New track and pointing models installed - and ripple measured

One regular CW inclinometry run was made Saturday afternoon (980411) - thanks Kimberly :

     dataset     Start time     Mean leg temperatures
                     HST        start   middle    end
   inc980411        13:30        2.5      3.6     4.4        
   inc980412        12:30       -0.2      1.0     2.0

The data was used to create a new model, using the yaw factors fL = -fR = 1.21 for reasons described in the report of 980408 . The difference in the track model parameter F1 (= yaw) between the model using fL = 1.21 and that using fL = 0.83 is shown here to include excursions with amplitudes reaching 5". However, even using the old values of fL, fR there would have been many notable differences between this new model and the previous one taken on 980224 . In particular, there are several 2"-4" spikes. The azimuths and corresponding wheel-joint interactions of these features are shown in the table below (wheel #1 is front right, #2 is rear right, #3 rear left, #4 front left) :

   azimuth    13    37    40    64   193   217   220   244
    wheel      2     4     2     4     4     2     4     2
    joint    9/10  3/4   10/11  4/5  9/10  3/4  10/11  4/5

[ We note that wheel #2 is the one most likely to be the source of the 0.5" ripple seen first on 980222 . The ripple is analysed further below. ]

This new model was installed in advance of the evening shift. The shift was dedicated to pointing, with the interval between 20:05 and 20:45 HST devoted to measures of the central bearing problem . The periods either side were used to collect enough allsky pointing to enable the improvement of the 7-parameter telescope model. 25 data were obtained before the central bearing work (data set A), and when a further 19 (set B) had been obtained after the c-b work the combined total should have been sufficient to accomplish the task. However, these two datasets seemed quite different in quality :

        dataset      HST         N       daz        del
                                      mean  rms   mean  rms
           A    18:15 - 20:00   25    -4.6  2.0   -1.9  2.3
           B    21:00 - 22:10   19    -2.6  1.3   -3.1  1.6
           C    22:35 - 00:50   31    -2.9  1.7   -2.7  1.1

and so a further dataset (C) was considered desirable. After much agonizing I decided that sets B and C were perhaps compatible and should form the basis of any update to the telescope model. The combined 50 data have rms scatters in (daz,del) of (1.6",1.4"). Running the data through the FIT7 program suggested that reasonably small changes to the parameters of the telescope model would give a model with an expected performance of (1.3",1.3"). This model was then also installed.

The pointing performance on subsequent nights is shown below :

 Collimation errors of JCMT/SCUBA from daily pointing logs
 
   1998 April    N     daz  +-       del  +-   typical seeing
                      mean scatter  mean scatter
 
       06       20    -4.3  2.7     -4.0  1.7       0.3"
       07       19    -2.8  2.2     -2.9  1.7       0.2
       08       16    -2.2  3.0     -4.1  1.8        -
       10       21    -3.8  2.8     -3.6  1.7       1.0

       11 - new models created and installed        0.4

       12       21    -3.9  2.6      0.1  2.5(*)    0.5
       13       24    -3.4  1.7     -0.8  1.6       0.6
       14       26    -2.0  2.0     -1.0  1.6       1.0
       15       20    -2.0  1.4     -1.1  1.7       0.3
       16       22    -2.1  1.8     -0.6  1.5       0.4
       17       26    -1.4  1.4     -1.7  1.5       0.2
       18       23    -1.6  1.8     -0.1  1.3       0.5

       (*) - strong thermal component to this elevation performance -
             correcting for thermal effect gives rms in del = 1.7

The data since April 11 suggest that the rms scatter in the daz pointing error may well reach the expected value of 1.3" under conditions of good seeing (0.1").

The last 7 listed datasets, without any thermal corrections, give a total of 162 data with rms scatters in (daz,del) of (1.8",1.7"). Adjusting the 7 parameters of the telescope model improves only del - and only marginally (to 1.6" rms) - while adjustments to fL = -fR in the creation of the track model do not improve either of these scatters. The adoption of fL = 1.21 +- 0.01 seems justified.

The ripple
One CW radial arm run, with the inclinometer on the radial arm to the right rear wheel, was made on Sunday (980412). This dataset still has 4 data channels in common with the 980411 data, and the rms scatters in the differences between the data in these channels - sigma-(TX,TY,LX,LY) = (0.20", 0.26", 0.20", 0.15") - indicate either great stability over this 24-hour period or insensitivity to the 2-3 degree change in ambient temperature.

A comparison of the tilt of the right A-frame with the measures of 971214 shows the ripple to be persistent, and with an amplitude of anywhere between 0.5" and 1.5" ( cf. the original estimate of about 0.8", as reported on 980223 . Extraction of a single representative amplitude using Fourier analysis is still desirable if quantitative monitoring of the defect is to be meaningful, and work continues in establishing appropriate coding.