Hi,RFDave,thank you for your reply.
I have read Perrott's thesis carefully, I found the multi-mod divider was designed skillfully. It was so powerful. So I try to implement the circuit. But when I construct the the first stage---div-by 2/2.5/3/3.5 ,I found the state machine(include two serial connected sb-pc-latch) can not reach the right state, The divider can not implement div-by 3 and 3.5. Is this true? Or it is my mistake?
I'm sorry, I was primarly interested in the system level stuff (Pre-emphasis, modulation, etc) rather than the circuit level details. I would say that the block diagram's in there may not be enough to recreate his design. If you google for Perrott, he is now a professor @ MIT, and has some more papers on his web page, that may explain it better. I suspect that there are better references in building a multi-modulous divider though.