All:

As David remarked, I did quite a lot of work on this matter of time delay a year or two ago. The work formed the basis for my talk at ALCON 2002 and the IOTA business meeting.  Some of it will be published a bit later in the ON.  Its out there in the queue of stocking stuffers.

The occasion of performing the tests was my beta test of Geoff Hitchcox' KIWI.  I got the GPS based KIWI together and it seemed pretty good.  I noted, however, that I could have constructed the circuit backwards, using an up going pulse when I should have used down going.  So I simply used KIWI to gate the oscilloscope and fed WWV into the y-axis signal amplifier.  The WWV pulse then occurs after the GPS one-second pulse.  Its a very stable display and one can study to his heart's content.

To make the work worth publishing I used a storage oscilloscope with digital output so I could graph the results. Here's the bottom line --

1.  The normal signal obeys simple, speed of light, predictions.

2.  As the signal fades so that the ticks are almost inaudible, the signal shows phase shift as the direct wave is combined with the refracted wave.

3.  I observed no delays greater than about  1.5x  the direct delay time. Any greater delays would have been inaudible.

For these tests I used a number of good receivers with a 100 ft long wire antenna at a height of about 50 feet. The vintage receiver was a Hallicrafters SX-28 of WWII and before vintage.  A competitive receiver was the best I could buy from RS.  Both showed the same results.

For our work we are not likely to be in error greater than 0.5x the delay calculated by the great circle method. This is much less than other errors we incur as we now time occultations.

Art Lucas, secy/treas, IOTA
5403 Bluebird Trl.
Stillwater, OK 74074, USA
Tel: 405-372-4506
Fax: 405-377-1474