SCUBA returned to service on UT the 26th, and the combined data from several nights since then are shown below

The pointing model seems fine (as judged by the lack of systematics between the residuals and azimuth and/or elevation), but the plot of the elevation residuals (del)-vs- time (centre-right) shows a strong trend of del generally becoming more +ve through the night. This is usually indicative of a trend with temperature . . .

The values of the (THI.tide) parameters TEMP_SLOPE and TEMP_MEAN_SLOPE have been adjusted in recent months on the basis of pointing data taken with RxA and RxB. These relate elevation pointing to the temperature difference between the front and back legs, and to the overall mean temperature of the antenna, respectively. They currently have values of 6"/deg and -0.3"/deg. Two recent large-ish datasets taken with RxA on UT20040824/25 are shown below :

Both parameters seem to be working fine, although there is weak evidence for changing TEMP_SLOPE and for using a 2nd order for TEMP_MEAN_SLOPE. Any such adjustments would need confirming over a longer time interval.

Elevation pointing data taken in these few days days are shown below :

TEMP_SLOPE seems fine for this (much larger) dataset, but there is strong (>8-sigma) evidence for adjusting TEMP_MEAN_SLOPE from -0.3"/deg to -1.0"/deg (or maybe something even more negative if an eyeball fit to the bulk of the data is used).

Focus
The recent SCUBA focus data also need incorporating into this discussion:

      Axis    No.of data     mean    s.d.      Notes
     ------  ------------   ------  ------    --------------
        Z         30        +0.05    0.08        1
        X         11        +0.26    0.23
        Y         11        +0.04    0.14

Note 1 : Z = (-0.016 +- 0.005) * T_mean + 0.19  s.d. 0.07  6 < T < 14

The mean value for the X-focus (+0.26mm) is quite different from those for SCUBA, RxA and RxB (approximately zero) prior to SCUBA's recent temporary demise, and for RxA & RxB since then (and as reported most recently on July 22). Results for RxA & RXB during August are shown below:

      Axis    No.of data     mean    s.d.    Notes
     ------  ------------   ------  ------  ------------------
      RxA_Z       117        -0.04    0.14     1
      RxA_X        66        +0.07    0.32
      RxA_Y        61        -0.05    0.16    

      RxB_Z        87        -0.09    0.14     2
      RxB_X        52        +0.04    0.21
      RxB_Y        52        +0.01    0.13

Note 1 : RxA_Z = (-0.030 +- 0.005) * T_mean + 0.20  s.d. 0.12  4 < T < 14
Note 2 : RxA_Z = (-0.026 +- 0.005) * T_mean + 0.09  s.d. 0.12  2 < T < 13

There is enough evidence (here and as reported over the past 2 months) to support adjusting the Z-focus by about -0.02*T_mean (i.e. to adjust DZT_COEFF, currently +0.070 mm/deg by -0.02 mm/deg, to +0.05 mm/deg), (and that was done today at 14:30 HST), but this should not impact the pointing. The X-axis data for RxA & RxB also show only weak trends with temperature, so there is unlikely to be any strong feedback from this that would solve the 'problem' with the elevation pointing.

20040901
I could stand it no longer : I've changed TEMP_MEAN_SLOPE to -1.0"/deg.