[time-nuts] Programmable clock for BFO use....noise

[ed breya]

Those programmable oscillators look interesting. I went to Cardinal the 
website to learn more, but they're pretty sparse on details. It looks 
like they make all sorts of crystals, OC, TC, and VT XO modules, etc, 
and these programmable ones, which are apparently PLL-based oscillators 
locked to an XO. They don't say much more, but if you asked, maybe 
they'd give some usable info.

There was also some mention of their own shortcomings with phase noise 
and jitter, circa 2005, and how the later generations are much improved.

I'd say that since the programmed frequency is at least XO-based, then 
its stability seems to be specified and can be good, but the noise etc 
of the associated PLL inside does not seem to specified. It may be OK 
for the application to just pick a certain frequency that would be 
easier on the PLL (if you knew more about it and the XO frequency they 
would use), and easy to divide down to your end results. I would wonder 
what the few-off cost would be for the appropriate base parts and 
programming.

They also seem to have custom crystal building service, which could be 
very handy, depending on the cost. I've often needed oddball frequencies 
for various projects, and wished it was easy and cheap to just order 
some up.

Another trick you can try, that I've resorted to a number of times, is 
to find two "standard" or common (or oddball ones that you happen to 
already have) crystal frequencies that you can mix to get the desired 
result. The combinations of various crystals and possible dividing 
ratios may yield something close enough to tweak in. An EXCEL sheet can 
help organize the info and choices.

Finally, of course, you can use DDS. This is nearly an ideal case for 
this, since you want to make stuff around 250 kHz, but necessarily must 
(for lack of in-range XOs) use a clock in the MHz region - maybe ten to 
a hundred times higher than the output, so easy to get a good sine out. 
Between the various XO clock frequencies available, and the program 
choices in the DDS, it should be possible to come up with a nice scheme 
to make whatever you want down in that range.

Ed