A 4046 based synth demonstrator

[anotherjim]

Not a pedal - too many controls for that. Only 2 chips tho' 
Schemo...



... edited. Not too big now. Right click & view image for biggerer.

Sounds -  one track for each of 6 wave sources....
Soundcloud playlist...
https://soundcloud.com/ashdalestudio/sets/4046synth

[soderstrom]

Great! 

[garcho]

awesome sounds!

[Kipper4]

awesome sounds!

Agreed Garcho.
I like 2 & 4 best.
No2 kinda reminds me of throat singing. Check out the Mongolians and Tibetans.
Good work Jim and thanks for sharing.

[anotherjim]

Thanks for the comments.
This is why I DIY, I can have something the mainstream producers wouldn't look twice at 

FYI, I played it from a Telcaster. Preferring neck p/u and tone low. Finger plucked (tips not nails) over the neck joint. No overdubs or other FX - it really can do several things at once. Controlling them is another matter.
Be warned, it can take several days to learn to drive this thing.

I think a description would be in order, it is long....

The Input:
A simple comparator using an op-amp. The output is a square wave. TLC272 used a CMOS type, as it can work with a ground reference. The output is a tone source for the filter as well as feeding the detector, the sync pulse and the 4046 signal in.

The Detector:
Actually isn't present on the unit the sound samples are from, it is only on my breadboard version at the moment. It's a basic diode charge pump. The voltage on the cap disables the mute transistor when signal is present. It's intended to mute when the input has decayed too much for the 4046.

The Sync:
A short pulse (the shorter the better) developed from the input positive transition momentarily resets the 4046 VCO while high at pin 5. It can with the wrong settings mute the VCO if it never completes a half cycle before the next sync pulse. With the right settings it can kind of PWM the VCO waveform or impose some of the input frequency when the VCO is out of track. Ideally, there would be a schmitt trigger logic gate involved to get a really fast pulse. The opamp is doing about 8us rise time at best and that limits how short the sync pulse can be. Also, while high, pin 5 disconnects the Mod source from pin10.

4046 Phase comparator:
Pin 13 provides the control of the VCO frequency. It compares the input with the VCO output and if different, either drives pin 13 high to make the VCO faster, or low to slow it down. If the VCO matches the input, pin 13 disconnects. A variable low pass filter turns the correction pulses to a smoothed DC average. With the Slew control large, there is a delayed reaction and the VCO frequency will cyclically over and undershoot the required frequency - hunting - producing a Thermin like warble. With Slew small, the warble is at audio frequency producing ring mod type sound. An additional control, Damp, can reduce the time delay to limit the warbling, but when the Phase Comp' does correct, the pulse rapidly changes the control voltage which can be a cyclic juddering in VCO tone.
With no input signal present, the VCO frequency will be driven to minimum and cut off.

4046 VCO:
The chip contains a voltage controlled oscillator. It has a frequency cap between pins 6&7 and a frequency range resistance off pin 11 toward ground. The cap value may be critical, I've tried a few different values from 4n7 to 10n with 2 different brands of chip. Switching cap sizes might be a useful addition. Varying the range resistor can force the VCO out of range of the input. The VCO control voltage is from a variable low-pass filter fed from the chips Phase Comparator.

Mod:
The VCO control voltage as a filter mod source. The tracking can be set so bad that this is an audio signal and will bleed thru the filter adding it's own sound.

The Tone Sources:
Fuzz:
The direct output of the input comparator. Some of the sounds in the samples of this don't even use the 4046 except some with the Mod switch on. Then the VCO control control voltage derived from the 4046 demod output (a buffered version of the VCO control voltage) sweeps filter frequency which is at its most useful with settings that cause the VCO to mistrack very badly!

Square:
The VCO output which is a square wave.

Saw:
The VCO timing capacitor wave. Pin 6 & 7 each produce a sawtooth waveform. These waves appear half wave rectified and offset from each other by half cycle. If they were combined, you would have a full wave Saw at twice the Square wave frequency. That isn't done here and pin 7 seems happier to drive a load than pin 6 for some reason although ideally a high impedance buffer should be used to prevent an external load effecting the VCO frequency.

Pulse:
Whenever the Phase comparator is correcting the VCO, there is a negative pulse on pin 1. Intended as a "lock indicator", this output is used as a tone source.

2f:
There is another type of phase comparator in the chip outputting on pin 2. It is a simple logical Exclusive OR of the input and VCO waves. When there is a difference, it can produce 2 pulses per cycle, hence I call it 2f. Octave up seems to be more obvious when a small VCO timing cap is used.

Mix:
This is the output of the phase comparator from pin 15. I call it Mix because it can be a complex combination of input and VCO signals.

The Pedal:
This option uses a synth type expression pedal. An on-off-on switch allows the pedal control voltage to affect either the VCO or the Filter. When applied to the VCO, it tends to force the VCO frequency away from tracking the input. This can create a harmonic sweep effect. I've used a TRS socket for my Roland pedal. Be aware that other brands can use different wiring to the jack. Roland is Sleeve=Ground, Ring=+supply and Tip=Control pot wiper.

The VCF:
Tim Escobedo's  Q&D filter - the non-oscillating version, since we don't have a VCA to shut it up. I've reversed the values of the filter caps - it sounds better to me that way. The input is extremely sensitive, so the source input is heavily padded down. Even so, the Drive control can -  you guessed it. It's output is pretty hot too, so is also padded down to the volume control.
With drive at minimum, you can still get audio if the Mod switch is on and VCO isn't tracking. The control LED seemed best with an ordinary 5mm Red 20mA type.

[deadastronaut]

f.....g awesome jim,..


just played your clips through mixer with delay....

nice work man..


i'm guessing it might be less glitchy with a comp in front?...very cool either way. 

[anotherjim]

Cheers Rob! Praise indeed from the demo master!

[cloudscapes]

good stuff, love how the filter sounds! raunchy!

this is something I want to do a bit of this year, cmos based synths (aka lunetta synths). I could program it as a far simpler application in a microcontroller, but that's not nearly as fun. and sometimes when you hook up the 'wrong' parts of cmos together, you get really cool sonic results! something hard to emulate in software.

[anotherjim]

I made no attempt to clean up the front end. It should be done, but it's a lot more work. Not for nothing have I called it a demonstrator. Idea is  - breadboard it, and see if there's a feature you'd like in a pedal. Using open preset pots and jumpers for switches, it shouldn't take long to assemble & then find the minimum control set for what you need. It's a big help to use an expression pedal.

It should be possible to make it nearly 100% 4000 series CMOS. I have seen a synth forum post somewhere that claimed the Q&D filter was possible with inverters, but not seen a schematic for that. I do suspect the 4046 input is prone to jitter with slow moving signals, so a schmitt trigger gate feeding it might well be a benefit.
On that, the Q&D isn't the deepest  filter out there, but I had one half of the op-amp unused. It does benefit from a rich source and with an expression pedal, would make a useful filter stomp box with not too many controls if used after a distortion.

[Marcos - Munky]

Very interesting sounds!

[pinkjimiphoton]

Not a pedal - too many controls for that. Only 2 chips tho' 
Schemo...



... edited. Not too big now. Right click & view image for biggerer.

Sounds -  one track for each of 6 wave sources....
Soundcloud playlist...
https://soundcloud.com/ashdalestudio/sets/4046synth

[Ben N]

Amazing! Jim, have you done anything else with this? Also--it doesn't look extremely difficult to lay out, but if you have anything...

[StephenGiles]

This might be of help.

[anotherjim]

Amazing! Jim, have you done anything else with this? Also--it doesn't look extremely difficult to lay out, but if you have anything...

I never draw layouts - just build from the schematic until its all on the vero or perf. With the rotary switch and all of the pots, there's a mess of wires. I put it in a project box - kind of a "squeeze it all in, put the screws in and don't ever open it" affair.

[StephenGiles]

Amazing! Jim, have you done anything else with this? Also--it doesn't look extremely difficult to lay out, but if you have anything...

I never draw layouts - just build from the schematic until its all on the vero or perf. With the rotary switch and all of the pots, there's a mess of wires. I put it in a project box - kind of a "squeeze it all in, put the screws in and don't ever open it" affair.

So do I, although my first ADA Flanger vero build was huge to allow trouble shooting if needed, then I rebuilt to fit in an old EH Memory Man case!

[pinkjimiphoton]

So do I, although my first ADA Flanger vero build was huge to allow trouble shooting if needed, then I rebuilt to fit in an old EH Memory Man case!
[/quote]


holy shit Stephen!! on vero? you may well be my new hero, dude

[StephenGiles]

There's a picture floating around somewhere!

[anotherjim]

I found a photo of my build.  Rockin' the '80s hobby magazine look. I may never find a knob for the D shaft rotary switch that suits it.

[Ben N]

Rockin' the '80s hobby magazine look.

Oh, my God, yes!!!

[Uoki-Toki]

Where is the output "Meditate" connected?