Date: Thu, 19 Dec 2002 11:46
From: Iain Coulson 
To: Donna DeLorm , Per Friberg 
Cc: Nick Rees 
Subject: Pointing report for Ops meeting

(Last week's comments by NPR in quotes ).

"Het. collimations will be added to SCUBA model"
- Limited pointings last night with Rxs A,B,W show good, good and fair 
  zeroing of offsets, respectively : del_W_D = 6", |others| <1".

- (Note added later) NPR's 'SCUBA' model is being used for all FEs.

"UKIRT TCS will be put into JCMT 
- see report by NPR elsewhere

"Thermal effects dominate possible flexure of the cabin (SCUBA vs Het)"
- This refers to the strong temporal variations in del seen 
  occasionally, which, on their own, suggest possible systematic
  thermal effects. The heating (insolation) of the front legs wrt to 
  the back legs is expected to generate a tilt backwards of the 
  antenna, which then requires correcting by tipping the antenna down 
  in elevation (-6" per deg of difference). We have a parameter in the 
  TEL coded to do this and mostly (for all but one night since 04 Dec, 
  for instance), the correlation between del and this temperature 
  difference is not significant. We also have the capacity to correct
  del for the mean temperature of the antenna legs (L), although we 
  admit to no rationale for this. Since 04 Dec the correlations between 
  del and L are as follows :

  fitting  del = a * L  + const    and noting the improvement, if any, 
                                   in rms and the sig-nificance of 
                                   the relationship :

     Date(UT)     a   +-     sig-ma  old(rms)new   temp range (C)   FE
     --------  ------------  ------  -----------   --------------   --
     20021205  -1.06  0.11    9.6    2.8"  1.5"     2.0 to 9.0      S
           06  -0.26    17    1.5    2.8   2.7     -1.0    8.0      S
           07  -2.08    86    2.4    4.9   4.4     -3.5    0.0      A
           08  -0.57    22    2.6    2.8   2.7     -3.0    3.0      S
           09  +0.93    26    3.6    2.4   2.1     -4.0    1.0      S
           10  -0.19    19    1.0    2.4   2.4     -5.0    3.0      S
           12  +0.79    41    1.9    2.4   2.3      1.5    5.5      S
           13  -0.66    49    1.3    2.1   2.0      5.0    8.0      S
           14  -0.79    15    5.3    2.2   1.6      3.0   10.0      S
           15  -1.08    24    4.5    2.5   2.1      5.0   10.0      S
           16  -0.70    24    2.9    2.3   2.0      4.0    8.0      S
           17  +0.86    34    2.5    3.5   2.9      4.0    9.0      S
           18  -1.06    22    4.8    4.2   3.1      1.0   10.0      S
           19  -1.28    28    4.6    4.2   2.8      1.0    9.0      S
     
  They are often (36%) very strong (>3sigma), but more often (53%) 
  negligible and sometimes (7%) of the opposite sign, so overall there 
  is no good explanation for the strong trends seen on any single night, 
  from which one must deduce that the correlation with mean antenna leg 
  temperature is something of a red herring.

New analysis
------------
From the table above one can see that the elevation residuals are still
(even since NPR's model adjustment of 13 Dec) quite poor, and azimuth 
residuals in the same period are (2.8, 1.8, 1.9, 2.8, 3.1, 2.3) : for 
the 165 SCUBA data since 04 Dec rms scatters in (daz,del) are 
(2.6",3.3"). These data show 
           del = (-0.83 +- 0.10)* mean_leg_temperature, 
correction for which reduces the rms scatter in del to 2.8".

The rms scatters can be improved to (2.4", 2.6") by optimizing the 7 
principal model parameters. No further improvement is obtained by 
including the Nasmyth offset parameters.

It is not possible to make these temperature corrections within TPOINT.
The FIT7 adjustments and their TPOINT equivalents are listed below :

  FIT7 parameter    :     1     2     3     4     5     6     7
  FIT7 adjustment   :   0.0  -0.7  -0.9   3.2   6.8  -1.4   2.2    
  TPOINT parameter  :    AN    AW  NPAE    CA    IA    IE    TF
  TPOINT adjustment :   0.0   0.7  -0.9   3.2   6.8   1.4   2.2

But . . .
----------
There's no point in implementing these changes if the temperature 
correction step is spurious.