By Chris Codella, W2PA
February 22, 2009 - updated 23 Feb 2009
There have recently been some articles in QST describing unusual HF propagation on 80 meters, along with some explanations for the phenomena. For example:
 Echo Delay Measurements, Martin Ewing, AA6E, Technical Correspondence, QST, March 2009, p54.
 Magnetospheric Ducting for Delayed 3.5 MHz Signals, Sverre Holm, LA3ZA, Technical Correspondence, QST, March 2009, p55.
In addition an article in 1980 offered an explanation, and was one of several citations in .
 The Magnetospheric Echo Box - A Type of Long Delayed Echo Explained, O. G. Villard, W6QYT, et. al., QST, October 1980, p11.
This one was the first to explain "long delayed echos" (LDE), that had been described for decades, as resulting from magnetospheric ducts that occasionally form high in the atmosphere. The article also describes the contributions of amateurs to the scientific work, observing and collecting data. Backed by measurements with orbital probes, this phenomena is well documented and presented as the proven basis for LDE. To summarize, a "duct" of lower density electrons is formed along the Earth's magnetic field lines from northern to southern hemisphere. When conditions are right, an HF signal can enter the duct in the northern winter, when the ionospheric layers are thinner, travel along the duct for thousands of miles, bounce off the denser ionosphere at the southern hemisphere end where its summer, and return to its origin via the reverse path. (Read the article for a lot more detail.)
Having never heard this before and certainly never expecting to hear it on 80m (of all places), I made some recordings over a period of about 45 minutes - roughly as long as the effect lasted - and present those results here.
The conditions were good on both 160m and 80m that night and I had been working into Europe and Central America quite easily. I was running 1500W into an inverted-V dipole suspended from a tree with its apex at about 60 feet above ground and its wire elements sloping down to about 6 feet above ground. The antenna wire runs north-south.
The figure below is a plot of the receiver audio for a sequence of three dots recorded at 0257z, February 16, 2008. The transmitted signal is indicated by the three, zero-amplitude sections indicating silence since these are the intervals when the receiver was muted. Measuring from the trailing edge of the silent interval to the trailing edge (i.e. the beginning of the decay of the wave form) of the received echo signal indicates a delay of 219 ms for all three echos. The darker gray shaded area shows the boundaries of one delay measurement. I used the trailing edges to eliminate any ramp-up of the outgoing RF signal and get a sharper edge on the echo. All switching is solid state - transistors in the FT-1000D and PIN diodes in the amplifier. Thus, I'm assuming little or no delay between the receiver un-muting and the termination of the outgoing signal.
Notice that the echos aren't completely continuous but sometimes appear split, especially in the case of the third one - this can be plainly heard in the recording. In addition, immediately after the mute interval, there is a short burst, difficult to hear on the recording. I'm not sure if this is a burst of noise as the receiver AGC engages, or some other echo with a very short delay (~20ms). The tone of the echo is quite pure and close enough to my transmitted frequency that any difference is not detectable by ear.
About half way through the event, at around 0311z, I began to notice that the echos also developed a slight frequency shift, just enough to be noticeable by ear, a shift I'd estimate at between 10 and 20 Hz higher in frequency. It was evident since it started out matching my sidetone frequency, indicating an identical frequency. After a few minutes it returned to zero-beat.
A plot taken at 0322 is shown below, a little while past the peak of signal strength - roughly S9 or so on the s-meter of my FT-1000D transceiver. The delay is about the same, at 214ms.
By 0330z, the echos had become significantly weaker, as shown below.
The first echo is clearly visible but the second and third are quite a bit lower. They continued to fade up and down for a while until the effect could no longer be heard at all - around 0345z. The delay was nearly constant during the entire time, and roughly 217ms near the end.
[Added on 2 March 2009]: I measured the delay between the ending of RF output and the receiver noise recovery (un-mute) at 4.7ms. So this means the above measurements underestimate the delays by that much, resulting in corrected measurements of 224, 219 and 222 ms for the three cases above.
During this time I also tried transmitting on both the 160m and 40m bands but could hear no such effect.
In , LA3ZA recently described magnetospheric ducting as an explanation for these kinds of echos. These ducts are basically 80m wave guides . He argues that K4MOG's echos  were probably also due to ducting and gives several reasons for this conclusion: A simple, around-the-world propagation path should introduce a delay of only 138ms, whereas K4MOG's were at least 165ms. Furthermore, the echo signals, like mine, were very strong - seemingly much too strong for a round-the-world trip on 80m. Also, citing the 1980 article by W6QYT , he calculates that at K4MOG's southern latitude, the delay would be only around 143ms and hard to distinguish from an around-the-world delay. Other factors were favorable as described in  and  - it was near a solar minimum, was in the northern hemisphere winter, and took place in late evening, before midnight.
The 1977 echos measured in Seattle and cited in  had delays of about 225ms, similar to the ones I recorded above.
Sverre, LA3ZA, kindly repeated his model calculations for my magnetic latitude (51.5 degrees, as opposed to my geographic latitude of 41.7), and came up with a predicted delay of 214ms - very close to my observed range of 214 to 219ms (or 219 to 224, corrected for QSK delay). His plot of the modeled magnetic duct path is shown below (thanks Sverre!). There is much more information about Long Delayed Echos and other unusual HF propagation phemomena on his web site.
To conclude, it seems likely that what I heard was the effect of my signal propagating out and back through one of these ducts. In 40 years of operating HF this is the first time I've ever heard it, but from now on I'll be regularly looking for it to reappear - at least until sunspot cycle 24 gets going and makes it even less likely.
Click the link at the bottom of the page to hear a snippet of the 0322z recording.
Copyright © 2009 Christopher F. Codella, W2PA. All rights reserved.