Parallel Tremolo /Ring Mod Chopper Brainstorming

[Mr. Lime]

Hi guys,

going through different tremolo and ring modulation circuits I'm currently thinking about a new project with parallel chopping paths and individual frequency controls. Probably even a modulation of the frequency controls but first things first.

Can anybody recommend me an oscillator with very wide frequency control to archive slow tremolo to ring mod oscillation? I think 3,5 Hz - 500 kHz would be perfect. How about the NE555?

My first attempt would be two CD4060 binary counter oscillators with switches to select divisions of the frequency range. Then I'd like to convert the square wave outputs to triangle and sine waves which control JFETs as varaible resistors in the parallel signal paths.

The selected divisions set the ranges of the oscillators and can be anything from tremolo typical phasing to ring mod kind of chopping.

There are still some points to think about.
Should the input stages be out of phase to each other to do some frequency cancellation? Then lowpass and highpass filtering on their inputs should be done.
Are there more preferred ways to convert the sqaure waves into other wave forms?


Thanks

[anotherjim]

Can anybody recommend me an oscillator with very wide frequency control to archive slow tremolo to ring mod oscillation? I think 3,5 Hz - 500 kHz would be perfect. How about the NE555?

I suspect nothing will do that range from a single control input and range switching (cap values) would be needed. Of the chips I know, the widest control sweep is the CD4046 and of those, the Philips HEF parts are/were best for me. It is possible to manipulate and stitch the (buffered) VCO timing cap waveform which is a disjointed sawtooth or triangle depending on how you look at it. See the Thomas Henry X4046 VCO for ideas.
https://www.birthofasynth.com/Thomas_Henry/Pages/X-4046.html

[Mr. Lime]

I suspect nothing will do that range from a single control input and range switching (cap values) would be needed. Of the chips I know, the widest control sweep is the CD4046 and of those, the Philips HEF parts are/were best for me. It is possible to manipulate and stitch the (buffered) VCO timing cap waveform which is a disjointed sawtooth or triangle depending on how you look at it. See the Thomas Henry X4046 VCO for ideas.
https://www.birthofasynth.com/Thomas_Henry/Pages/X-4046.html

Well range switching through caps is fine, there is no need to really cover the whole range with only one turn of a pot.
The Thomas Henry VCO indeed looks good but the wave shaping seems to be a lot of work.
The oscillator with very low parts count and great range I stumbled across is the ICL8038 with direct outputs for sine, triangle and square waves.
There are even options for wave distortion so this one looks very promising for my application.


Frantone Vibrato has an ICL8038 as well. Frequency modulation should be doable.

[ElectricDruid]

I think 3,5 Hz - 500 kHz would be perfect.

Do you really mean 500KHz?

3.5Hz to 3500Hz is 1:1000 range - easily doable. 3.5Hz to 35KHz is 1:10,000 range - quite possible (there are plenty of VCAs that offer this range of control, for example - use one as the frequency control element in your oscillator).
3.5 to 350KHz is 1:100,000 range - very difficult to do without range switching, and probably extremely over-sensitive to control if you did manage.

T.

[anotherjim]

You may find the XR2206 easier/cheaper to get.

[PRR]

> 1:100,000 range - very difficult to do without range switching

Dim memory says, IIRC, Bob Pease had an VCO with near that much range, based on log-junctions.

This is not it. This describes linear VCOs with 120dB (1,000,000:1) range after trim:
http://www.ti.com/lit/an/snoa594b/snoa594b.pdf

[ElectricDruid]

> 1:100,000 range - very difficult to do without range switching

Dim memory says, IIRC, Bob Pease had an VCO with near that much range, based on log-junctions.

This is not it. This describes linear VCOs with 120dB (1,000,000:1) range after trim:
http://www.ti.com/lit/an/snoa594b/snoa594b.pdf

Fair point. 1:100000 is 100dB, and there are plenty of VCAs that could pull that off, so you could design a VCO around them and you'd be there. Still, the control is going to be over-sensitive.

120dB is really going for it - still, Bob Pease was probably the person for the job. And the control problem would get another order of magnitude worse! You'd need a pot with a knob the size of a dinnerplate on it!

[Mr. Lime]

I did some math based on the Phonic Taxidermist (Maplin Voice Vandal) circuit which I have built.

If I'm correct then the CD4060n divides 4MHz into 4 divisions:

/16: 250k
/32: 125k
/64: 62.5k
/128: 31.5k

Those frequencies are selected by DIP switches and chop the signal.


If I'm using the XR2206 from the "Superfast Tremolo" thread as oscillator I would have a frequency range of 0.002 - 7.2Hz in Tremolo mode.
With a 10n cap switched the new range is 1.5 - 93 Hz.
So a 10p cap could give a range of 1.5k - 93k Hz which is probably good enough for ring mod kind of tones, isn't it?

Maybe such high frequencies aren't even needed, and my assumption of the 4060n values are completely wrong!

[ElectricDruid]

Why do you need the modulator to go above audio frequencies? Does that do something special?

[Mr. Lime]

Here's the schematic for the Phonic Taxidermist:


The oscillator chops the guitar signal, that's all..


f osc = 1 / 2.5 (R1xC1)

[ElectricDruid]

Love that multiplier. Ingenious. Deranged, but ingenious! 

[anotherjim]

With such high chop frequency, doesn't it rely on the clipping diodes to produce supersonic harmonics?
I wonder why the 4060 runs on 5v so needed the 4 transistors to level shift to the 4066?

[Mr. Lime]

Is this really a multiplier?

I think the input signal from a clean guitar is too weak to get clipped that much by the diodes..
The effect sounded best with no overdrive at all.

For what the transistors are there I have no clue. Mostly I used only one chopper at a time and left the other three switched off.
A simplified version with only a J201 JFET as chopper, which is directly connected to a oscillator might be sufficient. The delay part is anyway kind of crappy and can be left out.

Again I might be wrong but I think the freq range of the divisions with the 1M pot should be:

2k5 - 250kHz
1k5 - 125kHz
625Hz - 62,5kHz
312Hz - 31,25kHz

Now I'm interested if I would have a state variable filter and take outputs from HP and LP and chop them differently before they are summed up.
This could be done with either phase inverted or phase uninverted outputs.
Clean or band pass output could be used to blend some clean content in.

Any ideas if that's useful sounding? 

[anotherjim]

Looking at it a bit more, the inputs aren't meant for guitar, although the mic input gain doesn't seem that much. It's probably a riot with a drum machine.

A single chopper is a kind of multiplier by virtue of aliasing? Sub and difference frequencies. If you chop too fast, it's just a sampler and the sampling rate is removed downstream. I think that's why Tom was asking why the rate needed to be so high.

With 4 groups is there something else happening? The counts are altering the amplitude out as the 4 otherwise identical paths are taken in and out of parallel connection, and this produces an AM envelope at the slowest count rate (Saw-Ramp?). This may explain why the maximum counter clock rate can go so fast.

[Mr. Lime]

Jim is right, this effect wasn't designed with guitar in mind but I think with some tweaks it may be quite suitable as guitar ring mod.

There are a few videos on youtube which show all choppers switched on.


Taking the idea a little further, here's the filter part.
Outputs of LP and HP are used for the choppers while the BP is used for a kind of notch depth control.
The tonal possibilities should be very wide and it might be beneficial to add vactrols and an expression pedal to control the filter sweep.

Typical harmonic tremolos would only use one LFO and phase invert the swing of one JFET and I'm confident that this circuit could be a great base for that.
Not sure if it's capable for parallel tremolo and ring mod swinging, any opinions on this?




EDIT: Lol I realized that I'm really only looking for a bitcrusher circuit.. 
At leas a musical bitcrusher for a guy who thought he doesn't like bitcrusher circuits!

Anyway some suggestions would be very welcome! 

[Mr. Lime]

I decided to go another route and chose LDRs over JFETs.

Here's what I'd like to try, prototype PCB is on it's way because I hate breadboarding (and breadboarding hates me).

Still I'm not sure which LDRs exactly I should use for the highpass and lowpass outputs to work best with the bandpass-depth control..

Any advices for hte maximum off-resistances?

[Kipper4]

Have you tried the electric Druid Stomplfo as an lfo for the vactrols.
Great bit of kit for the footprint size.

[Mr. Lime]

Yes, actually I'm going to use the TapLFO 3 for this circuit.

[Kipper4]

As for the ldr it depends on what you want the last stage to do. Boost, unity.
Maybe experiment from a 5k bright up.
Might be worth experimenting with a variable CLR on the led too. Series 1k-10k variable-rail.

GL55xx series may vary quite a bit for ldr. Some matching may be required.
Might be worth popping an nsl32 vactrol too and again play with clr.

Does the tapflo 3 have a 50% duty cycle and 100%

Sorry for all the questions

[ElectricDruid]

Does the tapflo 3 have a 50% duty cycle and 100%

The TapLFO has a "wave distortion" CV input which alters the duty cycle from about 5% through to about 95% (can't remember the limits exactly - it stops a few percent from each end). This affects all the standard waveshapes, not just the pulse - hence the re-naming of what would have been a "pulse width" control to a "wave distortion" control.

Further details:
https://electricdruid.net/datasheets/TAPLFO3Datasheet.pdf