[ntp:questions] Re: AW: Re: AW: convert_rawdcf: parity check FAILED on my embedded Linux system

Hans Jørgen Jakobsen hjj at wheel.dk
Sun Oct 15 12:12:24 UTC 2006

On Sun, 15 Oct 2006 11:28:09 GMT, Torsten Krieger wrote:
> Hi Uwe,
> I made a small c program which reads the serial port /dev/refclock-0
> character by character, ntpd switched off. Shall I see one character by
> second (second = 1/60 minute) 0xf0 for low bit and 0x00 for high bit? If
> this should be the case then something went wrong with my serial interface.
> Ciao,
> Torsten
> Uwe wrote:
>>The DCF Signals is 1 pulse every 1 second. the 59th pulse is omitted to
>>mark the minute transition.
>>AAdditonal information is coded into the pulselength of the second pulses.
>>0.1 sec for a low bit, 0.2sec for a high bit.
>>on a 50Baud (0.02s/bit) serial line with inverted sense this is received as
>>HHHHLLLLLHHHHHHHH	S=StartBit, s=StopBit(s)
>>____S01234567s____  0xf0 for a low bit
>>____S01234567s____  0x00 for a high bit
I dont think its that simple in practice.
The DCF receivers I have used have some large jitter on both start and
length of pulse. Add to that noise impulses.
Try and print the bits and figure out what signal you get.

More information about the questions mailing list