[time-nuts] TV Frequency control - slightly O/T.

Magnus Danielson magnus at rubidium.dyndns.org
Mon Oct 8 20:56:36 EDT 2007

From: "Kit Scally" <kscally at bytecan.com.au>
Subject: [time-nuts] TV Frequency control - slightly O/T.
Date: Tue, 9 Oct 2007 09:28:55 +1000
Message-ID: <001a01c80a02$feb3d5b0$3601a8c0 at BYTECAN.com.au>

> ); SAEximRunCond expanded to false
> Errors-To: time-nuts-bounces+magnus=rubidium.dyndns.org at febo.com RETRY
> Just noting a few points made recently on this topic.
> Do not confuse the Studio colour subcarrier frequency
> (4.43 or 3.58 MHz) accuracy - or the digital equivalent
> clock frequency - with the TV transmitters carrier
> frequency of transmission.  They are totally independent
> of each other and may, or may not share a "common clock"
> source.  

Good point. I ment to comment on that.

> As others have noted, the TV carrier frequency need only
> be held to kHz, but is (and has to be) held to much
> tighter tolerances (Hz) for co-frequency (or "offset")
> operation.

For analogue TV systems, that may be true... yes.

> Geostationary satellite Doppler observed on downlink
> signals is far from random. A vectorscope shows this
> very dramatically.  It is entirely predictable and
> caused by diurnal changes in the satellite's position in
> it's 3 dimensional  "box" resulting in the spacecraft's
> altitude (Z plane) varying +/- tens of kms about the
> mean distance of 35,786 kms.  This daily oscillation is
> NOT corrected by the spacecraft's on-board thrusters,
> neither is the associated "figure 8" wobble in the X-Y
> plane. Gas-guzzling orbital corrections are used only
> when necessary, normally monthly. 

It should be mentioned that when the 8 meter parabols is not doing anything
"usefull" then train themselfs on the figure of 8 patter of various satelites
so that they have a recent number.

> In the analogue PAL/NTSC world, received satellite
> signals that are re-broadcast at a remote Tx location
> have their video passed through a 'frame store' which
> regenerates Fsc, line & frame sync from a
> locally-generated crystal oscillator, so removing the
> received incoming Doppler shift.  (This has tricked many
> a country T&F nut who thinks this source of 4.43 MHz is
> derived from a Rb or Cs source !)

Most frame-stores I've seen either have a PLL or is externally timed.

> Similar issues exist today in a digital TV & analogue
> transmission chain where the incoming SMPTE multiplex
> from the studio (carrying an embedded master clock
> signal from the studio) is decoded and a composite PAL
> signal generated for transmission.  In these instances,
> the incoming clock may be recovered to lock/steer an
> oscillator generating the required 4.43 MHz subcarrier
> frequency. It would be a bold system engineer who relied
> on the incoming digital multiplex's as his primary clock
> source to keep his Tx on frequency!  As Murray alluded
> to earlier, GPS technology have made this task simpler
> and considerably cheaper.

It has also made the TV broadcasting considerably more vunreble.

The DVB-ASI interface (which is used to bring the MPEG-2 Transport Stream to
the transmitter) is a mess timingwise, and no-one in their right mind would use
THAT for their carrier timing. It could have been good, but reality speaks
against it. With the use of Single-Frequency Network techniques, there is need
for PPS and 10 MHz at all transmitters. This is used both for SFN timing and
carrier frequency. Neither the DVB-ASI or SDI would allow the required timing
and stability. The SFN advantage makes system engineering of a national
broadcasting network easier, but it also raises new problems.

There are countries in which GPS can not be expected to work on a continous
basis. This brings "interesting" challanges for those doing modern broadcasting
networks. There are naturally solutions to be applied. GPS have started to
become a real system-hazzard for some. It has been too easy to just apply.
The backside of the coin have now started to show.

Oh, some of the new DVB-T transmitters even transmitt on channels so they may
locally kill the GPS signal on their overtone. :)

> When the USA, Australian/NZ & UK analogue transmissions
> are turned off in 5-10 years time (ha !), most of these
> "problems" go away I guess.

Well, in the meanwhile we are doing it here. We are learning from the mistakes
as we go along.


More information about the time-nuts mailing list