I sent this thread to a friend, and he replied: =======
> I bought a couple of pizza pans for the GPS Patch Antenna Improvement > Program... preliminary results: > > My 15+ year old Magellan patch antenna (actually made by Murata) > would barely function with the Tbolt. It was (barely) tracking three > satellites and going through long periods of holdover. A patch antenna is intended to be mounted flush against a ground plane. If he tried to use it unmounted, it would not have been happy. > I then mounted it on the bottom of a 16" pizza pan. It is now > tracking 8 satellites with no signal dropouts. OK. > If you are using a consumer/automotive type patch antenna, you might > want to try a pizza pan ground plane under it. I am doing a 48 hour > survey run to see how it performs compared to my other antennas. I > cannot compare it to itself without the pan since it would never > complete the survey. > > My geodetic grade antenna is a patch type antenna mounted on a 14.5" > aluminum disk. 14" diameter is a standard pizza pan size. I am > going to > try mounting the TAPR Motorola patch to it and see what happens. I > have a > good 48 survey with that antenna to compare to. Unless a ground plane is VERY large in comparison with the signal wavelength (about 8" for GPS L1), the gain _pattern_ of the antenna is strongly affected by the ground plane. (14.5" is emphatically NOT large in comparison with the L1 wavelength.) Not just the main lobe of the gain pattern, but also the minor lobes, including the BACK lobe(s) -- those directed at the ground rather than the sky. This strong sensitivity occurs because incident waves excite standing waves on the ground plane. Nothing interferes with the flow of electric charge/current from the upper surface of the ground plane around the edge of the plane to the lower surface of the ground plane, so the amplitude of the standing wave on the underside of the plane is usually as great as the amplitude of the standing wave on the upper side. Therefore, the antenna may have nearly as much gain toward the ground as toward the sky. If an antenna does not have much less gain toward the ground than toward the sky, then multipath interference from ground reflections will strongly perturb the observed group delays (code-modulation delays, or pseudoranges) and carrier phases of the signals received from the GPS satellites. The antenna pattern will also have fairly deep nulls in the upper (sky) hemisphere. So it's not a great idea to stick a GPS patch antenna on just any old sheet of metal. The ground planes used with commercial GPS antennas have dimensions deliberately chosen to minimize, or at least reduce to a tolerable level, the standing waves that could cause nasty pattern problems. (In an archive I have examples of patterns of GPS antennas with inappropriate ground planes. I also have computer-generated graphics showing the distributions of RF current on various ground planes. Measured gain patterns, and gain and current patterns derived by computer simulation, can be found in the open literature; but it'd take me a while to dig out references.) The best antennas for accurate position-determination by GPS have rather large and massive ground planes (e.g., 36" square) with special treatment of the edge; or the upper surface of the ground plane is ringed with quarter-wave chokes; or the antenna has no ground plane at all, and it achieves an upper-hemispheric gain pattern by means of arraying. I.e., the antenna comprises an array of several identical, circularly polarized, elementary antennas, or "array elements," spaced along a vertical axis; and the feedpoints of these elements are connected together by a manifold "feed," such that the signals received by different elements are combined with different relative amplitudes and phases in the input of the receiver. ====== -John =================== > Hi all: for the proper dimensions, see some of the sections of W1GHZ's > microwave antenna handbook to get the appropriate size for the pizza pan > choke. Mark may have gotten lucky... > Don _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
