Let's begin with the end result:
* The GT Theremin (http://mattb.alwayssleeping.com/images/theremin/theremin_32.jpg)
* Close up (http://mattb.alwayssleeping.com/images/theremin/theremin_30.jpg)
* Closer up (http://mattb.alwayssleeping.com/images/theremin/theremin_31.jpg)
* Inside (http://mattb.alwayssleeping.com/images/theremin/theremin_33.jpg)
* Filament transformer, tuning caps, internal supports for antennas, etc. (http://mattb.alwayssleeping.com/images/theremin/theremin_34.jpg)
* Detail on sliding contact for the removable antenna (http://mattb.alwayssleeping.com/images/theremin/theremin_35.jpg)
* Detail on one of the two stacks of circuit boards separated by a copper ground plane (http://mattb.alwayssleeping.com/images/theremin/theremin_17.jpg)
First of all, yes the blue glow comes from LEDs... Blackbody radiators have to be pretty darn hot to emit in the 470 nm band. :P The physical design of this Theremin was intended to create an eye-catching aesthetic to go along with such an interesting instrument. Thus the metal look, the LEDs, and the somewhat unnecessary use of a glow-discharge tube (the big tube glowing neon orange).
I wanted to build it with vacuum tubes because I had never worked with them before, they add to the aesthetic (sorry, but an op-amp design in a plastic box isn't very interesting to look at!), and I think that doing so somehow pays homage to Termen's original design and the instrument's popularity in sci-fi movie sound tracks. The design uses only vacuum tubes as active devices, though I did resort to using a single semiconductor diode in the power supply to avoid me having to build a (complicated) error-feedback voltage regulator late in the design stages. Obviously the construction methods are modern, with molded (rather than wound) inductors in the RF oscillators, protoboard and heat-shrink-wrapped header pins, lovely compact electrolytic decoupling and filtering capacitors, etc. Again, I wasn't necessarily aiming to create an authentic early-twentieth-century design.
Here is the full circuit schematic (http://mattb.alwayssleeping.com/images/theremin/theremin_full_schematic.pdf) if you're interested. It's entirely original, though I was influenced by Doug Forbes' and Art Harrison's tube theremin designs (kudos to them). I prototyped all the circuitry on several breadboards (http://mattb.alwayssleeping.com/images/theremin/theremin_10.jpg), but RF coupling issues were fairly severe with that setup so it still required some tweaking once it was fully assembled and in its present form.
The Theremin as shown was assembled over the past two and a half weeks or so, with the physical design being mostly dreamed up as we went, laid out in CAD in the evenings, and built during the day. The circuit design has taken me the past couple of months to conceive, simulate, and test. Some crucial components like the oscillators were simulated in EWB Multisim (http://mattb.alwayssleeping.com/images/theremin/12au7_colpitts_cp_v3.png) with some custom tube models that I found online and tweaked for my purposes. My friend is responsible for a lot of the good construction ideas, including the stacked circuit boards and the arrangement of the top.
So how does it play?
Pretty well! The pitch range is roughly 100 Hz - 3.4 kHz over a sense distance of around 12" (just guesstimating on the distance; I haven't measured in awhile). Volume dynamic range is about 28 dB, which is sufficient to bring the output levels from 1.5 Vpp max down to a few millivolts over the noise floor. There is a lot of harmonic content at the lowest frequencies, and the highest frequencies are obviously too closely-spaced to be very useful. I'd say the useful playable range is about