2005-12-31 Leapsecond mistake by HBG


Poul-Henning Kamp
phk@FreeBSD.org
Revision $Id: index.html,v 1.5 2006/01/07 23:34:54 phk Exp $

HBG leap second theory

The swiss radio station HBG transmits time signals at 75 kHz using a modulation which is close to, but not quite identical to the German DCF77 station on 77.5 kHz.

The link on metas.ch's homepage which looks like it should point to a pdf with the signal definition actually points to a description of their modem service so we only have the rather brief and clearly incomplete description on their webpage

The correct link is on the German and French pages and a bit of link manipulation gives this link for the english version.

This document describes the tripple and quadruple pulses and documents a sensible DCF77 like handling of the leap-second.


HBG leap second practice

Here is what I received from DCF77, HBG and Rugby during the leap second at the end of 2005:

If we zoom in on the minute before the leap second, we can see the double pulse described on Metas homepage quite clearly.

What they don't say on the home page is that they send a tripple pulse for hours and quadruple pulse for days.

And here is the actual leap second:

The leap second is inserted after the tripple hour marker.

Admittedly the Metas homepage does not specify how the leapsecond is handled in any level of detail, but this looks like a bug to me.

The next minute after the leap second is correct:

The plots are available also in PDF format.

The raw data is a 400GB 5MSPS*12bits capture of the entire VLF band which is available to anybody who wants it.

Valid HTML 3.2!