SUMMARY
The 's-correction' (for the elevation pointing problems occurring across transit) may not be working as intended.

To combat the elevation pointing problems associated with the reversal of direction in elevation - the 'transit' problem - a new version of tel_empirical.dat was installed on 16 May 2000. It was tested on 22 May 2000 , and seemed adequate at that time. That test was (somewhat asymmetrically) repeated tonight by tracking 3c279 from (az,el) = (138,57) at HST 18:28 to (191,64) at 20:15, using the map16 method. Seeing was about 1", but falling rapidly through this period. The data are shown below :

The azimuth data appear completely normal through transit with an rms scatter of 0.8", but while the elevation residuals are well behaved in the hour prior to transit, they still show the 4"-amplitude s-feature that the correction is supposed to remove.

Thomas Lowe confirms (June 12) that the use of TEL_EMPIRICAL.dat is reported at the time of the loading of the TEL task, and the contents of the file appear valid at this time.

2000 06 13
The test was repeated again by tracking 3c279 from (az,el) = (156,62) at HST 19:02 to (210,61) at 20:44, using the map16 method. The data contained many (14) individual integrations that did not yield valid centroids. The weather was quite poor, although the seeing averaged only 0.9". Many other centroid determinations (of the 150 original) have been discarded as being improbable. The remaining (98) data are shown below :

There is no sign in these data of a problem with the s-correction.

The rms scatters in daz and del are (1.6",1.6").

This is of some relief, given the measures of two days ago, but clearly the experiment needs repeating again in better circumstances to be sure.