Okay, so this picture kinda sucks. Didn't light it terribly well.
This is not the completed hookup, but shows the 12-Volt coil, power and keying-transistor detail before the rest of the wires are installed and block a clear view of the 12-Volt mod details.
The electrolytic caps in this pic are 1000uf 25 Volt, but we have found that 470 uf works just fine. They're smaller and take up less space.
The axial cap at the upper right has 6.3 Volts AC feeding into its negative lead. That wire is not yet installed as you see it here.
The shunt diode has its cathode (banded end) connected to the positive lead of the axial cap, the other end is grounded.
The series diode is strung from the positive end of the axial cap to the right-hand 'hot' coil lug of the relay socket. The radial filter cap has its positive lead on this lug of the socket and the negative lead grounded.
A third diode is connected in parallel with the relay coil to suppress the reverse-polarity "inductive spike" pulse that will come shooting out of the coil when you unkey. This could be unhealthy for the keying transistor if it were not suppressed with this diode.
The PN2907A keying transistor has its emitter lead on the left-hand 'cold' relay-coil lug. The collector is grounded. The center base lead has a 220-ohm 1/4-Watt resistor between the transistor lead and the yellow wire from the mike socket. We have since found that you can leave this out and connect the transistor's base directly to the yellow wire.
One detail that's not terrribly clear is a pair of .001uf disc capacitors. Any voltage rating 25 Volts or higher should be fine. Each goes to one of the sockets coil pins, the other side to the ground lug between the two coil lugs. This serves to prevent RF from leaking from the relay's antenna circuit into the coil circuit. Might not be necessary, but it's just conservative RF-design practice.
But that's the gist of it.
73