[time-nuts] Basic regenerative-divider questions

[John Miles]

> What's the crystal for?
> Crystal filters aren't usually necessary

The nature of the filter(s) is one of the questions that I'm hoping those
papers will help answer.  I was guessing that a crystal filter would make
the divider harder to start -- i.e., would require more gain from the
amplifier -- but would be good for attenuating any remaining noise at
offsets beyond the filter bandwidth.  Sort of like the 40 and 160 MHz
monolithic crystal filters in the 8662A's reference section (which is what
the circuit I'm working on will eventually replace).

> A divide by 8 conjugate regenerative divider has been built and tested
> by NIST.
>
> In principle it would be possible to generate F/2, F/4 and F/8 outputs
> simultaneously by adding parallel conjugate filtered feedback paths
> tuned to 3F/2, 7F/8, 3F/4, F/2, F/4 and F/8.
> However the difficulties associated with optimising the phase shifts and
> gains of all the filtered feedback paths may be more trouble than
> its worth.

Yes, I imagine I'll take the lazy way out and just run separate dividers in
parallel from a 2- or 3-way splitter following the OCXO.  Thanks for the
uploads and links.  Got some reading to do this weekend.

I rigged up a divider last night with a 220-MHz SAW filter and a
randomly-chosen Mini-Circuits mixer, MMIC amp, and 2:1 splitter.  I was
surprised at how well it worked without any tweaking.  It was kind of
surreal to see the f/2 output appear abruptly once the 440-MHz input reached
a certain level.  At +4 dBm of excitation there was nothing at the output,
but with +5 dBm at the input, a nice clean 220 MHz signal appeared out of
nowhere at +16 dBm.

-- john, KE5FX