On UT 20050818, the s-curve correction for the so-called transit error was disabled and Mars was tracked using RxB and the fives method. Tracking started at HST 03:50 at (az,el) = (107,69) and ended at HST 07:00 at (258,63). Transit at elevation=83 occurred at 05:15 HST. Ed reported that "tau was just over 0.20 and very stable." The pointing residuals are shown below:

The formal rms of the residuals in (az,el) are (1.5",1.2") although small (+-2") systematic errors around transit and at the end of the run are seen.

The performance is not perfect - but is not bad either. I'd prefer to collect more data before reflecting upon the utility of the current correction algorithm.