that incidentally spawned this sub-circuit). I imagine it would do OK. You'd just run the sub-octave sawtooth into the same circuitry, using the second half of the CD4013 and some of the available inverters on the CD40106 and more op amps. You could use three TL074s instead of two TL074s and two TL072s in that case.
I'd try it tonight, but my bench supply finally gave up the ghost last night. _________________ My Site |
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
|
Back to top
|
|
|
synthmonger
Joined: Nov 16, 2006 Posts: 578 Location: flada
Audio files: 3
|
Posted: Fri Mar 27, 2009 12:37 am Post subject:
|
|
|
Thanks Scott! That is a pretty slick and clean sub saw.
Are the MP102s acting as VCVR's to keep the saw in correct phase?
What vco are you using? |
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Fri Mar 27, 2009 8:16 am Post subject:
|
|
|
I'm using Rene Schmitz's VCO3 circuit.
The FET/op amp circuit forms a voltage controlled inverter. When the input to the gate of the FET is -V, the output of the op amp is non-inverted. When the input to the gate of the FET is 0V, the output of the op amp is inverted.
BTW, the voltage controlled inverter makes a dandy poor-man's quasi-ring modulator. It's sort of a cross between a Korg style square wave ring modulator and a normal ring modulator. In other words, if you feed the input of the circuit with any waveform, then modulate the gate of the FET with a -V to 0 square wave derived from a second signal (easy to do with a comparator), you get the quasi-ring modulated sum and difference of the two input signals. A ring modulator is a four quadrant multiplier, and this acts somewhat in the same fashion, quasi-sorta. Try it out, you'll be amazed. You want to modulate the gate of the FET with a square wave - the response would be non-linear with a level shifted sloped wave, though that might still be interesting. I suppose you could linearize the FET, but who's got the time?
As for making a sub-sub oscillator by duplicating the circuit - yeah, it works just fine. I've got it up and running on my breadboard. Probably a nice addition (which I have implemented on my breadboard) is to provide an offset trim to the input of the first sub-oscillator. If the input is not centered around zero, any offset from zero will become apparent at very low frequencies on the triangle wave - the two 'ends' of the triangle won't quite touch. Putting in that trim connects'em up slick as a whistle.
Actually, what I have on my breadboard is a circuit that derives five octaves of signals from a single sawtooth input:
/4 Triangle
/4 Sawtooth
/4 Pulse/PWM
/2 Triangle
/2 Sawtooth
/2 Pulse/PWM
X1 Triangle
X1 Sawtooth (the fundamental input)
X1 Pulse/PWM
X2 Triangle
X2 Sawtooth
X2 Pulse/PWM
X4 Triangle
X4 Sawtooth
X4 Pulse/PWM
In the simplest form of this module (listed above), I would like to have a three position switch for each octave to select the waveform and a mix pot for each octave to form a composite signal. Each octave could have it's own output as well. The problem is one of panel logistics - figure five toggle switches, five pots, then add in the PW and PWM pots for each octave and you quickly come up with a vast number of controls. I could easily fit it on one of my panels, but I think most other formats may not be able to accomodate so much control, especially when you throw in all the jacks.
I say simplest, because I have done other things to embiggen these signals, and actually have derived other waveforms. The toggles will ultimately be replaced by rotary switches, and I would have no problem filling up twelve positions on each rotary with the waveform permutations I have accomplished. When I say embiggenment, BTW, that means I've employed techniques to make the outputs sound as if they're not locked to the same frequency - I impose a bit of drift, which embiggens the sound.
There's another direction I also plan to go, which promises to be rather thrilling (at least for me) and somewhat less complex in implementation, but just as rich in output. All of this is why I plan to make the WaveThing a series of different modules.
I should get off my ass and make samples, I suppose....
Cheerios,
Scott _________________ My Site |
|
Back to top
|
|
|
synthmonger
Joined: Nov 16, 2006 Posts: 578 Location: flada
Audio files: 3
|
Posted: Fri Mar 27, 2009 8:56 am Post subject:
|
|
|
Scott Stites wrote: |
BTW, the voltage controlled inverter makes a dandy poor-man's quasi-ring modulator. It's sort of a cross between a Korg style square wave ring modulator and a normal ring modulator. In other words, if you feed the input of the circuit with any waveform, then modulate the gate of the FET with a -V to 0 square wave derived from a second signal (easy to do with a comparator), you get the quasi-ring modulated sum and difference of the two input signals. A ring modulator is a four quadrant multiplier, and this acts somewhat in the same fashion, quasi-sorta. Try it out, you'll be amazed. You want to modulate the gate of the FET with a square wave - the response would be non-linear with a level shifted sloped wave, though that might still be interesting. I suppose you could linearize the FET, but who's got the time?
As for making a sub-sub oscillator by duplicating the circuit - yeah, it works just fine. I've got it up and running on my breadboard. Probably a nice addition (which I have implemented on my breadboard) is to provide an offset trim to the input of the first sub-oscillator. If the input is not centered around zero, any offset from zero will become apparent at very low frequencies on the triangle wave - the two 'ends' of the triangle won't quite touch. Putting in that trim connects'em up slick as a whistle.
Actually, what I have on my breadboard is a circuit that derives five octaves of signals from a single sawtooth input:
/4 Triangle
/4 Sawtooth
/4 Pulse/PWM
/2 Triangle
/2 Sawtooth
/2 Pulse/PWM
X1 Triangle
X1 Sawtooth (the fundamental input)
X1 Pulse/PWM
X2 Triangle
X2 Sawtooth
X2 Pulse/PWM
X4 Triangle
X4 Sawtooth
X4 Pulse/PWM
In the simplest form of this module (listed above), I would like to have a three position switch for each octave to select the waveform and a mix pot for each octave to form a composite signal. Each octave could have it's own output as well. The problem is one of panel logistics - figure five toggle switches, five pots, then add in the PW and PWM pots for each octave and you quickly come up with a vast number of controls. I could easily fit it on one of my panels, but I think most other formats may not be able to accomodate so much control, especially when you throw in all the jacks.
I say simplest, because I have done other things to embiggen these signals, and actually have derived other waveforms. The toggles will ultimately be replaced by rotary switches, and I would have no problem filling up twelve positions on each rotary with the waveform permutations I have accomplished. When I say embiggenment, BTW, that means I've employed techniques to make the outputs sound as if they're not locked to the same frequency - I impose a bit of drift, which embiggens the sound.
There's another direction I also plan to go, which promises to be rather thrilling (at least for me) and somewhat less complex in implementation, but just as rich in output. All of this is why I plan to make the WaveThing a series of different modules.
I should get off my ass and make samples, I suppose....
Cheerios,
Scott |
Quasi is always good! I tried out the "digital-mixer' schematic I found in the CMOS cookbook which gives a similar ring modulatory sound. I will definitely try out the fet/amp combo this weekend.
I was making something similar with saws mixed with sub-octaves pulses that have PWM in a DC mixer. Then using a saw to triangle converter to get sub-triangles. Though I could hear the tick just a little too much, kinda aggravating.
My design is based off the Korg whatcha-ma-call-it keyboard. This lead to way to many controls. I think I'm going to make the panels either 1 or 1 & 2 suboctaves and find a way to gang them together for more sub octaves.
One fun thing I found was modulating the width of the pulse wave with a triangle lfo, then mixing the pulse wave and saw together. It has one of the coolest sounding waveforms I've ever heard.
Did you try getting 3rds or 5ths? |
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Fri Mar 27, 2009 10:43 am Post subject:
|
|
|
Yeah, this would be a bit different than the digital mixer, in that either input can be any waveform you want, with the caveat that the input to the FET should be converted to a negative shifted square wave.
Quote: | Did you try getting 3rds or 5ths? |
Nope.
Cheerios,
Scott _________________ My Site |
|
Back to top
|
|
|
Ricko
Joined: Dec 25, 2007 Posts: 251 Location: Sydney, Australia
Audio files: 27
|
Posted: Sat Mar 28, 2009 9:47 am Post subject:
|
|
|
Scott Stites wrote: |
Actually, what I have on my breadboard is a circuit that derives five octaves of signals from a single sawtooth input:
|
The Roland SH-3A has a function quite like that. (It is missing from the schematics posted on the web.) 5 sliders for levels of different 8ves (2' to 32'') each with a 3-pos switch to select square, 25%? or saw. The 8' has a PWM capability (labelled "Chorus"). Sounds great.
I have a VST emulation of it available at extra.schematron.com
free, with some extra modulation capabilities.
There also was a really rare (Italian?) synth that had two oscillators, each with this kind of mixer arrangement, too. |
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Sat Mar 28, 2009 3:15 pm Post subject:
|
|
|
I wonder if it's anything like the SH1000/2000? Those pretty much do the same thing, only the waveforms (pulse, sawtooth and square) are available in different footages on tabs rather than sliders. Dunno if it's derived the same way; they do it in a way very similar to how Ian's Double-Dekka works, only the Double-Dekka allows one to shape the waveform via sliders.
Funny you should mention it, a video of the SH1000 is what sent me down this road.....
Cheerios,
Scott _________________ My Site |
|
Back to top
|
|
|
loss1234
Joined: Jul 24, 2007 Posts: 1536 Location: nyc
Audio files: 41
|
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Sat Mar 28, 2009 7:05 pm Post subject:
|
|
|
Yeah, Loss, it's a series of circuits more or less, building blocks so to speak. These circuits won't work with anything other than a sawtooth input, I'm afraid. They might do some crazy sh*t to an external input, though.
Synthmonger - I found that if I hooked straight into my D8 recorder, I could hear a bit of buzziness on the triangle waves, particularly the sub-sub - I put in filtration, which cleaned it up quite a bit - I'll have to add that to the schematic.
Triangles alone would make a fairly simple faceplate - one could just have five mix knobs, in input, and what output would be desired. Don't underestimate the humble triangle wave when it roves in packs! Here's my first sample of the WaveThing - it uses all five octaves of triangle waves only. This sample uses only one VCO, processed by the WaveThing. Right now I'm just using my mixer module to mix in and out the octaves. So it goes VCO->WaveThing->Filter->VCA. Then it's going through my Lexicon MX200 delay to space things up.
A static Klee sequence is driving the VCO and EG. The only variation is me mixing in and out the various octaves as the tune progresses. One take, no overdubs, etc.
It was a good way to pass a snowy morning, and it sounds to me like the weather looked outside....
Cheerios,
Scott
Description: |
WaveThing: 5 Octaves of Triangles derived from VCO driven by Klee Sequencer |
|
Download |
Filename: |
WaveThing_5_O_triangle_1_hq.mp3 |
Filesize: |
12.75 MB |
Downloaded: |
1478 Time(s) |
_________________ My Site |
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
|
Back to top
|
|
|
loss1234
Joined: Jul 24, 2007 Posts: 1536 Location: nyc
Audio files: 41
|
Posted: Sat Mar 28, 2009 10:15 pm Post subject:
|
|
|
man this thread has gotten all kinds of ideas rolling in my head
Scott...let me ask you...how do convert a signal (for the circuit you were talking about above) for the JFET into a negative going square?
I was having some trouble with this same issue earlier this season when i was working on an aliaser/ crusher thing...
you mentioned a comparator. so let me see if i understand this...you feed you r signal into a comparator and then invert the out? or you compare it to a neg voltage?
thanks _________________ -------------------------------------------- check out various dan music at: http://www.myspace.com/lossnyc
http://www.myspace.com/snazelle
http://www.soundclick.com/lossnyc.htm http://www.indie911.com/dan-snazelle |
|
Back to top
|
|
|
loss1234
Joined: Jul 24, 2007 Posts: 1536 Location: nyc
Audio files: 41
|
|
Back to top
|
|
|
loss1234
Joined: Jul 24, 2007 Posts: 1536 Location: nyc
Audio files: 41
|
Posted: Sat Mar 28, 2009 10:44 pm Post subject:
|
|
|
i cant wait for the rest of the circuits...very intrigued by your idea of imposing drift, and also your Other waveforms you speak of
Can you possibly explain more on how this works?
i see that the waveform comes in, gets squared and feeds a 4013...i see the u1 is wired up as a differential (?) amp but i dont know much about those...ahh hell...even the idea of a voltage controlled inverter is amazing.
soon to the breadboard!
there are so many ideas!! _________________ -------------------------------------------- check out various dan music at: http://www.myspace.com/lossnyc
http://www.myspace.com/snazelle
http://www.soundclick.com/lossnyc.htm http://www.indie911.com/dan-snazelle |
|
Back to top
|
|
|
frijitz
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
|
Posted: Sun Mar 29, 2009 12:43 am Post subject:
|
|
|
Scott Stites wrote: | In the simplest form of this module (listed above), I would like to have a three position switch for each octave to select the waveform and a mix pot for each octave to form a composite signal. |
Hi Scott --
That's a very clever circuit.
Have you thought any about including dynamic mixing? Say a 3-input MXR/VCA for each octave (to mix a combination of the three waveforms and VCA the sum)?
Quote: | When I say embiggenment, BTW, that means I've employed techniques to make the outputs sound as if they're not locked to the same frequency - I impose a bit of drift, which embiggens the sound. |
So this would be like the sawtooth phase-shift animator?
Ian |
|
Back to top
|
|
|
synthmonger
Joined: Nov 16, 2006 Posts: 578 Location: flada
Audio files: 3
|
Posted: Sun Mar 29, 2009 4:14 am Post subject:
|
|
|
Scott Stites wrote: |
It was a good way to pass a snowy morning, and it sounds to me like the weather looked outside....
Cheerios,
Scott |
I wish it was snowing here ;C Too warm to even wear a hoodie here.
I love the track! Very smooth and fluid like. |
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Sun Mar 29, 2009 7:44 am Post subject:
|
|
|
Here's one with the four lowest octaves set to PWM and the top octave is a triangle. It's some rather in-your-face PWM, driven by a klunky Klee sequence. _________________ My Site |
|
Back to top
|
|
|
|