Help me build Buchla/Serge style Capacitance Touch Keyboard?

[a.b.o.z.]

anyone ever built synapse circuit from boards akrearke posted?

[TheAncientOne]

The Japanese/Buchla circuit is intriguing. I built a few things like this, when I first did synth DIY, back in the 70's though I got away from analogue and drifted in digital getting interested in things like PRBS circuits and early hybrid oscillators.

A few thoughts.

The twin-Tee oscillator should, in theory run at about 10KHz. R3 seems tiny though, a more usual value would have been half of the 15K in the other limbs. I wonder if it's a misprint, and should be 6K8.

R7 is most definitely a misprint or copier error. There were never any easy to get 60R resistors, 68R or perhaps that is actually the 680. At max collector current, 68R makes about 100mV difference at the emitter.

The response is going to be influenced by the number of stages, and the characteristics of T4. The initial bias for all the sensor circuits comes from the roughly 5V at the junction of R11 and R12. All the base currents, (low though they are) are coming through R11, and will vary a little with the setting of the trim pots.

If I do breadboard it, I'll start with the oscillator and one stage , just up to T5, (the other transistors are just buffers and LED drive). I might make a small bet that T4 may have to be selected or graded, and I suspect that the setup procedure might involve shutting down the oscillator to set trim 2, before adjusting T1 with it one. I also feel that adjustment of anything with a 47pF series cap is going to be 'touchy' in quite the wrong sense.

It's a very early design, and bits of it are quite horrible, (Buchla or not - we all start out messy), and due to the capacitive coupling effect, it probably won't be consistent day to day, or person to person. Also, in those days, EMC problems were minimal. A mobile phone might play havoc with the operation if it was nearby. A modern version might have to include some small chokes, to kill off RFI pickup.

On another vein of touch keyboard design, the use of the 74150 is due to an exploitable 'quirk' in the input stage of 74 series TTL logic chips. Back in the 70's this was a well used trick, but it won't work with CMOS equivalents or even LS (SChottky) TTL. It was used in the Synthi KS keyboard too, and a few magazines did touch controllers using the trick, though often using (I think I'm remembering right), things like the 7404 inverter. The use of the 74150 is why the KS has 32 keys, (2x74150 for the notes, and another for the scale shift keys). If you get the KS drawing from Graham Hinton's website, you can see that the 'start', 'stop' and 'random' pads use one input of a 7400. By the way, I think I can see at least one error in that drawing; The capacitor connections to the scanning oscillator, E28, near the bottom on the left hand edge seem wrong, I think the 330uF is not in parallel with the 1uF, but to a separate pair of inputs to provide a start delay. If you want to make a 'Synthi' style keyboard, there are suitable shift registers still around, though a bit of extra logic and a CMOS RAM chip might be just as easy a redesign. An old 6116 would give 2K steps, or better time resolution.

I would be interested in building a latter day KS type keyboard. Using a more recent touch trick, and a cheap micro for the scanning and sequencer. If it were to use something like Arduino, then we'd have little sequencer factory, where users could alter the code to get effects they wanted, like ping-pong and pendulum, or simple arpeggios. Something cheapish to make to give us all a simple test/portable keyboard.

Back in the day, I made my first keyboard using self adhesive copper foil. Provided one could connect to it, the self adhesive aluminium foil, like the tape used to alarm windows, could be laid up as pads, say stuck onto perspex, with a layer of adhesive clear film over the top.

A few thoughts anyway.

[ericcoleridge]

[piedwagtail]

...Most touch sensors should be debounced with a long time
constant, preferably a second or so. This prevents any nervousness or
hesitancy from being entered as multiple inputs. Always use only
the absolute minimum sensitivity you need for any touch sensor and provide as much protection as possible to the sensor.

A capacitance-operated touch system for electronic-music
keyboards is shown in Fig. 5-16. An input square wave is used as a
clocking signal. Normally, the inverter delays the clocking until after the
D input has accepted a new value, so an output low results.
Touching the sensor at the D input adds 300 pF or so of capacitance to
ground. This gives us a resistor-capacitor delay network that slows
down the waveform reaching the D input. In fact, it slows things
down so much that the clock gets there first, giving us an output-
high condition. The input and output waveforms for this circuit are
shown in Fig. 5-16B. The choice of clock frequency determines the
debouncing you will get. Always use the lowest possible frequency
for any particular application. For many electronic-music
applications, 500 hertz or higher is a good choice.

Don Lancaster

http://www.amazon.co.uk/CMOS-Cookbook-DON-LANCASTER/dp/0750699434/ref=sr_1_1?ie=UTF8&qid=1323390355&sr=8-1



Clock buffering will be essential for each key and post buffering for the 4013.

.....sequencer factory.

Robert

---

Shot020.png
Description:
Filesize:	78.88 KB
Viewed:	178 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[xi-boy]

i am probably blind but i cant find the type of the three transistors in the synapse touch circuit.
can i use bc547 transistors for them ?

edit: bc547 work fine

[squarewhite]

oh yes, if someone is making the pcb, i'm interested! would be nice either in 8 or 16 keys. let me know.

[xi-boy]

[squarewhite]

there's an error with the link. Can you send again?

[xi-boy]

[squarewhite]

thanks! the link works!

best