🚀 go-pugleaf

I am building a class-C amplifier for 2275 hz (yes - audio frequency),
as shown in the schematic below:

                                        V+
                                        |
                                    +---+---+
                                    3       |
                                    3       |
                                    3       |
                         +----------3       -
    +-+                  |          3       _
    | |                |-+          3       |
    | |  Signal In ----|<+          3       |
    | |                |-+          3       |
 ---+ +-----------       |          +-------+                     
                       -----
                        ---
                         -



The application is a transmitter for a locator beacon, where the
magnetic field of the inductor is picked up by the reciever's tuned
citcuit.  This allows a micro-power transmitter to operate over a
distance of 50 meters for about 1000 hours using a nine volt alkaline
battery.  I want to mazimize the magnetic field generated, and
minimize power input in a highly reliable circuit, while keeping size
and weight down (it should fit in a large shirt pocket).

Before I get to my real question, I will describe my circuit in case
anyone has any better suggestions.  It starts with a crystal fet
oscillator, which is divided down to 2275 hz with a cmos divider.
This signal is shaped into a low duty cycle pulse (with adjustable on
time for tuning purposes) by a couple more cmos gates, a resistor and
capacitor, and the mosfet output transistor is switched in the class-C
configuration shown.

Now my question is how do I optimize the inductor in the output tuned
circuit.  Clearly, I want the largest inductance possible while
keeping resistance reasonably low.  So far my best coil is 32 gauge
wire wrapped on a ferrite rod (from an AM antenna).  This has
acceptable size and weight, along with reasonable values for
inductance (50mH) and resistance (10 Ohms).  I am doing something
right, since the LC circuit develops several times the power supply
voltage, and the signal can be detected at a moderate distance given
reasonable low power input.

I am convinced that a tapped inductor is necessary, since otherwise,
the signal clips when it reaches ground (even with a bipolar
transistor).  I have tried various configurations, and it appears that
tapping the center of the inductor yields the cleanest signal across
the circuit.  It there a reason for this?  It this necessary for
optimum efficiency?  Theoretically, what is going on in this circuit?

Also, is an AM radio ferrite core acceptable here?  I can cook my own
core out of low freq ferrite powder, but this is a pain and I would
prefer to avoid doing so unless really necessary.

Finally, the coil produces an audible buzz.  I assume this is due to
vibrating wires and will go away when I pot the final version in
epoxy.

Thanks for any information.

Ken