Jeff Liebermann wrote:
I live in a fringe reception area for the german time
signal at 77.5kHz
Spain? According the coverage map at:
http://www.ptb.de/cms/en/ptb/fachabteilungen/abt4/fb-44/ag-442/dissemination-of-legal-time/dcf77/reach-of-dcf77.html
Yes, the signal is probably not very strong and you're in the skywave
only region.
Good observer, yes, Spain. Also the Pyrenees are in the way blocking most ground signal.
and have many clocks that use this signal but because
a marginal reception they do not always sync.
That sounds like the older amplitude modulated system. DCF77 also
transmits a phase modulated signal,
https://en.wikipedia.org/wiki/DCF77#Phase_modulation
I knew about the phase modulated encoding on the signal but I am not sure if any of my clocks use that. The older ones sure do not, time ago I dissassembled the oldest alarm clock and scoped the signal output from its radio board and I could see the 0.1/0.2 sec signal drop.
There is a long wave radio transmitter in France at 162 kHz that carries time information by phase modulation of its carrier:
https://en.wikipedia.org/wiki/Alloui...ve_transmitter
I find it an efficient way to use existing resources, also you can check for reception by listening with a LW radio.
The curious thing about this is every night I have a different set
of clocks sync, they seem to decide at random when the signal was
good enough.
Kinda sounds like you're getting some local interference. It doesn't
take much computah or switcher noise to trash the signal, especially
when it's weak in the first place. If your devices have a loopstick
for an antenna, they are directional, with the strongest signal
perpendicular to the loopstick. I was having random updates until I
moved my WWVB clock away from several switching power supplies.
Yet, if I had one clock next to a noise source I would expect it to sync less times than the rest. But no, there appears to be a random distribution.
The oldest one is 13 years old and is an alarm clock I check daily,
in these years I have had it three times take an incorrect time/date
probably from interference. It either takes the right time or does
not sync, three times in 13 years seems not bad for the simple parity
bit protection the protocol uses.
It's a bit more than just the parity bit. Some chips require that the
correct time be received successfully more than once before it will
sync. This is a function of the chip design.
That's what I belived, but then how did it receive incorrect data? Did it actually receive the same wrong data twice in a row? The algorithm could also check for a reasonable deviation against the current setting.
I think, however, it is somewhat dangerous to have the clocks sync
at night *after* you have checked they are set correctly
Sorry, but you don't have a choice as to what time to sync.
Maybe a setting to enable sync only during weekends or only when the alarm function is disabled.