Quantizer/Analog Shift Register take II

In 1972 as a resident at CalArts, Fukushi Kawakami made four modules as additions to the school's rather extensive Buchla 200 system. The modules are a Control Voltage Switching Matrix, two Control Voltage Integrators and what I believe is the world's first Analog Shift Register. 

http://barryschrader.com/the_fortune_modules/

Since then, the world has fallen into disarray, computers have taken over, analog modulars have gone in and out of favor several times and those four Fortune Modules have ended up in Grant Richter's hands. Somewhere in the middle of all that, Serge made an analog shift register and wrote about it (under the nom de plume Arpad Benares) in Synapse. Even before the Fortune Modules, Buchla had made a rather amazing Control Voltage Integrator called the 155, but that's another post.

Anyway....

Analog Shift Registers are a bank of Sample and Holds. In fact, using only the first output, it is a Sample and Hold. When a pulse is applied, the CV on the input is stored on output one. Whatever was on output one is moved to output two, and so on. 

In 1997 I made 2 copies of a module that was a dual four stage analog shift register as well as an 8 channel voltage quantizer. One of these is still in daily use over at OSI music and the other is rotting on a shelf in my shop. When this photo went around the forums and blogs some people suggested that rearranging the panel to allow the analog shift register outputs to be quantized via shorting bars would be a good idea. Point taken.

 

I had a couple of ideas of my own that could make it a better module. Sadly, it got back-burnered and never saw the light of day until now. The new version has rotary switches to select the scale to quantize to. The ASR outputs can be plugged into the quantizer with  shorting bar. There is no longer a "slave" switch to chain the two ASR's together, but a cable and shorting bar can now do that too. Some new ideas have come up as well, like using the quantizer to look-up the voltages from the "random" voltage sequences from the 266. It's obviously not as glamorous as an oscillator or filter, but it will come in useful to some people.

Musings on Voltage Processors

Some time in 2007, the idea of "building a better voltage processor" came up between Reed and me. The problem goes like this....



256
pros: good for mixing, inverting and scaling. Huge output voltage swing possible, for black knob era systems where you really need 15 volts to open some parameters.
cons: All it can do is mix and invert. Only 2 sections. No indicators. All black banana jacks. (just kidding Rick)


257
pros: able to mix, scale, invert, crossfade, manually offset, VCA
cons: no gain, can only mix 2 CVs. Because it uses PWM for the crossfade, it doesn't like audio rate CVs. Only 2 sections. Confusing nomenclature.

256e
pros: able to crossfade. invert, scale, mix (kinda). Breakpoints!
cons: only will work with up to 10 volt CVs. No real mixing.



The first idea was to make a quad voltage processor module that had 2 sections of 256 clone and 2 sections of 257 clone. That seems like a cool enough idea, but then I thought, "what about making a design that can do what both do and including 4 of those?"

The idea is to have the ability to mix 3 CVs, crossfade, scale (from a knob or another CV), offset. It should have an LED to show it's state. I had a design made up and a front panel worked out, then my laptop was stolen by a Gypsy in Barcelona. These print outs are all that remain. Anyway, I am ready to dive back in and I will get something together.



It will be like a 257, but with 2 reversable CV inputs. The PWM crossfader will be made from newer, faster parts. So the cutoff frequency will not be so low. The 257's triangle oscillator, for the PWM, is 21kHz. I think I'll try to take it up to 100kHz or so. This shouldn't be a problem for modern analog switch chips. I think I'll use center detented pots for the reversing CV inputs. Here's the 257 schematic, BTW.



With this configuration, you can mix 3 signals, one with voltage controlled gain. It's not the 4 signals of a 256, but it's close. It can do everything a 257 can do, with a bit more speed. It's 4 processors in one module space. It'll have LEDs on the outputs. the output will not swing a full 15 volts, because I don't want a 24 volt supply line (200e and many 200 systems don't have one). It won't even try to do the breakpoints, we'll leave that to the 256e.