Hi Gedas (and all),Thanks for the links to both your TDF sideband pix above and your previous video capture of the 162 kHz TDF carrier with sidebands. Thinking about TDF sidebands has become a bit of an addiction for me--particularly useful for inducing sleep at bedtime. What I could not understand was how TDF’s phase modulation (PM) would induce a roughly 80 Hz wide modulation envelope as seen in your original video capture (your TDF Digital Data post on 2-4-2026). Your newer deeper dive static screenshots here show that the wider modulation envelope is actually made up of many well-defined frequency spikes.
The Wikipedia page (thanks, Glenn) (en.wikipedia.org/wiki/ALS162_time_signal) provides two seemingly conflicting format descriptions in the same article. The first one, near the top (Signal Format section), states that TDF’s PM modulates the carrier by +/- 1 radian (about 57.3 degrees of a 360 degree cycle; 2 x pi radians in 360 degrees) over 0.1 sec, implying a rate of change of 1 rad/0.1 sec = 10 rad/sec = 10. This would lead to a square (not gradual) carrier frequency shift of (10 rad/sec)/(2 x pi) = 1.59 Hz for both sidebands. However, more detailed information further down in the Phase Modulation Pattern section (below the big multicolored table) clarifies that the 0.1 sec modulation period actually consists of 4 x 0.025 sec subperiods, each of which has a 1 rad advance or retard (up, down, down, up) such that over the 0.1 sec period there are 4 x 1 rad shifts but no net change in phase, so a rate of phase change of 1 rad/0.025 sec = 40 rad/sec within each of the 0.025 sec subperiods. This means that the square carrier frequency shift for each of the 0.025 sec subperiods is (40 rad/sec)/(2 x pi) = 6.366 Hz. Call it 6.4 Hz. With both U and L sidebands, this would account for a roughly 12.8 Hz envelope width, extending from about 161,994 Hz to 162,006 Hz. So that’s progress but still does not explain the much wider roughly 80 Hz modulation envelope around 162 kHz seen in Gedas’ video capture.
Clearly, that “extra width” of the observed 80 Hz modulation envelope includes a lot of the smaller spikes seen in Gedas’ deeper dive static screenshots. But where do those come from? Intriguingly, there’s a statement that “Additional non-timing data is sent by phase modulation during the rest of each second.” Yikes! A reminder that the above time data PM stuff only accounts for the first 0.1 sec (or 0.2 sec in the case of a binary 1) of the entire second! So, what is the nature of the transmissions during the remaining 0.8-0.9 sec of each second? A hint is provided by the fact that TDF modulation and formats are reported to be highly similar (though not identical) to German time station DCF77 on 77.5 kHz. Again using Wikipedia (en.wikipedia.org/wiki/DCF77), it states that “for 793 ms beginning at 200 ms, each time code bit is transmitted using direct-sequence spread spectrum. . .using 15.6 degree phase shift keying.” (Note: 15.6 degrees = 0.27 rad). Also, the subperiods used by DCF77 for each spread spectrum phase shift are much shorter than the 25 ms periods used by the TDF PM time encoding at the beginning of each second (see above), running for only 120 cycles of the 77,500 Hz carrier, or roughly 1.6 ms for each spectrum jump! This then leads to a rate of phase change of 0.27 rad/0.0016 sec = 169 rad/sec, which then leads to a 26.9 Hz carrier deviation (169 rad/(2 x pi) = 26.9) for each spread spectrum jump, leading to a U + L sideband modulation window of about twice that or 54 Hz. Now, this is all DCF77 data, but if TDF uses something similar, such as keeping either the 120 cycles “bite” (of the 162,000 Hz TDF carrier) OR the 1.6 ms spectrum jump timing the same as DCF77, I think we’re getting close to being able to account for Gedas’ observation of a roughly 80 Hz wide total modulation envelope for TDF.
Lastly, since we’re talking about very short time periods with square wave jumps of PM applied to the carrier, the presence of absence of any given frequency spike seen in Gedas’ instantaneous screenshots would depend on being “inside” a given 25 ms or 1.6 ms window (or whatever) at the precise moment he triggered the screenshot. So, practically speaking, no two screenshots would be the same, which is my interpretation of his 5 different screen shots.
And if you’ve read this far, I’ve got to hand it to you. My biggest concern now is, what thought processes will I use to put myself to sleep tonight?
73 to all,
Bruce WA1HGJ