LF: DBF39 phase jump experiment

[MarkusVester@aol.com]

Dear LF-group,

. All I can say now is that the phase jump is always a multiple of  25.7xx° (or so), but WHEN these jump's occur I can't say.

perhaps I should explain my observations:

The phase jumps on the signals of DCF49 (129.93) and DBF39 (138.83) are caused by the phase-continuous FSK modulation during the telegrams, which come irregularly, typically once every 16 seconds. This why on a hi-resolution spectrum the carrier looks like a 60mHz wide noiseband without a central line, and its frequency cannot be determined accurately in a conventional way.

A while ago I had concluded that with 340Hz shift and 200bd datarate, the phaseshift of each single bit excursion from mark to space would be 340/200 periods or 17/10*360 degrees. This has the consequence that the carrier phases in the intervals between the telegrams would always remain on a fixed raster of 36 degrees.

Detailed measurement using SpecLab's IQ-save revealed that there indeed is such a raster, however the bit-phaseshift is 47/28 periods. Frequency shift is as nominal, but the duration of a space-bit is slightly short (4.937ms), and the following mark-bits are prolonged to keep the average baudrate. Apparently all telegrams contain an even number of space-bits, so the phase raster is 1/14 of a carrier period or 25.714 degrees.

You can confirm that in a little experiment: Tune the station to a low audio beat frequency (eg 130Hz) on a stable receiver, and watch at the spectral band with 10mHz resolution (Argo 90s). Then switch back to full band view, and increase the audio signal to overload your soundcard input so that many harmonics are visible. Now take a detailed look at the 14th overtone (1820Hz) ... you will see a sharp line !! - That way, I could  measure the carriers of DCF49 as 129930.015Hz and DBF39 as 138830.033Hz. They seem to be stable at least on a 10mHz scale, but we currently don't know whether they are free-running or somehow locked to UTC or GPS for longterm stability.

This also means that one could actually use DBF39 to study ionospheric effects on signal phase, provided the SNR is still good enough to identify in which of the 14 possible phasestates the carrier is in ( >10dB in 1/16Hz bandwidth). Alternatively, one of us here could publish a log of phasejumps for postprocessing.

Another one worth observing is the 160mW transmitter-intermodulation on 135360.031Hz, a combination of 2xDCF49 minus precise 122500Hz from DCF42's pilot: carrying twice the frequency shift it has a 1/7-period phase raster.

Have fun playing,

73 de Markus, DF6NM

DBF_FAST.zip
Description: Zip archive