Pointing
The plot below-left shows the entirety of pointing data for SCUBA during May, with plots for the nights of UT20050531 & 20050601 shown alongside:

The data show that the current model exerts very good control over azimuth residuals, and, on some nights, produces similarly good elevation pointing performance. Both the single-night plots above, however, show negative trends of the elevation residuals through the night - as mean temperatures fall. As seen in the analysis below, this may be a systematic effect requiring an adjustment of the parameter TEMP_MEAN_SLOPE.

The plots below show the entirety of pointing data for, respectively, RxA, RxB and RxW_C:

Even ignoring the occasional obvious outlier, there are trends of both azimuth and elevation residuals with elevation (centre frame, upper line and centre frame, centre line, respectively), confirming results from previous months that a separate heterodyne pointing model may be justified. Last month we said : "We continue to wait for SCUBA to return to service in order to more fully quantify, by experiment, these apparent, elevation-dependent pointing systematics." Well, SCUBA was out of action for much of May again, but is back in operation now, so perhaps these experiments can now be performed.

Thermal - i.e. analyses of the elevation residuals as functions of Receiver and antenna leg temperatures:

  FE  No.of data   fit                              Temp range
  --  ----------  ----------------------------     -----------
   S     115      del = (+0.23 +- 0.65)*T_diff     -0.7 to +0.6
                      = (-0.24 +- 0.07)*T_mean       4  to  15

   A     204      del = (+1.45 +- 0.67)*T_diff     -0.9 to +0.7
                      = (+0.25 +- 0.08)*T_mean       2  to  12

   B     182      del = ( 0.00 +- 0.62)*T_diff     -0.8 to +0.7
                      = (+0.17 +- 0.07)*T_mean       3  to  14
 
   W      31      del = (-1.51 +- 1.01)*T_diff     -0.7 to +1.1
                      = (+0.08 +- 0.20)*T_mean       3  to  11

Significant but inconsistent trends of elevation residual with mean leg temperature are seen; the slope of the trend in the SCUBA data being opposite to that in the heterodyne data. MEAN_TEMP_SLOPE would need to be made FE/Rx-dependent and used (and confirmed) with the appropraite pointing model.

This could get v.v.complicated . . .

Focus

      FE  axis  No.obs.    mean   s.d.  Notes ('significant' trends)
                           (mm)   (mm)
    ----- ----  -------  ------- ------ -----------------------------
    SCUBA   Z      56     -0.26   0.09   
            X      24     +0.12   0.20
            Y      24     -0.01   0.23

     RxA    Z      97     -0.28   0.10   1
            X      74     +0.07   0.21  
            Y      73     +0.05   0.21

     RxB    Z      75     -0.28   0.10
            X      66     +0.08   0.23   2,3
            Y      66     +0.09   0.21

     RxW_C  Z      10     -0.07   0.04
            X       9     -0.11   0.29
            Y       9     +0.29   0.17

Notes

1. RxA_Z = (+0.014 + 0.004)*T_leg     temperature range +2 to 12
2. RxB_X = (-0.55 + 0.13 )*cos(el)
3. RxB_X = (+0.039 + 0.008)*T_leg     temperature range +3 to 14

The RxB_X trends are essentially identical to those seen last month, while this month's RxA_X-vs-cos(el) relationship has a slope of -0.20+0.13: somewhat of a reversal from the trend tabulated last month.