The AT90S2313 has two pins which can be used as the inputs of an analogue comparator.
I propose to create an LC oscillator by placing a parallel resonant tuned circuit across these pins and by supplying energy via a resistor from an output pin (which would normally be tristated) when the comparator switches state.
I am thinking of a 20kHz frequency with a capacitor between 10nF and 1 micro Farad with a 3V supply.
Do I need to protect the inputs from the stored energy in the tuned circuit when power is removed ? Would it be good (or bad) practice to include a resistor between the tuned circuit and a comparator input. The configuration I suggest leaves the comparator inputs floating - is this a bad idea and if so what should I do about it ?
I would like to achieve a stable frequency - as I would like to detect changes in the frequency when metal comes into the vicinity of the coil. Should I be employing a different oscillator configuration - for example a "centre tapped" capacitor like in a colpitts oscillator.
On a more theoretical note, presumably such an oscillator will generate electromagnetic radiation - which I suppose one could model by a resistor in series or across the tuned circuit. Is this a misunderstanding on my part - or is it so negligable that that is why one doesn't see it discussed in models of a tuned circuit.
Many thanks for your comments.