[Next] [Previous] [Top]

The Effect of Antenna Covers On GPS Baseline Solutions

Conclusions

With the spherically shaped antenna covers, the thickness and mounting of the covers can cause variations in the height solutions (figs. 8 and 9). Errors as large as 10 mm are seen when a one-quarter inch dome is mounted so that the antenna's mean phase center is not located in the center of the dome. If a one-eighth inch dome is mounted so that the phase center of the antenna was in the center of the cover, the error drops to 4 mm or less. Both error bounds above are for solutions with troposphere parameters estimated at one of the sites. Generally, the noise of an L3 height solution with troposphere parameters estimated at one site is about 2 mm. The horizontal components of the baseline solutions do not noticeably change for these solutions, but results in Scandinavia show that there is up to a 10 mm change in horizontal baseline components with and without antenna covers (Jim Davis, personal communication).

The magnitude of the errors is also a function of the elevation cutoff angle used for processing (fig. 5). L3 height solutions using the Trimble conical dome, and with troposphere parameters estimated, vary by more than 30 mm between 10 and 20 degrees elevation cutoff. Differences in solutions that vary as a function of elevation cutoff angle, and differences in height solutions with and without troposphere parameters estimated implies that antenna covers cause errors that are mismodeled as tropospheric delay. This is important to scientists who use GPS to accurately monitor atmospheric water vapor.

In addition to the antenna covers, the metal mounting plate of the Trimble conical cover appears to cause a baseline height error as well. Both the Trimble Dorne Margolin choke ring antenna and the Trimble L1/L2 GEOD GP antenna are affected by this large mounting plate (figs. 5 and 6). The mounting plate causes the largest signal when it is attached under an antenna like the Trimble L1/L2 GEOD GP (i.e. one that uses a ground plane to suppress multipath instead of choke rings).

As is the case with many errors in very high precision GPS surveying, it is possible for the effect of domes to cancel out at the single difference level. For very short baselines with two antennas using the same cover type, a common error is introduced to both stations which is subtracted out when the single difference observation is formed. However as baseline length increases, and the two antennas observe a satellite at different azimuth and elevation angles, the errors no longer cancel at the single difference level. It is also likely that the observations from one station will be processed with data taken from another station with a different type of antenna cover, or no cover at all. In these examples the effects will not difference out, and an error will be introduced.

Ideally, it would be best if no cover were needed over an antenna. However, some permanent GPS sites require their use to keep rain, snow and animals away. In this instance it appears that the best solution is one where the cover is thin, nearly spherical in shape, and mounted so that the radius of curvature is centered near the mean electronic phase center of the antenna. While still producing an error in the vertical baseline component, the effect is minimized.

[Next] [Previous] [Top]

Last modified Thursday, 17-Nov-2005 03:58:00 UTC

Home | About Us | Contact Us | Support | Search | Facility | PBO | Education & Outreach

Comments: webmasterunavco.org
© 2009 UNAVCO, Inc.