37 data were obtained during the night, most during an intensive 2 hour period starting 04:00 HST. (Az,El) space is well (~uniformly) covered. Opacity was stable at ~0.10 throughout. The combined data are displayed below after the removal of collimation offsets of (dS,dZ) = (-4.6,-0.6) :

Residuals show some systematic trends : in particular dS-vs-Z (centre-left frame). Formal rms's in (dS,dZ) are (5.6",3.1") - rather horrible compared with performances seen earlier in the year (see even the report of a month ago). Results from GL865 and GXMon (##23,24) seem most deviant, although removing them only reduces the residuals to (5.2",3.1"). A full fitting/optimizing to all 37 points of all 7 TPOINT parameters improves the rms of (dS,dZ) only to (4.2",2.7") - tolerable enough in dZ but nowhere near nominal in dS. For completeness: the suggested changes to the model were

       coeff       change     value    sigma

  1     IA         +7.663   -254.30   12.844
  2     IE         +0.167    -19.18    2.820
  3     NPAE      +25.868    +48.13   13.037
  4     CA        -27.060   -185.01   17.331
  5     AN         +2.440    +10.80    1.016
  6     AW         +0.441    -14.70    0.951
  7     TF         +0.275     -3.98    3.964

It may be a complete red herring - but poor performance in azimuth is also a feature of the current beam shape (see HARP commissioning reports of Nov/Dec 2006).

20061214
13 RxA pointing data logged on UT20061214 (an early point on Saturn was removed as >5sigma deviant) gave rms's in (dS,dZ) of (2.0",2.4"). This is after accounting for collimation offsets (CA,IE) of (1.7",-0.8") from the current model (which dates from 29 Oct). Despite the results from yesterday, but mainly because of some chopper gain changes made during the day of the 13th HST (CAW) it appears that the azimuth performance is correctable, and that the model is still useful if not actually totally valid.