It’s been two and a half weeks since my last post and there have been many subtle, and not so subtle, changes to the design of the Magnetic Cello. Lets start with a dramatic (and still in progress) change to the body of the instrument:
This is a model of the neck and body of the next iteration of the Magnetic Cello. Lawrence Le, an architecture student, designed the shape and laser cut 1/4 in. board to build it. There are many changes that need to be made, but we’re going to give the cello a light, strong, ergonomical, and good looking design.
Having Lawrence’s help on the physical has given me more time to work on the electrical. Here’s where the subtle comes in: most of my time these last few weeks have been spent optimizing and making small changes to the string selection and pitch circuits. I replaced the optoisolator string selection system with a demultiplexer, meaning less power drain and a simpler circuit. I used the zero adjust of the voltage controlled oscillator to add a knob that controls the absolute pitch of the instrument while retaining the relative pitches between the four strings. And I tested the instruments ability to stay in tune over a few hours of play or a few days of rest. None of these changes are huge, but they are steps to creating a refined circuit.
Talking about refined circuity, one thorn that has always bothered me is the quality of the voltage controlled amplifier. In my last post, I was able to combine the signal from the oscillator and coil in a way that could be picked up from any audio amplifier. It has noise problems, but it seemed passable. But after testing the circuit again today, it turns out that there are major reliability and noise issues with using optoisolators to short out a current to voltage converter. May be why no one does it that way. So I pulled out the oscilloscope that the electrical engineering department is lending me this summer.
After alot of furious googling and quite a bit of guessing, I realized that it would be a much better idea to bring in the FET transistor output of the oscillator before the operational amplifier, using a voltage divider. The results were noise but consistent. After a few shots in the dark, I found the right places to place capacitors or diode to cut out noise. The result is this:
Surprisingly, this setup has very low noise, and there is no hum when the instrument is on but not being played; I actually accidentally left the circuit on because I had not heard the gnarled sound that usually pops up when no note is played. It uses the internal circuitry of the VCO chip, meaning low part count and simple design. And on top of it all, some experimentation with the oscilloscope showed that the capacitor across the operational amplifier (the component with the value “To Taste”) could be used to control tone. A small value leads to a reedy square wave, and a large value leads to a more fluty triangle wave.
I’ve said this before, and I’ll probably say it again, but I think I’ve finally overcome the last ‘big’ problem with the Magnetic Cello. I still need to replace the 9v battery power supply with a AA supply, but, unless the new amplifier finds a way to act up, the circuit is finally almost complete.
The instrument has gone pretty far since my first sketch three years ago:
(Full disclosure: I can’t draw this well. I traced over a cello that was in a Macy’s advertisement.)