Following the adjustment to TEMP_MEAN_SLOPE on 01 September (see report), SCUBA pointing has been free of temperature effects, while that for RxA and RxB shows systematics - as expected, I suppose - with the mean temperature of the antenna. (All pointing data between UT 04 & 17 Sep):

  FE    No.data   del -vs- temperature                   Temp. range
 -----  -------   -------------------------------------  -----------
 SCUBA    183     del = (-0.03 +- 0.46) * (T_f - T_b)    -0.7 to +1.6
                  del = (-0.05 +- 0.11) * (T_f + T_b)/2    4  to  13

  RxA      50     del = ( 1.89 +- 0.73) * (T_f - T_b)    -0.6 to +0.7
                  del = (-0.47 +- 0.11) * (T_f + T_b)/2    6  to  15

  RxB      19     del = (-1.29 +- 0.72) * (T_f - T_b)    -0.5 to +1.4
                  del = (-0.50 +- 0.18) * (T_f + T_b)/2    8  to  15

T_f = mean temperature of 4 probes on the front legs, etc.

So, currently, heterodyne pointing may suffer approximately 1" of elevation pointing error for each 2 degrees of uncompensated temperature change - i.e. if you let the temperature change that much without doing another pointing.

Do we need a focus-station-dependent version of TEMP_MEAN_SLOPE !?