Here is the block diagram for the long wave IR theremin design. A full size expanded version is available by clicking the image.
IR Sensor Discussion
The IR sensors are an array of thin film Bismuth-Antimony thermocouples surrounded by Xenon gas. Active junctions in the center of the detector receive radiant energy through a thin film filter which eliminates wavelengths outside of a 8-14 micron area. Non-active junctions are shielded by the TO-5 can and have reasonably thick bond wires leading to the sensor leads in order to provide some level of heat sinking. The Xenon gas is a bit of a thermal compromise. It was chosen to provide slow albeit decent enough response times for music purposes, but more so to provide greater stability, and less of a headache for dT/dt (temperature/time) events.
High Gain Amplifier
I normally like chopper amps, but since this is a music app, and I dont know if I have any in the junk box I’ll likely go with something more conventional like an OP77. The IR sensor signal is very small, on the order of a few microvolts, so low drift op amps are a must…but obviously noise is a concern as well. The Vos drift spec of 0.3uV/C is a real concern… but if the input amp is kept separate and close to the isothermal block assembly, most of this concern will be alleviated.
The offset/gain block
At this point, the signal is high enough, I can do some serious op amp modifications to it without getting into too much trouble. The biggest issue here is not to run into a wall during processing. The 555 pulse position modulator needs an input range of roughly 0.1-0.9VDC for full scale performance.
The 555 Pulse Position Modulator
The nice thing about the PPM is that it is very non-linear, as will be the output of the long wave IR sensor, add in the issue of even tempered musical scales, and life will be interesting. Eyeball wise, this should work out ok… being Spice is worthless in this topology, I wont know for sure until its up and running and I can make some measurements.
555 PWM and the Voltage Controlled Amplifier
Granted, such is a bit kludgy. However, the intent is to use parts I have sitting around… and the PGA’s I have dont exactly lend themselves to the output provided via a duplication of the above gain/offset design as an input stage. Then again, if this was a production design, I’d just bring the high gain amp output into a ADC and let a processor or even an FPGA work out the non-linearities, as well as creating the output waveform on the fly.