SUMMARY
The amplitude of the central bearing defect has reached its lowest value since measurements began - just in time for its replacement next week. Confirmation will be attempted at some time in the days remaining.

Mars was tracked using repeated RxA3I fivepoints, with the empirical corrections left enabled, from :

• (az,el) = (116/29) at 20:28 HST to (132/46) at 21:56

The elevation residuals are very well behaved throughout, with an rms scatter of 0.4". This is also the value of the seeing as measured by the SAO phase monitor and is therefore the expected noise on the azimuth residuals also.

The azimuth residuals show an overcompensation by the empirical correction for the defect expected at azimuth 125. There are also systematic structures in what ought to be baseline measurements prior to the event. With the elevation above 30 degrees or so throughout, and with the good behaviour of the elevation residuals, it is hard to attribute any of this to atmospherics.

However, any baseline slope across almost 10 degrees of azimuth in the immediate vicinity of the event is no more than 1" and so will be ignored hereafter. Peak-to-peak excursions amount to 4.7", and allowing for the lag of the logging process behind the center of the fivepoint, the event occurs at (az,el) = (125.6, 40.6). This agrees well with TEL_EMPIRICAL, which has this event centred at azimuth 125.65 degrees.

The observed error then corresponds to an error at the horizon of 4.7"/cos(40.6) = 6.2". The current version of TEL_EMPIRICAL.DAT has a peak-to-peak amplitude of about 12.7", as established on 990403 , so an overcompensation of 6.2" implies a true amplitude now of 6.5". The additional structure, which is significantly larger than the noise, undoubtedly implies changes in the defect profile as well as in the overall amplitude.

The new data are added to the collection .

Note - I shall provide a more user-friendly version of the fives.icl routine a.s.a.p.

990429
Remo installed a new version of TEL_EMPIRICAL.dat, but attempts the next night to check it stumbled. The first attempt, at the same azimuths as above, discovered a typo in the new version of fives.icl, and after correction, the second attempt was reported, by TBL, as producing " . . some negative data values which are playing havoc with the fitting routine" :

We'll try again.