Help with Data of fiberoptic

[Ryan7777]

Hey guys and gals,

I need to trick some circuit cards for a large and expensive piece of specialty test equipment that uses fiber optics to communicate (do to the fact the cards are floating at high voltage). the data is in the form of 24 bits or pulses, that vary between a narrow pulse width and a wide pulse width.
i.e. only two states, wide or narrow that stand for 1 or 0. so its like a 24 bit serial packet, but instead of high-low, its wide or narrow (no, its not pulse width modulation, in the sense that it only has two states).
the narrow bits are 260 nS (yes, nano) wide and the wide bits are 1 uS wide.
the spacing between bits is 1.30 uS between narrow bits, and 520nS between wide bits and any other bit (wide or narrow). the total width of a given packet is around 35uS which doesnt change. I'd like to be able to vary the 24 bit number with say a pot connected to an ADC input. seeing how were down into the nano second range here it doesnt seem like it can be done in PBP. but you guys tend to be pretty crafty. I'm not real how forgiving these cards are timing wise, they contain some standard logic to process the data and feed it to one ADC and one DAC. no processor on their end, the control head has an 8051 family chip in it i think. I'm open to any idea's. I've already tried a pattern generator with no luck, as it can only do limited NRZ - RZ patterns..

[Bruce]

It should be do-able with a 15.360MHz crystal, but you would need some
assembler mixed in with your PBP code for critical timing sections. And PBP
timing would be a tad off (~4%) since it only directly supports 16MHz.