SUMMARY
Nighttime pointing in good conditions with the new TCS has allowed some corrections to be made to the various TCS pointing models. Prospects look bright for the progress to completion of this project.

A diary of the events of the 1st shift of 29 Nov 1999 (UT 19991130) follows.

17:30  Firmin & me doing TCS tests - no Kimberly. Sky blue.
 
       Hardware changes made to accommodate new TCS

       TPOINT model setup using parameters from radio_telescope.dat :
       AN   =  #1 =   10.3  
       AW   = -#2 =  -17.1
       NPAE =  #3 =   58.8
       CA   = -#4 = -169.8
       TF   = -#7 =   12.6
       IA   =          0.0 expected   
       IE   =         93.7 following apparent usefulness of this value
                           last time, although we must be prepared to 
                           find it close to zero tonight if this is an
                           incorrect assumption.

       Latest track model also loaded.
18:10  tcs software loaded   tau_cso 0.078 seeing 0.8"
       RxB set to last night's last freq : 356.73425 Vlsr=-24 LSB dual SSB
18:20  slew to Mars - smooth. point #0001  (uaz,uel) = (12.9,1.2)
       focus (dX,dY,dZ) = (0.21,-0.16,-0.75) repoint   (14.5,3.3)
                    dZ                -0.77
18:40  Mars el=35  bmsw sample sig 16.3  S/N 2000
                   nosw R beam      8.15      600
                   nosw L beam     -8.15     -675 
18:45  Uranus (az,el) = (219,43) point     (uaz,uel) = (13.6, 0.0)
       Mars             (226,33)                       (14.3, 3.7)
       g34.3            (262,25)                       (14.3, 6.7)
       g45.1            (270,32)                       (12.1, 3.6)
       w75n             (314,50)                       (11.6, 0.6)
       ngc7027          (316,54)                       (11.1, 3.0)
       bllac            (327,61)                       ( 7.9, 0.8)
       n7538irs1        (356,48)                       (11.7, 1.0)
       g34.3            (265,17) one xtra point logged (14.2,12.0)
       Mars             (234,25)                       (14.8, 1.1)
       Saturn           ( 93,50)   20" diam            (11.6,-4.8)
       ......           ( 94,52) after focus checks    (10.0,-5.1)
       L1551-irs5       ( 80,29)                       (13.3,-1.0)
20:00  crl618           ( 59,30) tau_cso = 0.041       (13.8,-4.0)
       gl490            ( 29,39) seeing < 0.5"         (12.8,-7.7)
       w3(oh)           ( 19,43)                       (12.9,-7.7)
       L1551-irs5       ( 80,33)                       (13.8,-4.3)
       0420-014         (104,31)                        no result
       n2071ir          ( 95,15)                       (11.6,11.5)
       Saturn           (100,63)                       ( 8.5,-2.5)
       Uranus           (241,22)                       (11.2, 5.5)
       n7538irs1        (341,44)                       (11.1,-2.9)
21:00  w3(oh)           (373,46)                       (11.5,-8.3)
            carousel not following properly in 5th quadrant
       crl618           (419,43)                       (12.6,-5.9)
       Uranus           (245,15)                       (11.5,13.8)
       Saturn           (111,73)                       ( 8.4, 0.0)
21:26  logger_close - the raw data are shown below :

Note the strong systematic of del-vs-el (centre plot) and the weak systematic of daz-vs-el (upper center). Data are analysed with FIT9 and TPOINT .

22:00  analysis complete.
       raw data N=29   (daz,del) = (12.2 +- 1.9, 1.0 +- 5.9)
       Solution from FIT9 : 
       delta   p1   p2    p3    p4    p5    p6    p7  
             -1.7  -2.6  1.7  -3.0  10.4  -7.1  10.7  
       giving fitted rms's of (1.8,4.2) - a quad sum of 4.6"
       Solution from TPOINT :
       delta   AN   AW   NPAE   CA    IA    IE    TF
             -1.7   2.6 -4.3   2.6  12.1   7.1  10.7
       giving a total rms fit of 4.6" 

       Allowing for the sign changes described below, and for the
       artifactual changes in parameters 3,4,5 (NPAE, CA, IA), 
       the correspondence between the fits is quite good.

The result of the FIT9 analysis is shown below :

Note that the weak systematic has been corrected by a change to one or more of the pointing model parameters, but that the strong systematic remains.

  
       Also, sign differences are expected for p2/AW (seen), p4/CA (seen),
       and p7/TF (not seen), but not for p6/IE (seen)

       A fairly strong daz-vs-el systematic is removed by the fits,
       but the strong del-vs-el systematic does not disaappear.
       This suggests that either there was extreme anomalous refraction,
       or that the refraction model within the new TCS is incorrect.
       However, the total absence of a refraction correction would
       generate elevation errors of ~140" at el=15 degrees . . .
       which we do not see.

22:45  The TCS uses SLA_LIB refraction (SLA_ATMDSP). Firmin worries then
       that the parameters are not being fed into SLA_LIB properly -
         - the temperature is going in OK as degK
         - the humidity was going in as a fraction in (0,1), but
         - the pressure was going in as mmHg rather than mbar :
           eg 464 instead of 628.5(mbar).

       To first order the pressure enters the calculation of the
       refraction 'constant' A as 0.371*p where p is the % pressure
       difference from 624mbar. The value of p generated by this mistake
       is -25.6 (%) rather than +0.6 (%), leading to an error in A
       of 9.7. The refraction goes as A*tan(ZD), so for data lying
       between ZDs of 20 and 75, errors in elevation pointing of
       between 3" and 36" may be expected : a span of 33". This is 
       about 1.5x the span observed, but is in the right ballpark.
       
       Firmin changes the refraction code for mbar
       new logger file opened & pointing resumes :
22:10  Saturn (245,75) point & focus (uaz,uel) = (3.4,-0.5) !! much better
       n2071ir   (121,56)  (9.5,-13.5)
       crl618    ( 24,72)  (3.1, -6.2)
       oj287     ( 77,27)  (9.2,-18.8)
       n2071ir   (136,64)  (9.5, -8.8) 
       saturn    (262,60)  (6.6, -7.2)
       w3(oh)    (339,41) (12.0,-10.5)
       n7538irs1 (330,19)  (8.8,-19.9)
       crl618    (348,73)  (1.8, -1.0)
       Saturn    (265,53)  (8.3, -9.8)
       w3(oh)    (335,38)  (12.1,-11.1)
01:10  logger_close - raw data shown below :

Note that the systematic errors are completely different from those above - they also look a lot more amenable to solution within FIT9 or TPOINT . The FIT9 solution is shown below :

and the suggested changes to the models are as follows :

       Solution from FIT9 : 
       delta   p1   p2    p3    p4    p5    p6    p7  
              1.3  -1.5  32.0  38.7  38.1  3.4  -23.9
       with fitted rms's in (daz,del) of (1.6",1.7") - a quadratic sum of
       2.3".
       Solution from TPOINT :
       delta   AN   AW   NPAE   CA    IA    IE     TF
              1.0  1.9   15.6 -22.0  26.0  -3.2  -23.6
       The solution leaves a total rms of 2.0", and so the TPOINT
       model we should start with next time should be
              AN =   11.3  
              AW =  -15.2  
            NPAE =   74.4  
              CA = -191.8  
              IA =   26.0  
              IE =   90.5
              TF =  -11.0  
       
01:10  A DAS 500 spectrum of w3(oh) (#0077) at 12CO:3-2 gave a 
       peak of 13K in each polarity cf 25K expected !