Back to Projects
The Electronotes ENS-76 digital keyboard appeared in EN#68 and my re-design appeared in EN#92.
A single-buss keyboard was scanned by four 74150 16-to-1 multiplexers into four inputs of a 74151 8-to-1 mux (not shown here). In my revision, the buss connects to ground through a 47 ohm resistor. This is low enough to be a logic '0' to the 74150 inputs but the current from the '150's input pullup resistor (61 of them for a 5-octave keyboard!) of any pressed key causes a voltage drop across the 47 ohm. This is detected by a LM311 Comparator and supplies the Gate or "Any Key" signal. I thus eliminated the rather intricate and overly complicated, in my opinion, method of deriving a Gate used in the original design. Schematic
OpAmp OA-1 is a differentiator. It issues a brief pulse whenever the DAC voltage changes OR the Gate rises or falls. This pulse triggers the first 555 and the falling edge of that triggers the 2nd 555 ONLY IF the Gate is high. This is the Trigger signal, which is actually slightly delayed from the rise of Gate. I see now that I might've eliminated the 2nd 555 by connecting GATE to pin 4 of the first 555. (Or maybe I tried that and it didn't work reliably. I don't remember.)
Resistor R1 at OA-2 should be a bit less than 'R' in the R/2R Ladder so that trimpot TP2 can be adjusted for 1V/Octv. OA-3 inverts the voltage from OA-2 to give the proper polarity and resistor Rx value is selected to give the desired RANGEing. OA-4 completes the CV circuitry with Portamento.
The original design used two 7475 4-bit latches, followed by a 7404 and a 74C04, inverting the bits twice to buffer the latch outputs and provide 0 and 1 levels to the R/2R ladder more consistent than those of TTL outputs. Somehow, the fact that the 7475 has /Q outputs, needing only one inversion, was overlooked. I used 74175s because that's what I had. A 74C174 6-bit latch would've further reduced the parts count because its outputs would not need buffering. Perhaps that function in CMOS didn't exist yet.
I also made changes to the original Clock and Counter circuitry, providing a Data Enable (DE) strobe that occurred during the 2nd half of each count. This eliminated the (again) rather intricate switch de-bouncing method of the original, which relied on 555s too much for my taste. It was bad enough that I'd used two of them in the trigger circuit.
I later crippled my brilliant re-design while preparing it for future 4/8-voice operation -- which never materialized :-(