I've been fascinated by the concept and use of phase for a long time, really starting I think when I was introduced to the lockin amplifier as a graduate student. I knew about phase before that of course, but it was just then that I got a glimmer of how useful it could be. At its deepest levels phase is one of the characteristics of quantum mechanics that underlies our understanding of the physical world. But right now I want to explore its measurement in electromagnetic radiation, and in particular HF radio waves.

I'll use my Flex5000 radio as the detector and the NIST standard broadcast station WWV located near Boulder, Colorado, which broadcasts on 2.5, 5, 10, 15 and 20 MHz as the source to be measured.

I recorded about 50 seconds of the 5 MHz signal, with the Flex software depositing a wav file of the 48KHz sampled I/Q data on my computer. I read this file into the program Mathematica, where I did the analysis and plotting.

Here is a plot of the phase and amplitude of that signal over 50 seconds.

There's potentially a lot to be learned from these plots, but first I need to do some checking about sources of error. There are three possible sources of error in the phase plot.

First there is the source - WWV. But WWV is the NIST standards station and its phase, as transmitted, is about as constant as you can find on the RF spectrum. They are responsible for keeping the nation's official time and use Cesium atomic clocks whose frequencies are uncertain at the 3 * 10^-16 level, which amounts to about 1 second in 100 million years. The rf signals transmitted by WWV are normally stable within about a part in 10^12, which means the 5 MHz signal is constant to within about 5 microherz. That level of variability is undetectable on the scale of the above plots.

But of course that signal has to propagate from Colorado to Boston, and that happens when the 5 MHz rf radiation refracts off the earth's ionosphere, which is a very dynamic environment and significantly affects the received frequency. A rough rule of thumb is that skywave signals may vary by a hertz or so from their transmitted frequency, and the deviation is quite variable with many effects contributing - Doppler shifts due to ion cloud motion, changing electron densities driven by solar activity and time of day, interference between multiple signal pathways, and so on. If I'm trying to get a constant phase reference signal, this variability is a problem. But in fact I'm interested in exploring the effects of propagation, so its not noise - its data.

Before I can attribute the changes to propagation I need to make sure that my receiving system is stable enough that it does not affect the phase I measure in the above plot. I don't have an atomic clock to provide a local reference for my receiver, so I make do with what the Flex radio gives me. The Flex uses a TXCO - which is a temperature compensated crystal oscillator - and although they don't provide detailed specs another rule of thumb is that a reasonably designed quartz oscillator is good to about a part in 10^10 over short periods of time, such as the minute long plot above. They can drift a good deal more than that, due to temperature changes for example, but if I'm interested only in changes of phase in the plot above, that should be good enough for now.

There is often an additional source of uncertainty when using the computer clock to digitize the audio signal coming from a receiver, but in my case the digitizing is done by the Flex radio itself, which is presumably synced with the same crystal and hence accuracy as that of the RF mixer signal that determines its basic frequency stability.

So provisionally I can say that the variation in phase in the above plot is all due to propagation effects between Boulder and Boston, and I can start working at understanding what the signal is telling me. Stay tuned.

Ultimately I'm interested in a careful study of this type of data because Flex has announced a new transceiver family - the Flex 6000 series. These receivers will have a GPS disciplined oscillator which means that they will be able to measure absolute times to microseconds with frequency stabilities comparable to the Flex 5000 that I'm using now. I'm excited by the possibilities of coordinated measurements between a number of these stations. Since they are transceivers, they will be able to transmit as well as receive. Think about the possibilities of realtime tomographic measurement of the ionosphere by cooperating amateur radio operators. There are commercial and government ionosondes that do this now, but the number of stations is limited, and there is the real possibility that hams will be able to make some interesting contributions to our understanding of the ionosphere. And it should be a lot of fun in any case.

Tangled Spaces