Monday, 22 August 2016

Turning drawings into control voltages

This part of the design took months and months. Turning felt tip pen drawings into digital logic is unfeasibly difficult! I went through about three or four prototypes on breadboard with varying degrees of success. I had a lot of very generous free assistance from Robin Iddon and Roger Dealtry, who coached me on the necessary logic and prioritisation circuitry through these experiments, and I got very close, but not quite close enough. I was trying out different types of flip-flop circuits: SR, JK, and D-type. Essentially I needed a circuit which would read and latch onto the last read input, but that would also reliably reset all the other inputs so that only one input would be 'read' at any one time. From there I could use analogue switches and resistor ladders to select and output control voltages, which would then be used to control pitch and volume etc. In the end I had to go back to the drawing board and old fashioned logic timeline graphs to analyse what was going wrong, and then what needed to happen instead, and finally the penny dropped and the optical reading circuits were born. I wanted them to be simpler with a reduced gate count, but as I had finally something working properly I just went with it rather than trying to make it less complex first. Once the circuit was working I took the time to design proper PCBs, as I would need seven of them for the final machine. This was by far the most complex circuit I had ever designed and also by far the most complex PCB design I had made, so I had my fingers firmly crossed as I switched on the PCB version for the first time. I had about three hours of exasperation as it just would not cooperate, and then I realised I had not soldered one tiny resistor which was essential for the start up process (a simple RC sub-circuit to overcome start-up glitches and to make sure the machine came on silent awaiting the first instruction). After fixing that, and a doing a bit of calibration on the optical inputs it finally worked. I made some small adjustments to the layout and then ordered one PCB for each of the input parameters of my Mini-Oramics machine.

The first optical reader PCB. The multipin header sockets connect to the phototransistor readers and also the indicator LEDs which allow you to monitor if the drawn inputs are being read correctly:

Last of the breadboard prototypes:

No comments:

Post a Comment