Last time I looked at simultaneously recorded WWV signals on 10 and 15 MHz to see if there was any correlation between their phase variations due to propagation. The phases in each varied, but there didn't appear to be any common features in those variations. Five Mhz is evidently so large a frequency separation that the corresponding refracting portions of the ionosphere aren't very correlated.

There aren't any WWV frequencies that are closer together than 2.5 MHz, but we aren't stymied yet. WWV transmits more than a carrier - there are a number of signals that are imposed on the carrier by amplitude modulation. Here is a waterfall plot of about two minutes of the 15 MHz signal I looked at previously:

Waterfall spectrum of WWV on 15 MHz over two minute period.

The carrier is present at 0 Hz - this is the signal that I've been analyzing. There are two sidebands at 100 Hz: these carry BCD coded time information and I may explore these later. Of more interest are the two tones - 500 Hz during most of the first minute and 600 Hz during the second minute. You can also see the 1000 Hz minute ticks, and some wide band voice at one and two minutes. There are 1 second ticks of 5 cycles of 1000 Hz with 40 msec interruptions in the 500 and 600 Hz tones during those ticks, but those are hard to see at the plot scale and I'm going to ignore those for this analysis.

The tones provide two synchronized signals that are 1000 and 1200 Hz apart - this is at the other end of the frequency separation scale and I expect that there will be correlations between the sampled ionosphere patches at this scale. The signals are a little inconvenient in that they are not continuous, unlike the carrier, but its something to work with.

I computed the phase of each of those tones and the carrier during the time they were present, with the following results. The initial phase of each is arbitrary, so I shifted each to start at 0 radians. First the 500 Hz tone phase variation:

-500 Hz,  0 Hz (carrier) and +500 Hz WWV tones. — -500 Hz, 0 Hz (carrier) and +500 Hz WWV tones.

Next the 600 Hz tone phase variation:

-600 Hz, 0 Hz (carrier) and +600 Hz WWV tones.

And to put them in context, I plotted each tone segment on the full two minute carrier plot, again each segment phase starting out at the carrier phase at that moment:

Phase change of carrier and 500 / 600 Hz tones.

What can we say based on these plots? First the phase evolution on the three frequencies are highly correlated, but definitely different - this is encouraging. Without more data I'm reluctant to attribute this to propagation differences at these narrowly spaced frequencies. The phase of each of the signals changes at a different rate and since the plots are phase as a function of frequency, a constant slope would indicate a frequency offset. The frequency offset needed to give the roughly two pi (one cycle) phase change in 60 seconds is only 1/60 Hz which would correspond to a doppler velocity of about a third of a meter per second. Its certainly doesn't seem unreasonable to me that there might be a velocity shear of this size between two nearby patches of the ionosphere.

There is always the possibility of a problem with the analysis routines that I'm writing and I want to see more data before I can attribute my measured differential phase evolution with an ionospheric condition. This will take more samples, and over longer periods of time. Its unfortunate that the tones aren't continuous so I can't track them continuously and have to reset the phase zero at each segment, but you work with what you have.

And so I will title this post with the velocity shear number, but put a question mark at the end. Maybe I should put two question marks.

Tangled Spaces