[sdiy] ot: long wave radio reciever for time/frequency

[Magnus Danielson]

From: "Czech Martin" <Martin.Czech at Micronas.com>
Subject: RE: [sdiy] ot: long wave radio reciever for time/frequency
Date: Thu, 3 Apr 2003 09:38:08 +0200

Hi Martin,

You seem to have bumped into an area which I care quite seriously about and
therefore doesn't quickly shut up on... ;O)

> Yes, but I do not want to use the AM time code, just the carrier frequency.

The original DCF77 carrier is somewhat flawed. This have lead to a phase
modulation being used which allows receivers to remove some of the disturbances
of multi-path reception which relates to surface and atmospherically reflected
waves. I just *happends* to have documentations on the DCF77 including an
article about the phase-modulation enhanced receiver.

I would actually like to build receivers for all kinds of standard frequency
transmitters since I am a bit of a precission-frequency-loony. The DCF77
transmitter is run of a R&S Rubidium clock (I have one too!) which is being
constantly maintained by comparision with Cesium clocks.

> Indeed, the colour carrier of the "ZDF" station in Germany is locked to
> atomic clock (for some internal use).
> But that would mean hacking my TV.

You could easilly get a TV-tuner and extract it. This way you would be able to
ZAP TV-channels all you like.

The PAL color-carrier is to be +/- 1 ppm of the nominal frequency (in relation
to UTC if you follow the papertrail) and not change faster than 22.6 ppb/s.
This gives a good reason to have the station under a stable clock. However,
the color-carrier is to be in a strict frequency relation to the vertical and
horizontal sync, so the full system is run from the same source. There are
really good reasons in the TV-production and contribution networks to keep
things in really good sync, since when you are not you have to go through
frame-stores which re-sync the content, but frequency differances causes
constant slips and the framestores only reduces the hurdle but there is a cost
in quality to pay for that, so in practice you want really stable things.
Running TV-houses of GPS is how a stable reference is become cheap these days.

The downside of using the PAL color carrier directly is that multipath flaws
is not compensated. There is a sine-burst transmitted on a line which aids
receivers to basically get a TDT-measurement of the transmission path and be
able to compensate out the largest reflections (or indeed most) by means of
FIR-filtering. For precission reception of frequency, where you want to
minimize wander due to multipath effects, you want to operate in a filtered
environment. I don't think that standard TVs actually process the color carrier
this way, since their gain is only in removing ghost images and not in the
stability of the color carrier.

In PAL is also the color-carrier and line frequency interacting such that over
a 4-frame (8-field) sequence their phase relationship completes a full cycle.

Within TV-production is timecode embedded into the signal on separate lines.
This VITC might actually be in transmitted material. I have not heard any
details that they are or isn't. I know that some code is sent in US on NTSC
and that this caused many video-owners great greif since many stations have had
them way of chart and did not even know they where transmitting it and where
to adjust it. I have not dug into the facts if this indeed is the VITC or some
other TC. There was an interesting article in the IEEE Spectrum on that some
time back.

Anyway, if you know you have a stable and good transmitter from ZDF (I know
for a fact that they have _REALLY_ good senior technical staff in and near
them - I've worked with some of that people) you could also use other parts of
the transmission. You could be using the carrier of the full transmission,
since it needs to be as good as the color carrier if you think about it.
I sometimes move my TV antenna over to my spectrum analyzer and poner over the
spectra. The NICAM stuff is easilly spotted for instance.

Cheers,
Magnus