Yesterday's video showed a very fussy version of Tetsuo Kogawa's 1 transistor FM transmitter, which worked after a fashion, but which seemed really squirrely. Almost any motion of anything caused the circuit to behave rather badly as capacitance changed, and I picked up a considerable amount of hum. Today, I rebuilt the circuit onto a piece of one sided copper clad PCB material, and it worked much better. Hardly any hum, and much less finicky. I didn't even try to add a clip lead: what you see below is the circuit just operating with whatever signal radiates from the PCB.
I'm still getting multiple copies of the output across the FM broadcast dial, so I am not sure that it's really that great of a circuit, and I'd be terrified of trying to amplify this and send it over a greater distance lest the FCC come hunting me down, but it at least works, and wasn't very hard to debug, once I got the difference in pin layout for the 2N3904 sorted out and redid the layout a bit.
Check it out!