[ntp:questions] gude.info

Rob nomail at example.com
Tue Mar 24 09:58:36 UTC 2009


Towli <towli at oldboyz.dk> wrote:
> During the weekend - saturday evening 18.27 according to the syslog - it 
> synchronized with the (poorly recieved) Dcf77 signal and claimed the date to 
> be 27/3 2013...
> I sendt it back to the seller today, but was wondering if this is a certain 
> sign of a defective clock, or just a sign of a realy messed up dcf77-signal 
> under poor receptions conditions...

It probably is a combination of a messed up signal and poor software in
the clock.
The DCF-77 signal provides only limited error detection possibilities, and
a designer of a device that uses the signal should know about that, and
design suitable countermeasures against wrong time values in the product.

For example, when the date 24/03/2009 has been received for quite some
time, and then a datagram with date 27/03/2013 is received, the clock should
not jump to that date.  It should ignore the value and receive another
datagram.

When the clock apparently does not do that, your safest bet is to return
it (not for replacement, but to cancel the order).  Using this unreliable
device will certainly cause trouble sometime in the future.




More information about the questions mailing list