I think the design is now stable enough to provide a schematic:
As you can see, there's not much to it! A 74AC86 exclusive-OR gate serves as the oscillator, changes the carrier phase under control of the microprocessor, and provides high output current to the tank circuit. (I may change to the 74HC86 to reduce high-frequency emissions, but that will require increasing R4 and R5 to reduce the output current, giving less field strength. Right now the field strength is just about the same as my AMB and MRT transponders.)
One interesting thing that I mentioned previously about the design is that the tank is driven differentially (bridge drive), which substantially reduces the second harmonic level in the output spectrum. This could help with meeting emissions requirements.
The values of CT1 and CT2 might have to be changed to match whatever tank inductor you use. My original perfboard prototype used 16 turns of 30 gauge Kynar (wire-wrap) wire, while the PC board has 20 turns of traces (10 on each side). The ratio of the two capacitors will set the tank Q (CT1 larger gives higher Q), while the series combination of CT1 and CT2 sets the resonant frequency.
R1 was originally a limiting resistor for the shunt regulator inside the intended microprocessor, the PIC12HV609. It is shown as a diode because I have been using PIC12F508 microprocessors left over from another project. The diode drops the input voltage by about 0.6V to keep the 12F508 happy (it has a maximum operating voltage of 5.5V), and as a bonus protects the transponder from reverse polarity. Using the 12F508 (or almost any 12F series device other than the 12HV609) means the supply voltage to the transponder is restricted to 6.1V maximum. That's okay for virtually all BEC circuits, but is not compatible with a 5-cell NiMH or 2-cell LiFe battery. I stopped using receiver batteries a long time ago! Adding a separate Zener diode would let the transponder operate from the same large voltage range as the PIC12HV609 permits.
I am considering changing to the PIC12F683 for its larger memory and EEPROM.
The 47uF cap is probably not necessary for use with ESCs having linear BECs. If the BEC is a switcher (like a voltage booster for 1s LiPo cells) then it helps filter out voltage ripple. But further testing might reveal that it is extraneous.
J2 is a jumper block to select the transponder ID number. Currently the code supports two ID selections.