SUMMARY
Uranus was tracked for a couple of hours to check the pointing and track models. A change to the generation of the F1 term of the track model, from 1.21 to approximately 1.0, is suggested.

Uranus was tracked with SCUBA, using the map16 procedure, on a night when tau was too high (> 0.15) to do the scheduled observing :

• from (az,el) = (134,39) at HST 18:52 to (177,53) at 21:08.

The data were taken at a time when the H6 pixel, in the inner ring of SCUBA's LW array, was very noisy, but it does not seem to have biased the centroid determinations. The resulting pointing residuals are shown below left :

• The raw rms scatters in (daz,del) are (0.8",0.4"). The seeing during this period was approximately 0.5" throughout.

• The excellent results in del suggest that the F3 term in the model is well determined from inclinometry (it is, after all, the easiest of the 3 terms to calculate, being simply the TMU-Y readings appropriately scaled).

• There are fairly obvious systematic errors in the azimuth residual, suggesting an error in the F1 term ( yaw) of the track model. The figure above right shows the behaviour of the F1 term in this azimuth range, together with the wheel-joint interactions, shown as blue lines. There are wheel-joint interactions at these azimuths :

            Azimuth   Wheel   Joint 
                               
            139.09      4      7/8  
            142.09      2     14/1  
            147.29      3      4/5  
            150.29      1     11/12 
            166.09      4      8/9  
            169.09      2      1/2  
            174.29      3      5/6  
            177.29      1     12/13 
  

  The first set of 4 correlate fairly obviously with the systematics seen in the azimuth residuals, although the data in the second half are not free of error. Let us divide the data into two halves, with mean (az,el) of (145,45) and (172,52). When the amplitudes of the F1 excursions in each half, approx. 20" and 10", respectively, are multiplied by the cosine of the mean elevations, 0.71 and 0.62, the expected amplitudes of the uncorrected residuals are 14.2" and 6.2", respectively. The observed excursions ideally will be zero, but are measured here as 3.2" and 1.0", approximately, giving apparent scaling errors of 22% and 16%, respectively.

  Currently the A-frame inclinometry data are scaled by a factor 1.21, a change, made in April 1998, from the previous value of 0.83. We have been waiting for the correct circumstances to confirm this value. The last unfinished part of the Track Improvement Program has been this redetermination of the efficiency of the inclinometry system. It has, however, not been possible to perform the necessary experiment (tracking such as this) without having two repeatable inclinometry measures bracketing the tracking data, so that there could be confidence in the stability of the underlying track model. The required stability has only recently been acquired, due to the 3t loading of the front wheels in May 1999, and despite the reduction in frequency of inclinometry runs due to the carousel drive problems of the summer.

  If the inclinometry scheduled for Oct 06 shows little change from that obtained on 990830 , then we may indeed be able to use these data (and 4 other sets taken since May) in this way.

  However, it may be worthwhile to repeat the above experiment soon (tonight, 990924 ?) with a scaling factor of 1.0 in order to assess the likely impact of adopting such a value permanently.