Archive of the former Yahoo!Groups mailing list: MOTM
Subject: One More Stupid Idea...
From: mbedtom@...
Date: 2002-06-18
Forgot to jot down one more idea before I hit the "send" button...
Linear Controller: Build a wood "box" to house a 30" long slot in which a machined aluminum "U" channel is fitted. At each end of the "U" channel is a nylon pulley, mounted perpendicular to the "U" channel. A chain or steel wire attaches to each end of a shuttle that rides in the "U" channel slot. The shuttle is fitted with a pointer to indicate pitch and a microswitch-based gate/trigger switch. Inside the box, this chain or wire is routed around an idler arm & pulley at one end, and a Gurley encoder at the other. The encoder is connected to a small PC board that has a PIC processor and an accurate D/A convertor. As the shuttle is moved back and forth in the "U" channel, the encoder shaft is rotated in either direction and provides positional information to the PIC processor. The PIC in turn, converts the positional information to a control voltage via D/A convertor. The PIC also generates gate/trigger signals in response to the microswitch installed on the business-end of the shuttle.
The Gurley encoder is the heart of the system. It provides up to 1800 pulses per revolution of the shaft. That translates into a very fine resolution for position sensing. The Gurley encoder also has quadrature outputs which can be used to interpolate position finer, by a factor of four. If one assumes that the gearing in the shuttle and encoder causes 10 shaft revolutions of the encoder for a full transit of the shuttle assembly, one could resolve 1 part in 72,000 of positional information. That is better than 16-bits of resolution - far better than anything one would need for a linear controller application.
There are many other brands of optical encoders out there, but Gurley is on the high end of resolution and accuracy. They are pricey though. But they are unaffected by dust as they are sealed quite well and are not particularly temperature sensitive. That combination should provide a linear controller with an accuracy better than the rest of the synth and a lifetime of use. Personally, the idea of a "ribbon" based controller doesn't seem very reliable to me. I cannot imagine that a resistance element would last very long or be very stable when time, dust, temperature, and humidity are factored into the equation. Even the choice of wood on which the resistance element is affixed, would affect long-term performance.
The encoder of which I speak, looks like a small motor that might be found in a radio-controlled model car. Inside, there is a precision circular disc made of glass that is fixed to the encoder shaft. On that disc, microscopic "spokes" have been deposited through a photographic process. These "spokes" interrupt a light beam to indicate a small amount of rotation. Each time the light beam is interrupted, the encoder puts out a logic-level "blip". There are actually two or more light sources and "beam interruption detectors" which are strategically located inside. They are mechanically offset slightly so that the signal from one "beam interruption detector" is skewed from the other "beam interruption detector". That generates two signals that are out of phase with respect to each other (by 90 degrees). With the proper electronics, it is easy to determine which way the shaft is turning and by how much. With the use of some software trickery, one can multiply the resolution as long as the shaft is moving.
I think that a linear controller that is based on an optical encoder is the way to go. Calibration would be a snap. The shuttle could be made with a very low mass - that would be important to provide crisp "action". For smoothness, I'd put a Teflon "foot" on the shuttle. That would make the action silky smooth and last a very long time. Lastly, I'd put a knob or "handle" on the shuttle, fitted with a ball-bearing race. That way, the knob or handle would be held firmly by the performer, but allow rotation as the angle between the "U" channel and the performers hand changes. No carpal-tunnel syndrome for me, thank you.
Cheers!
Tom Farrand