Differentiator VC Feedback Project


So, I want to make a Differentiator pcb with additional voltage controlled feedback.
I normally feedback it with a summing mixer, but I think CV control will make it even more interesting.
The effect of the feedback is that the edges repeat themselves several times until chaos ensues.

I paired the Cadet Multiplier schematic with the Differentiator and changed some stuff
The design is not finished yet and must be tested.
Right now I’m thinking 1 pcb for pots and jacks, 1 pcb for the main circuit and a panel pcb.
All sandwiched together to decrease depth. 8HP if it fits

  • feedback path switch (normal / inverted) - be to tested
  • 4 edge filter inputs (instead of 6) - to save space. If it fits, this can be changed to 5 or 6.
  • switches for feedback routing
  • not added yet, but interesting: diode clipping distortion or more filtering in the feedback path?

example of feedback edges: (in a patch, no cv yet)
The white dots = edge of the green shape routed though a Differentiator and feedbacked with a modified summing mixer.
(red & blue channel = camera feedback)

Remark: this project is directed towards DIYers and will be open source.
—>> All input and help is welcome!



Current schematic:

This will be my first time to do a pcb project with sandwiched pcb’s.

note: I left the 4 quadrant switch in. this has cool effects



testing with separate modules. fast screensots
The dots appear when the CV in modulated with a VCO.
The signal needs some boosting to get the repeats, so I have to add this in the schematic (or change some values / make a gain control

The feedback path routing adds nice features, so must be implemented on the front panel; with a rotary encoder or multiple positions switch (4x1)

When several inputs are used, the image get more interesting details (picture 2)



Complex CV possibilities.
pic1: minimal setting
pic2: half way
full on: chaos (not shown)



a 100k pot (VR3) in the amp stage of the differentiator works pretty good.
it needs a range limiter (r38) for the low end, which value I have to determine yet.
The Chaos tipping point is tricky. but will be solvable with the right resistors.

I’m getting a lot of cool images already :slight_smile:

I wonder if r41 & r42 (or r6) can be skipped. they were final outputs, but not anymore
the same with r44, r47, r45, r46.

and is r3 still necessary?



the differentiator is one of my favorites I think of it like a little utility mixer basically

a similar idea to the nearness module but for video

this module seems like a neat furthering of that concept. I’m always up for more feedback I’ll happily be following along but have no technical knowledge to lend



This is magnificent. I can’t wait to see how this ends up. Sign me up for one when it’s ready!



Following with intense interest!



Looks great! As far as skipping some parts, I would leave those resistors there to provide some isaolation from the contacts of the mechanical switches. (Are you using a rotary switch here?)



The one in the middle is a toggleswitch,
The 4x1 will be either a rotary or a 4position slide switch.

What about r3?

1 Like


Only sort of half understanding what all is going on here, admittedly, but I wonder if there is a way to introduce variable time/delay to the feedback path for thicker & thinner edge feedback effects (if edge feedback is indeed what is happening here)?



@reverselandfill R3 keeps the input of U3A from floating while the switch is momentarily in a no contact state (which will prevent ringing or other problems you want to avoid.)



that makes sense, I had not thought about it that way.

Do you see any other points that can use improvement?



R7 and R8 should be ferrite beads and not resistors.
It would be ideal if the switches were replaced by analog multiplexer (74hc4053, etc) but then you need to add a +5V power supply and some input protection to the lines. For a DIY design I would not worry about it. For a production design, it’d be worth the fix.

What is the intention of the SW and SW2 pads and the switch S1? My suggestion would be to turn the LT1251 into a 4 quadrant multiplier – basically crossfade between the inverted and non-inverted signals. U3A and U4B could be eliminated, as you would send the differentiator buffer amp (U4A) to the non-inverting input on one half of the LT1251 and to the inverting input on the other half of LT1251 in parallel.

1 Like


If you are looking for more ideas for functions, one would be to use a rectifier to split the positive and negative sides of the differentiator output and make them selectable (left edge only, right edge only, etc.)

1 Like


r7 & r8 are already ferrites, I don’t have footprints for ferrites :slight_smile: