Video synchro drift

[Michel Bardiaux <mbardiaux@xxxxxxxxxxx>]

I have at last completed my measurements of long-term drift of the
timing of broadcast video frames w.r.t. the 'ideal' 25fps of PAL. Over
10 million frames (10 exp 7), I observe a drift of 14millisec/24h on one
particular channel (maybe others have a different drift). And no, this
is not simply the system clock drift, it is synchronized with a DCF77
receiver every 30 secs.

I dont have PAL norm texts available. Can anyone, broadcast
professionals in particular, tell me whether this value is reasonable?
Or could it be some more subtle artefact of the measurement process?

Season greetings,
--
Michel Bardiaux
Peaktime Belgium S.A.  Bd. du Souverain, 191  B-1160 Bruxelles
Tel : +32 2 790.29.41