I've been working on the design for a frequency divider to complement the Thunderbolt I recently bought from TVB (thank you Tom, it's working very well as far as I can tell, though of course I've no other standard to compare against).
Thanks to lots of advice and guidance from Bruce Griffiths (many thanks again Bruce), I've got the design near completion. I'm not aiming for NIST or equivalent perfection in terms on minimising jitter and other noise, but would like to at least make a at least a half-way decent job of this. I'm now thinking ahead to the PCB requirements,with the caveat that I've only ever designed one PCB before and that was a single layer board done using double sized mylar and sticky black tape (Yes, it was a good many years ago). Now to questions: 1. Surface mount or through hole? I don't have a re-flow oven (or even a hot air soldering system), so my inclination is to use through hole CMOS (74HC163s with 74AC glue logic and flip-flops), with the surface mount restricted to the clock shaper using a BAV99 and either an ADCMP600 or MAX999 and surrounding components. Will using through hole cause me grief? 2. How many layers? In an ideal world with money no object, if I understand the current art correctly, I think I'd probably aim for a five layer board with Vcc, Digital Ground and Power Ground being separate internal planes, and trace routing on the top and bottom of the board with as few vias between top and bottom as possible. Does that sound right? Do you think I can safely restrict myself to two layers, and if so does it make most sense to make one side of the board digital ground, and route everything else (Vcc, Power/Analogue Ground, and signals) on the other side. Or is there a better approach (always assuming that a two layer board is a viable option). Cheers Dave Partridge _______________________________________________ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.