Micropowered Theremin?

Many of the circuits I've been messing with lately draw very little current, which makes me wonder how low one might possibly go in constructing an otherwise viable yet low- or micro-powered Theremin?  Hams have these sorts of projects / contests, which can be both instructive and fun.  And a micropowered Theremin could be easily powered by a battery, which raises the possibility of eliminating a serious noise source (wall warts are now most often switched power supplies) as well as increasing portability.

Here is a candidate ~200kHz oscillator that draws less than 1/2 mA with ~60Vp-p at the antenna (in LTSpice):

I need to bench it, but similar circuits using CMOS have worked well in the past. 

The dual JFET VCA circuit over on another thread here is similarly quite low powered:

I'm thinking a supply voltage of 3.3V (+/-1.65) is a good target (maybe too low?).

No plans!  Don't know where this is going, if indeed anywhere, but it's fun to think about.

[EDIT]

The goal with low powered oscillators is that they don't poop out or stall when the antenna is grasped.  This can be realized with sufficient drive current / reduction in antenna voltage swing, and an insulating sleeve on the antenna doesn't hurt either.

One reason to go with opamps in a micro-powered analog Theremin is that the specifications are much more well defined (tighter) than for discrete FETs and such.  Opamps also give push-pull drive, where a single FET can only pull one way.  CMOS is also an option, but when linearly biased buffers and gates (particularly the higher speed families) can draw more current than an opamp while essentially doing the same job.

A lot of it obviously comes down to the specific circuit.  There would likely be very little difference in current usage between an opamp oscillator and a single FET Clapp or Colpitts.

LOL ;-)  FredM cant log in, but I can - and we share the same password! (Wonder how I know that? must be telepathy! ;-)

Dewster, I will probably repeat some of the things discussed in our private emails..

I completely agree that a good op-amp like the LM6132 beats CMOS hands down for linear operations - and as you say, "when linearly biased buffers and gates (particularly the higher speed families) can draw more current than an opamp while essentially doing the same job".. In terms of remit of this threads topic - "Micropower" - Opamps are a good choice.

In terms of cost / performance though, CMOS looks to win hands down.. Unbuffered CMOS clocks in at about 8p per inverter (each inverter can be configured as an inverting amplifier with the addition of 2 resistors, each amplifier can have a useful gain of at least 5) wheras the LM6132 clocks in at about £3 (£1.50 per amplifier).

When it comes to quality, the cost of a few LM6132's isnt significant - But I have been getting better performance for things like oscillators and mixers from unbuffered 4000 and 74HC series CMOS than I can get from opamps, even the 6142 which is twice the price of the 6132 (The LM6142 is the closest to an ideal op-amp I have found).

Oh, there are many linear (analogue) things you cannot do with CMOS - For any application requiring accurate differential inputs or rail-rail IO  CMOS is quite useless, and the self-biasing configuration limits CMOS amplifiers to AC coupled applications mostly.. But for theremin speed (<2MHz <10MHz BW) AC amplifier and filter / mixer applications it seems to do everything required.

But whilst CMOS is not excessively power hungry (if care is taken), opamps win for micropower applications, and beat simple discrete semiconductor implementations as well.

Fred.