1. Analyzing the Tracks/ 2. Analyzing the Numbers / 3. Conclusion/ 4. Forum Discussion

Starting out in the small streets of Nogent/Marne just East of Paris, with a lot of signal obstruction. The Sony system is not able to keep up, with the position jumping around quite a bit, likely resulting in some poor guidance. Of interest too is the markedly improved performance of the Globallocate based HP hw6515 with the beta 1.29 software version, even compared to v1.28 that had brought some good improvements compared to the original v1.19. The SiRFstarIII based units that have a fix (that rules out the G100) do very well too, although some "rounding" of the track can be seen during changes of direction.


Driving around the local hospital with buildings on the side for some strong multipath "bouncing" effect but other than that the environment is easier than above and apart from the G100, the sytems give similar results, with the tracks being "pushed" by reflected signal, but thanks to "map matching" (aka snap to road), guidance should be unaffected. The G100 now has a fix, but accuracy is poor and the track it recorded bears little resemblance to the actual one.


Succession of sharp turns around a block of apartments. Same observations, the SiRFIII units "rounding" the turns, especially the G100. The Sony has a "lapse" between the apartment blocks.


Round and round we go around the roundabout surrounded by trees. A good way to get a sense of the regularity of a track, and previous comments are confirmed here. Solid SiRFStarIII performance especially for the Globalsat BT-338, as the Asus drifts a bit and improved performance for the v1.29 HP6515 vs v1.28

The striking data at the top of the table are the invalid fixes, i.e. no position available, for the Sony but even more significantly for the G100, meaning no guidance for 35% of the time, unacceptable when the competition does so much better. The number of satellites in the fix below that shows the "headroom" available before the fix is lost (<3) so other than the G100, there is a good buffer. Signal levels are a bit more difficult to interpret, but generally speaking, the higher the maximum level, the better, as that indicates a more sensitive antenna, when comparing the same chipset. Clearly the G100 is at a disadvantage compared to the other SiRFstarIII systems. The data for the v1.29 6515 looks "odd" and that could come down to the "beta" status of the software.