Thanks Jim,

I would be helpful to understand what causes these phase variations. And
in a next step it would be nice to learn how to predict / estimate the
phase variation. This would allow to correct the phase, for getting a
(valid) decode. But that is probably not easy at all and it would also
rise the question if using 3rd party propagation data is allowed to get
a valid decode at all...

73 es HNY, Stefan

Am 29.12.2017 21:53, schrieb [email protected]:
> Stefan and Paul,
>
> I had been guessing issues in or near the receiver, as in your comments below, and the near field reference sounded like a great idea.
>
> I thought I would mention that on a few occasions I've been surprised by atypical phase variations on minutes or hours timescales for signals along similar paths, down the US east coast over land-water-land-water-land-water-land... paths at distances up to 1500 km or so, when concurrent signal phase in (for example) Florida (mostly water path) and inland (all land) at roughly similar SSW bearings looked stable. On one such occasion I checked phase sensitivity of the high-variability path (many land-water transitions) using LWPC, and compared with sensitivity of the mostly-water and all-land paths using LWPC. The LWPC comparison was qualitatively similar to the empirical comparison: high sensitivity of phase to ionospheric variations along the path with many land-water transitions. Guessing this is not the explanation for VO1NA-Forest night-to-night phase variations, but I thought I'd mention it, perhaps for future reference. I thought briefly about ice variations along the path but guessed that any ice would be far too thin to matter.
>
> Also thought I would mention that in general, for path lengths shorter than 1000 km or so, lower VLF frequencies generally have larger groundwave-skywave-interference phase variations on minutes/hours/diurnal timescales than do higher VLF frequencies, but this should not generally be a factor at 8270 Hz for distances much greater than 1000 km, i.e. not a likely factor on the VO1NA-Forest path.
>
> 73,
>
> Jim AA5BW
>
>
> -----Original Message-----
> From: owner-[email protected] [mailto:[email protected]] On Behalf Of DK7FC
> Sent: Friday, December 29, 2017 10:43 AM
> To: [email protected]
> Subject: Re: LF: VLF NA Ebnaut
>
> Am 29.12.2017 08:51, schrieb Paul Nicholson:
>
>> I've examined the software at Forest and all appears to be working
>> fine, at least as well as my own.
>>
> OK Paul, thanks for working that out.
>
> A near field reference for testing the phase stability is certainly a good idea. It should be there all the time ideally. Then one could even correct the faulty segments. A safety redundance.
>
>
>> If we see distant
>> MSK phase changing (which we probably will), is it caused by
>> the rx or the propagation? Alpha signals are a bit tricky
>> to use when there are two transmitters interleaved in the signal being
>> averaged.
>>
> In my Alpha RDF spectrogram, the colours (indicating the phase) are reproducing day by day quite nicely. There are 2 glitches. One is when the TX switches the phase at 0 clock Moscow time. The other glitch is at
> 0 UTC. Then there is the slow change of the phase between day and night.
> I am in another location and it is another path though.
>
>
>> Perhaps just the weakness of the signal at Forest?
>> Eb/N0 -0.1 = S/N 7.5 dB in the 30.6 uHz bandwidth in which the phase
>> is measured.
>>
> Could be the reason. But you are comparing just 2 or 3 days. On my 1 character message at 17470.1 Hz there were no unexpected phase changes
>
> Another idea could be a damage in an electrical component in the high impedance preamp. Maybe a water drop that changes the capacity somewhere, leading to a phase shift. This would not affect the timing but the phase Just a guess.
>
> I find it most useful to get an overview of the system performance by looking on a 'wideband' spectrogram, covering e.g. 0...24 kHz and half a day or a day. You will immediately see if something changed, like the signal levels, the phase and timing and sample rate correction (when running an Alpha plot for example). Or you will see if a new QRM source appeared or disappeared. All this without a particular search for eventual errors, just the usual daily view on that spectrogram. I would feel blind without that :-)
>
> So in the end, is it propagation or not? For me it sound unlikely that it is propagation. On 17.4701 kHz, over such a distance, maybe. But not on 8270 Hz on the path between VO1NA and Mike.
> Uuh, another strange idea: Maybe a faulty component in the preamp that changes the phase frequency dependently, like a capacitor moving the edge frequency of a LPF or HPF and changing the phase hereby?
>
> 73, Stefan
>
>
>
>
>