KM6I, Gordon, in Palo Alto also heard echoes of his own signals and recorded them. In his blog he analyzed the delay from the output of his transceiver and found 157 ms. He found that to be so close to the round-the-world time for signals of 138 ms, that he assumed that to be the cause.
I don't agree, so I took the location of W6FB at locator CM97ah (Santa Clara) as a starting point for computing delay. This is latitude 37.31 and longitude -121.96 and gives a geomagnetic latitude of about 42.5 degrees. Then I put it into my program for computing path length along geomagnetic field lines assuming a height of the reflecting ionosphere on the opposite side of 100 km. The result is shown in the figure and predicts a delay time of 126 ms. My estimate of uncertainty is +/-5 ms.
The delay value is slightly less than 138 ms and easy to confuse with a round-the-world path. The challenge with estimating delays like this from the signal is that amateur transceivers may have an unspecified delay between start of transmission and start of sidetone. Measuring on the audio output as done here, measures the sidetone, not the actual RF.
I discussed this source of error in my 2009 QST article "Magnetospheric ducting as an explanation for delayed 3.5 MHz signals." Therefore the measurement shown above may fit with 138 ms just as well as with 126 ms, it depends on the actual transceiver's delay.
Other properties of the echo, such as the amplitude of the echo which according to W6FB at times was louder than the direct signal, also point to the duct theory as the explanation.
Others have heard such echoes also:
- 165-168 ms, K4MOG, February 2006, 80 m. Listen to the signal here. This event is also the object of the analysis in my QST article referenced above.
- 210-220 ms, G3PLX, November 2006, 160 m. Listen to the signal here.
- 214-219 ms, W2PA, February 2008, 80 m.
- 237 ms, OZ4UN, January 2009, 80m, Listen to audio file where also OZ7BQ very close by heard the echoes.
- About 200 ms, G3ZRJ, January 2012, 80m. Also heard by GW3OQK, 100 km apart.