The components were mounted on foamboard and wired on a perfboard suspended in parallel by the leads of the components. Pieces of foam were inserted between the foam faceplate and the perfbord to make the instrument more structurally solid. The perfbaord circuit was implemented using lots of thin jumper wire, and a ribbon cable was used to bring points to my breadboard on which the arduino was mounted. Some additional circuitry was left on the breadboard (for the encoders).
The inspiration for the layout and combination of sensors was a basic granular synthesis algorithm. As it turned out, the algorithm, which originally operated on sample files, was altered to synthesize sound particles. This made it much easier to debug the circuit, and turned out to be a very nice sound source to run though some other granulation software I had previously written.
Early on, I ran into one major difficulty that took up most of the development time. After everything had been mounted and wired up, I was having trouble getting data from the FSRs. I tried to trace back the problem through different parts of the circuit leading the FSRs. My critical error was in assuming that since the problem exhibited itself in the same way for each FSR, that the bug must have been in my circuit. In fact, the problem was caused by different issues for each one. One of the FSRs was being compressed by the foam I had inserted between the faceplate and the circuit board, and was thus acting like an open circuit. The other FSR was simply faulty, which I finally verified after ripping it from the 'finished' construction. Replacing the component was delicate and tedious. In fact, I broke off the leads of my replacement part, and had to solder wires directly to it! I have worked with circuits before, but this served as a reminder of the importance of checking and double-checking circuits as you go, no matter how simple the project at hand.