I've for a long time been interested in timekeeping devices. On a
business trip to Germany back in 2001, I came across so called
'atomic' watches. The European version (DCF77 signal) of these
timepieces include a receiver tuned to 77.5 kHz, the DCF77 timecode
carrier frequency. Although pretty interesting, I'd decided it
wouldn't make much sense to acquire one of these since the DCF77
signal reach is limited to the European continent. I would opt for a
solar powered 'regular' timepiece instead. Fast forward to Sept '09,
on visit to Berlin, I came across a watch I just had to have. Despite
being an 'atomic' type, I acquired it and thought I'd figure some way
to replicate the DCF77 back home in Montreal. So began my search for a
solution.

There are in fact a few readymade DCF simulator solutions out there.
These short range, low power devices are meant for use in trade shows,
shops and the like and may be used to relay any number of time
sources. Although I couldn't find North American suppliers, they can
be purchased directly from European sources. They devices are pricey
however.

In essence, there are two problems to solve. Time code simulation and
signal transmission. After some searching, I hooked up with Steve from
Novatech Instruments, Inc. His background combined with H/W & S/W
integration expertise were the right ingredients. But what really made
the day on this project was his motivation and genuine friendliness.

That being said, the technical issues associated with the development
of a low cost simulation solution were numerous, as were possible
development avenues. The end product offered here is a correct blend
which achieves both functional and cost criteria.

The circuit is basically a low power transmitter with a carrier
frequency matching that of the actual DCF77 transmission in Europe.
Timestamps (time + date) are picked from the PC's internal clock. Goes
without saying that the clock may be synchronized from any number of
sources. It's then coded per the DCF77 format and sent over the
carrier.

So two things. Get the transmitter online (flick the pwr switch) and
start up the DCF77 Code Gen. The procedure consists of running the
provided dfc1_sds executable. Inputs to the program include a 'fudge
factor'. It is in fact a pulse timing correction, as most PC's won't
run the transmitted timecode at the required accuracy. Nothing to do
with actual time signal, but rather keeping the transmitted pulses
phased to what the receiver CPU (the timepiece) is expecting. This
correction (or fudge) factor is arrived at experimentally and may be
different from one PC to the next (~ 98.5% and closer to 100 with
'older' PCs). Trial & error using the timepiece + DCF simulation kit,
this effort may be wrapped up within a half hour.

I've recently moved the breadboard circuit to PC board. The end result
is the ability to remotely synchronize a DCF77 capable 'atomic'
timepiece, even when far away from the Old World.

Many Thanks to Steve and his handiwork.

Pierre
Montreal, Canada