[ntp:questions] Re: Zone Problem?

Martin Burnicki martin.burnicki at meinberg.de
Wed Nov 5 15:08:32 UTC 2003


Daniel Dotta wrote:

> Hi,
>     I instaled a GPS reciever in a serial port at my Linux PC. The service
> is running ok. My clock is syncronized with the GPS.  My only problem is
> with the hour diference between the GPS and PC.
> I am using UTC time in GPS and Linux (no time zones), but I am still
> having a diference. See this:
> ntpq> clocklist
> status=0000 clk_okay, last_clk_okay,
> device="Meinberg DCF77 C51 or compatible",
> timecode="\x02D:03.11.03;T:1;U:13.40.17;    \x03", poll=9, noreply=0,
> badformat=0, baddata=0, fudgetime1=10.000, fudgetime2=0.000, stratum=0,
> refid=DCFa, flags=0,
> refclock_time="c350cb31.00000000  Mon, Nov  3 2003 12:40:17.000 UTC",
> refclock_status="TIME CODE; (LEAP INDICATION)",
> refclock_format="Meinberg Standard",
> refclock_states="*NOMINAL: 00:09:48 (100.00%); running time: 00:09:48"
> Why I have a diference between the hour that appear in timecode field when
> I compare this with refclock_time field? How can I set the fields to show
> the same time?

The timecode string displays the ASCII time string which is sent by the
radio clock device. This is currently Central European Standard Time,
UTC+1h. The status fields in the time string (the spaces before " \x03")
indicate that time zone, and thus the offset of 1h from UTC. 

The NTP parse driver subtracts the time offset and supplies the resulting
UTC time to the NTP kernel, so your computer's time is correct. 

The output of some ntpq commands should be the only places where you can see
the incoming local time, and it would be the only thing that would change
if you open the radio clock housing and change the DIP switches to let the
clock output UTC in the serial time string. The NTP parse driver would
determine an offset of 0h instead of 1h in this case and your system time
would be as correct as it is now.

Hope this helps.



Martin Burnicki

Meinberg Funkuhren
Bad Pyrmont

More information about the questions mailing list