Friday, 14 October 2011

Bitsy - Stepped CV Generator/ Recorder

A DIY Ring Counter based on Ken Stone's Gated Comparator circuit.



My inspiration for this module was Grant Richter's Wiard Noise Ring - a random CV generator with a form of memory. While I liked its ability to capture an eight note sequence, I found it hard to control and wanted some method to reliably input my own data.

I first thought an analogue shift register - a chain of sample + holds - might provide the solution. Thanks to a prod in the right direction from Matthias of Fonitronik, I discovered Scott Stites' Klee sequencer and later Rob Hordijk's Rungler and Ken Stone's digital shift register circuits.



Using a CMOS chip like the 4015 or 4021 won't allow you to sample your own CV stream, but the end result is the same. A series of zeros and ones is converted into a stepped CV via a DAC, in this case an R/2R network. Feeding the last stage of the shift register back to the input allows you to hold a sequence. Here's how it sounds:



I built Bitsy as a companion to the Wogglebug, which provides the clock and data, via its burst output. Like the Klee, it can also be programmed via a sequencer or manual gates. To simplify the design, I left out the input comparators of Ken Stone's circuit. A switch is used to manually hold a sequence. The last stage is also available, allowing VC over new/ old data selection via a sequential switch like the Doepfer A-151. An LED reflects the voltage level.



I used the DIY Layout Creator to plan my perfboard and a 6HP Makenoise blank as faceplate. This was the first time I'd made a parallel board, which meant using wire connections for the jacks etc. However, the benefit of this was I didn't have to be totally accurate when drilling holes.

Thanks to Matthias for the hint and to the designers of the original circuits which inspired this project!

6 comments:

Brian said...

Hey nice! I've been thinking about my Noisering a lot the last few days due to a thread on Muffwigglers. Just this morning I was eyeballing the Chance Input and wondering why in eighteen months I've never been able to make sense of what it does. I've input various other CV sources, but I've never made sense of what's going on -- and rarely does any input there seem to affect the output anyway. Maybe this is what you mean by the Noisering being "hard to control?"
So I'll ask, what exactly are you shooting for when inputting your own data, and what data are you expecting to input? An sequence of some kind? Or CV to influence the sequence that your bitsy outputs? Again, this is one area of the Noisering that is a mystery.

Citric Acid - Acid Jack Rec. said...

Hi,

nice..can you please share your layout or do you plan a comercial use for it ? will be nice to rebuild it :-)

Anonymous said...

Sounds very nice, what fx are you using for that nice reverb / delay sound?

Navs said...

Cheers!

Brian, the main problem I had with the NR is that it has a mind of its own! Setting the thresholds for Chance and Change is fiddly. In a NR thread at MW, Grant describes a 'compensation' circuit that injects a new bit of data into the string when the shift register is either full or empty (all ones/ all zeros). This makes sense musically. Ken's circuit doesn't have this, so Bitsy and the Gated Comparator can 'flatline' i.e. there is no movement.

In terms of data, you could imagine an 8-step analogue sequencer: each step can be programmed to be low (off) or high (fully CW). Send this as digital data to a shift register and you have a measure of control over the 'notes'. This is the principal behind the Klee, as I understand it.

BTW, Bitsy isn't meant to be an improvement on the NR, which is a complete instrument/ system in itself. It's just that once I started researching the topic and understood its workings, I thought it would be an interesting project.

Navs said...

Citric Acid, the schematic is Ken Stone's Gated Comparator, minus the input comparators and individual stage buffers, LEDs & switches.

The clock and data inputs each go to their respective pins via a 1K resistor and diode, even though the 4015 is supposed to be protected. If you have the space, buffering/ squaring-up the inputs is probably a good idea.

Q4B is 'multed' to both the R/2R and the switch for the hold function. You could buffer this using the second half of the output TL072, but I found it worked without.

Pay attention to the 4015. The first time I powered up the circuit the loop function didn't work. I thought I had made a soldering error until I swapped the chip for the one I had used on the breadboarded version - this worked.

Navs said...

As with nearly all my sounds, the delay is the Electro Harmonix Stereo Memory Man w/ Hazarai ;)