[time-nuts] WWVB Chronverter update progress
eb at telight.com
Tue Aug 28 19:11:30 EDT 2018
If you're going to reference it from a GPSDO anyway, why worry about a
TCXO reference (and power too, for that matter). You can easily make the
60 kHz from the 10 MHz.
For example, with two 74HC390s and a 74HC86 you can make 50 kHz and 10
kHz and mix them with one EXOR section of the '86 to have 60 kHz
available. Some fairly simple bandpass filtering should select and clean
it up sufficiently. Two other sections of the '86 can be set up as
inverters and self-biased as amplifiers - one to convert the 10 MHz sine
reference to logic, and the other from the filter output to logic, if
needed. And, you'd still have a divide by 5 and an EXOR left over to
fool around with.
The same parts and process can be done at a higher frequency and then
divided down afterward. (5+1) MHz/100, and (500+100) kHz/10 would work
too. It depends on what frequency you prefer for the BPF. If you go high
at 6 MHz, you then have the option to make a crystal filter from readily
Going the high way also provides for higher logic frequencies that are
more or less in sync, in case you want to do any I-Q modulation type
stuff - you can even use synchronous counters instead, to really make sure.
Overall, I think I'd recommend going at 5+1= 6 MHz, filtering with 6 MHz
crystals, then dividing down to 60 kHz, with 2f and 4f clocks available
for I-Q use. I sketched out a quickie circuit that's quite simple and I
think would do. It would take two HC390s for the dividing, as before.
Each HC390 is two divide by 10 counters, including a 1/2 and 1/5 in
each, usable separately. So, with two parts, there are four 1/5s and
four 1/2s available.
Here's a verbal process description: 10 MHz sine convert to logic with
HC86, 10 MHz/2=5, 10 MHz/10=1 with first HC390, add 5+1=6 with EXOR, BPF
6 MHz, convert to logic with HC86, 6 Mhz/25=240 kHz=4f, 240/2=120
kHz=2f, 120/2=60 kHz=f, with second HC390.
More information about the time-nuts