Following the SCUBA2 engineering and subsequent inclinometry we have acquired 50 RxA pointing data in three nights (UT20080412, 13, 15) of admittedly rather poor (grade 5) weather. Individually, they show statistics as follows

         UT       FE       No.     CA  +- dS       IE  +-  dZ
                                      (daz)           (del)
       20080412   RxA       18     +1.2 +- 2.3     -0.5 +- 3.0
       20080413   RxA       13     +0.2 +- 2.2     +1.4 +- 2.3
       20080415   RxA       19     +1.7 +- 2.7     +3.6 +- 3.2

and combined without any adjustment for nightly collimation differences they behave as follows:

Formal rms's in the (az,el) residuals (dS,dZ) are (2.5", 3.7") and there are significant trends seen in the residuals as functions of both azimuth and elevation (see the centre, upper-left and upper-center panels, for instance). Some doubt may exist about the quality of data taken in such weather, but a full 7-parameter fit seems warranted nonetheless. A new model is obtained with predicted rms's of (2.2",2.7"):

and this model has been installed (HST 11:15 15 April 2008).

 TPOINT       old        new         approximate
  term       value      value    change  significance

  IA       -263.61    -254.55      +9       1.5 sigma
  IE        -52.28     -54.21      -2        1
  NPAE      +53.56     +62.65      +9        1
  CA        -85.00     -96.35      -9        1
  AN        +17.76     +20.01      +2        4
  AW         -4.90      -6.45      -2        3
  TF        -10.05     -15.88      -6       2.5

The RxA data from the subsequent night are added to the table, as are statistics from the combined HARP data taken during 4 nights in the week thereafter:

         UT       FE       No.     CA  +- dS       IE  +-  dZ
                                      (daz)           (del)
       20080412   RxA       18     +1.2 +- 2.3     -0.5 +- 3.0
       20080413   RxA       13     +0.2 +- 2.2     +1.4 +- 2.3
       20080415   RxA       19     +1.7 +- 2.7     +3.6 +- 3.2

       20080416   RxA       15     +0.7 +- 1.8     -0.1 +- 2.4

     20080416..22  H        49    +15.9 +- 2.1    +31.5 +- 3.0
       20080425    H        13    +15.7 +- 1.6    +32.2 +- 3.4

[ The nominal collimation offsets for HARP are currently (+15",+35"). ]

Even without a working K-mirror model, HARP pointing suffers errors only slightly worse than those for RxA.