Thermal
-------    
                                                                  
  FE  No.of data   fit                              Temp range
  --  ----------  ----------------------------     -----------
   S      23      del = (-1.88 +- 1.04)*T_diff     -0.7 to +0.7
                        (-0.87 +- 0.28)*T_mean       4  to   8

   A      61      del = (-1.81 +- 1.08)*T_diff     -0.6 to +0.6
                        (+0.04 +- 0.14)*T_mean       2  to  10

   B     166      del = (-0.18 +- 0.56)*T_diff     -0.2 to +1.3
                        (+0.16 +- 0.09)*T_mean      -2  to  10

   W      43      del = (+2.30 +- 1.73)*T_diff     -0.6 to +0.4
                        (+0.44 +- 0.13)*T_mean       0  to  11

Only the SCUBA and RxW relationships with T_mean are significant 
(>3sigma, and only just, and based on rather few data). Overall there's
nothing significant and consistent across the FEs to justify
changing the value of mean_temp_slope.

The RxA and RxB data show no concerns about the applicability of the current pointing model (that from 15 December 2004).

Focus
-----              
      FE  axis  No.obs.    mean   s.d.  Notes ('significant' trends)
                           (mm)   (mm)
                                                                        
    SCUBA   Z       8     -0.14   0.06    
            X       3     +0.25   0.29
            Y       3     -0.12   0.22

     RxA    Z      29     -0.13   0.10      1
            X      23     +0.09   0.20
            Y      23     +0.04   0.16

     RxB    Z      63     -0.11   0.11      2
            X      48     +0.04   0.21
            Y      50     +0.14   0.18

     RxW_C  Z      15     +0.09   0.08      3
            X      17     -0.28   0.28
            Y      18     +0.29   0.19

     RxW_D  Z       5     -0.06   0.14
            X       7     -0.18   0.40
            Y       7     +0.06   0.21

                                                    Temp range
                                                   -----------
Notes : 1 : RxA_Z   = (+0.018 +- 0.007)*T_leg         0 to 10 
        2 : RxB_Z   = (+0.022 +- 0.004)*T_leg        -2 to 10
        3 : RxW_C_Z = (+0.017 +- 0.004)*T_leg         0 to 12

The elevation dependencies seen last time have apparently been corrected successfully, but the Z dependency upon leg temperature is too significant to ignore. The SMU parameter DZT_COEFF has therefore been changed (back) from 0.050 to 0.070 (mm/deg; it had this value between Oct 2003 and Aug 2004). There may some minor impact upon the zero-points of these relations : to be adjusted next time if necessary.