[time-nuts] WWVB d-psk-r update. Seems a reasonable solution

paul swed paulswedb at gmail.com
Thu Mar 28 21:01:27 EDT 2013


OK it has been a while and I have promised to share results. In the next
few days I will put whats on paper into a schematic and share. I won't
dwell on all of the stuff from October to now. But several front end rcvrs,
and several analog Costas loops were built using MC1496s and another using
AD633 multipliers. Various other glue.
Nothing to rave about.

What I will detail is what I have done with quite common parts. But what it
isn't is a kit or absolute or anything else. It can be made better. Don't
doubt that for a second.
So what I do have are 4 pretty simple and reasonable parts of the solution.
It really does work even on the east coast. Though MSF can really cause
issues ocasionally. One key part the costas loop is a design I stumbled
across from 1971 designers casebook electronics magazine ca herbst. Was
cleaning out old papers. The article left a lot of detail out. But the
principle is solid. It was surprising at how well it worked the first time.
So what I will share with the Time-nuts are the following schematics.

A 60 KHz TRF receiver. Takes a 1uv signal and increases it to 2-4 V PP with
AGC and soft limiter. It includes a 60 Khz xtal filter to set the passband
and a tuned stage or 2. There are quite a few possibilities on the
internet. Tried a few. This is my version. But others would work just as
well.

Digital costas loop using nothing more the a 74hc86 quad exclusive or.
Resistors and caps to integrate.

The 12 Mhz VCO. Why 12? Thats the crystals I have. Boy can this area use
some help.

The divider. Cmos 12 Mhz to 0 and 90 degrees at 60 Khz. This is straight
out of any digital application manual. No magic at all.

When combined this creates a complete WWVB costas loop solution that
delivers a 60 Khz output locked to wwvb without the phase modulation. A
byproduct is the digital data from the new signal. Not used. Not even to
drive a led.

A path could be added that literally inverts or not the original AM signal
so that spectracom 8170s and Truetime DC60s would work and recover time
correctly. That is a piece O cake.

So I will get this into a schematic to share over the next day or two.
By the way on performance the system handles sun rise and sets and keeps
ticking. Using a fluke 207 driven by a Rb ref to watch for failures. This
is recorded on a 4 channel chart recorder program on a laptop. I have used
the system to drive 2 X HP VLF 117s and they track as w
Lots of areas that might be improved AGC control and time constants. Adding
a opamp to the 0 degree phase VCO control voltage and also adjusting Time
constants. Though the 12 Mhz oscillator is really standard parts just could
not find something reasonable and off the shelf. Digikey or Mouser stuff.
something with 4 pins and a few $. By the way the control voltage is
.5V/.1Hz.

Regards
Paul
WB8TSL


More information about the time-nuts mailing list