Next: Acknowledgements Up: No Title Previous: Assessment of Track

Discussion

Can answers be given to our questions in Section 4 regarding our methods and assumptions ?

Our Inclinometry-with-Shims (IWS) experiments showed that, within observational errors, up-down motions of the antenna wheels yielded A-frame tilts as expected from rigid geometry.
Our IWS experiments also showed that, given the measured A-frame tilts, the amounts of PITCH and ROLL generated were 0.91 that expected assuming rigidity, but with an error large enough to encompass a factor of 1.0.
Production of ROLL from A-frame data proved intractable. A comparison with the ROLL measured on the TMU was not possible and so the validity of our method for producing YAW from the A-frame data was not confirmed.
The No-Track pointing experiments show that on average the pointing corrections generated by track model are correct, but that, under certain circumstances, the model may be as much as 20% from this optimal performance. The factors determining this variation must be identified and appropriate corrections to the model made.
Systematic errors in the inclinometry data, 1 to 5 quadrant effects, and sudden changes in zero-point, may all be due to track irregularities, providing further justification for the Azimuth Track Improvement Project.
Desired pointing performance of 1 rms in each coordinate requires track errors be reduced from their current value of between 0.5 and 1.0 rms to <0.4.

Many of the assumptions that go into the production of the model from inclinometry data are supported by the experiments we have performed during 1995, while others have been challenged by the data. Our improved inclinometry system reveals a high degree of similarity between various measures of tilts (eg between TY & RY, and between TX, RX & LX, Fig.2), that, on the one hand, support the assumption of rigidity implicit in our modelling algorithms, yet challenge the assumption of independence of motion of the A-frames. The relationships between wheel motion and A-frame tilt, and A-frame tilt and PITCH & ROLL support the quantitative consequences of the assumption of rigidity, while the apparent variability of the track profile and the inapplicability of a single track model at all times gives cause for concern unless we can correlate the variation with some other parameter (elapsed time, temperature etc).

Much is still on-going : the failure to extract ROLL from A-frame inclinometry data prevents a decision on whether the current location of the inclinometers on the bottom beams of the A-frames is sensible or not. The strong correlation between RY and TY (Fig.2b) strongly supports the current locations as being valid, yet the steps (Fig.3) present in A-frame data, but absent in TMU data, suggest a change may be expedient. Plans for the near future include moving the inclinometers from the A-frames to the elevation bearings for these reasons. We may then require a new method for calculating YAW; one that avoids the possible flaws in our current method.

A desired global pointing performance of 1

rms in each coordinate requires an improvement in the track model performance from a current value of between 0.5

and 1.0

rms to <0.4

. Achieving this may be done both by improving the quality of the track, by the removal or suppression of the track joints, which will prevent or reduce the shocks that might be the cause of systematic errors in inclinometry data, and by identifying and quantifying the causes for the inapplicability of the track model so generated at other times or under other circumstances.

At the time of writing (October 1995), the Azimuth Track Improvement Project is about to commence, with projected completion of the engineering in January 1996. Current aims of the engineering, based on these recent experiments and analyses, are to suppress the effects of the joints in the track. This will reduce the peak-to-peak errors in pointing, globally and locally, and remove a source of possible variable error, viz the dynamic component of the joint effect. The additional benefit may be the removal of sources of systematic effects upon the inclinometry data, and hence an improvement in the global rms quality of the models.

A more complete specification of the desired pointing performance, which will include a global rms scatter of 1

in each coordinate, is the subject of work in progress, and will (must) be completed soon.

Next: Acknowledgements Up: No Title Previous: Assessment of Track

Iain Coulson
Wed Mar 12 15:06:27 HST 1997