|
|||
| Joint Astronomy Centre |
|
|
|
| Show document only |
|
||||||||||||||||||||||
|
____________________
|
||||||||||||||||||||||
|
Pointing: central bearing report 980705
July 05, 1998
In June 1998 it became clear that the updated corrections installed in April were not adequate. Pointing errors of up to 3 arcsecs (@ EL=0) were observed at the azimuth ranges associated with the central bearing problem. Moreover, the pointing errors appeared to show a 'sine-shape' similar to, albeit much smaller than, the shift introduced by the central bearing problem itself. Data AnalysisOn June 22 Scuba pointing observations were done on Uranus covering 2 of the affected azumith ranges. The data is shown in the figure below as the blue and cyan points. Note: all measurements in the figure have been corrected to 'horizon' (EL=0) values and have been folded into the same azimuth range. Analysis of the data was delayed because of software development needed to get a precise azimuth calibration of the shift. The dashed line shows the fit to the data which was installed at the telescope on July 2 at 4:29pm.Most noticeable about the fit is that it implies a 0.27 degrees slip in azimuth. This slip may well have been present in the April data as well, but those data were not as precisely analysed for this effect. The amplitude of the fit is 10% less than in April, suggesting that the 22.5% increase over the original value installed at that time probably was too large. 
After the installation of the solution on July 2 more tests were done that same and the following night. The data is shown in the figure as the magenta and red points. The red line at the bottom of the plot shows the actual residuals observed on 07/03 and indicates that the new fit in general was adequate except for a single excursion of -2 arcsecs at Az = 239 (modulo 360/22). However, the data show that shift in azimuth of 0.27 degrees has indeed resulted in a much better fit. The 'final' fit to the data is shown as the solid line which has been custom tailored to follow the 07/03 data mostly. This fit incorporates deviations from the standard double-gaussian used so far. The implied(!) reduction of the residuals is shown as the blue line at the bottom plot. This fit was installed at the telescope on July 4 (sic!) at 4:03pm. Unfortunately the next central bearing test on July 5 had a very large rms because of bad weather. Nevertheless, those data show no indication of any anomalous excursions. ConclusionsWhile not rigorously tested yet, the newly installed fit should reduce pointing errors resulting from the the central bearing problem to within than 1 arcsec at the horizon. The implied peak-to-peak amplitude of 11 arcsecs is close to, and perhaps somewhat less than, the value initially found last year. Naively one would assume the effect to widen in azimuth and to decrease in amplitude as edges get smoothed and slopes become more gentle over time. I see little indication that the effect is getting "worse" as has been feared to date from the amplitude versus time plot. However, I can not explain that plot except to note that re-balancing of the telescope has taken place on several occassions and that small errors in the perceived amplitude tend to become emphasized up by the division by cos(EL). Interestingly inbetween the 06/22 and later observations the telescope was in fact re-balanced and the data clearly hint at a change in amplitude of the "positive" peak.At present the slip of the phenomenon by 0.27 degrees in azimuth remains unexplained, although in principle there is nothing that could prevent such slippage from happening. I would like to emphasize that the above curves almost all have an underlying "baseline" subtracted. Determining this baseline currently is the limiting factor for the accuracy of the fit. The "high" amplitude of the positive peak of the 06/22 data could have resulted from an incorrect baseline, which for those data consisted of the sum of a ramp plus sine function. From the plot it can also be seen that very little baseline data is being taken. Moreover, quite a few observations are start too late. This is a serious problem which hampers the analysis. SuggestionsWhile the amplitude of the problem may not be a major concern anymore, the azimuth slip should be. I strongly suggest that regular monitoring (i.e. on a monthly basis) with Scuba remains in effect. I strongly urge that the phenomenon is tracked sufficiently in azimuth to enable a fit of any baseline and, if possible, that at least two of the affected az ranges are covered. This requires 2.5 hours of observing time. Daytime can be used provided that conditions are quite reasonable in terms seeing (probably < 1").Dream mode dataFor completeness, the figure below shows some central bearing data obtained using the Dream data mode. The fits should be ignored(!) since there appears to be a substantial non-linear baseline but not enough range for a fit. Nevertheless the number of data points (about 2500 over the whole phenomenon) should make very detailed monitoring and analysis possible.
Remo Tilanus Last modified: Mon Jul 6 13:55:03 HST 1998 |