Check the pointing after the nov05 azimuth encoder failure

    The azimuth encoder failed on 14nov05. It was reset using the index. A source was tracked after the fix and the azimuth error had an offset of a few arcseconds. With a single source it is difficult to separate out a model/source offset from an offset in the current position of the encoder.
    The pointing/calibration  data from 20nov05 through feb06 was used to check the azimuth encoder offset. Only feeds above 1 Ghz were used (lband and above).

    The plots show the pointing errors for the calibration data taken nov05-feb06 (.ps) (.pdf):

• Fig 1 Pnt error by src:  The colors/symbols show different sources. This includes the data from all the recievers.
• 1st plot: Za error versus za.
• 2nd plot: Za error versus az.
• 3rd plot: Az error versus za.
• 4th plot: Az eror versus az.
• Fig 2 Pnt error by receiver: The colors/symbols show different receivers. These plots are ordered a little different than the first set: first by za and then by az.
• 1st plot: Za error versus za.
• 2nd plot: Az error versus za.
• 3rd plot: Za error versus az.
• 4th plot: Az eror versus az.

Conclusions:

• There is no offset in the azimuth encoder.
• It is interesting to note the structure in the azimuth error as a function of azimuth. These are great circle errors so it is probably not an azimuth encoder runout of the rack gear. These features have been present for awhile (see model 15).  The pointing model can not fit these variations. Hopefully the azimuth encoder table would fix this.

 home_~phil