SUMMARY
All-sky pointing and a tracking test (of the 's-correction') were carried out during grade 4 weather.

The tracking results possibly show a 2" step in the elevation pointing residuals, but located at about azimuth 195 - this is so far removed from transit to require some other explanation, or label. Alternately, the data may be interpreted as showing only rms noise levels of 0.9" in elevation. Other peculiarities in the data are discussed.
New all-sky pointing data (N=28) are disappointing, confirming recent serendipitous data. Correction for elevation motion is attempted but proves ineffective.

Transit Tracking
The 'transit' problem was noticed first in April . A correction was installed on 16 May, but checks since have invariably shown other peculiarities that question the efficacy of the s-correction, either in amplitude or location, and thereby leave uncertainty about the undelying cause of the poor elevation pointing performance around transit.

This latest test was done during conditions of poor transparency in place of scheduled observing. The phase (seeing) monitor was not working at the time. Mars was tracked for 90 minutes, using the SCUBA map16 method, with the EMPIRICAL correction activated :

Transit occurred at elevation 59.3 degrees. The raw residuals are 0.7" and 0.9" in azimuth and elevation, resp.. There is a possible 2" step in del, but at azimuth 195 ! This is too far from transit to warrant labelling as a 'transit' effect. There is also a similar change in azimuth performance at the same time, implying some other, global structural defect, or possibly an atmospheric effect.

The performance of the s-correction within 10 degrees of transit seems fine. Other explanations are needed for the other defects in the tracking seen here. The glitch at azimuth 172 in both coordinates, for instance, is not caused by a track joint : there are joints at azimuths 169 and 174 . . .

All-sky pointing
Recent all-sky pointing with SCUBA had shown rather poor performance in elevation seemingly uncorrelated with temperature or transit effects. The new data (N=28) are similarly disappointing. The 7-parameter pointing model installed on 21 November had performed well enough until the recent FTS run and one wonders if the realignment process has impacted the pointing subsequently. 28 data were collected, mainly all before the above transit experiment. The raw residuals are shown below :

The azimuth performance is nominal, although a systematic trend with elevation seems possible. The elevation performance is decidely poor. A systematic trend of del with azimuth is apparent although the paucity of data in the 5^th quadrant leaves its form unclear.

A much larger dataset taken during the course of regular observing in the past week is shown here, and reveals a possible sinusoid in the plot of dev-vs-azimuth. However, it is possibly confused by a variation as a function of time (temperature ?), with an additionally confusing break at shift change. Since the observers during the week have possibly used the same pointing sources at the same time each night such data may be subject to selection effects. The usefulness of this larger dataset to effect a model change is therefore questionable and we concentrate for the moment on using the dedicated data of the 26^th.

Despite the generally satisfactory tracking through transit seen above, the underlying problem causing the 'transit' effect and the need for the 's-correction' is the reversal of direction of elevation motion and the subsequent presumed hysteresis in the elevation bearing/drive. Thus, each slew of the antenna from one source to another may provoke a reversal and therefore a pointing error. However, this ought then to be corrected by the 's-correction'.

The 27 slews involved here comprise 8 UXU events, 6 UXD events, 5 DXU events and 8 DXD events, where the 3 letters characterise the elevation motion before, during and after the slew, U is up, D is down, and X is either. The elevation residuals in each group showed mean values of -0.1", +3.8", -4.1" and +0.6", respectively, with s.e.m.'s of about 1.2" in each case. Those slews not creating a change in elevation track direction impact the pointing negligibly. The amplitude of the impact (+ or - 4") from those slews that do change the tracking direction is essentially that of uncorrected pointing data, as described, for instance, on 08 Apr. The sample sizes here are statistically small, and correcting the handful of relevant UXD- and DXU- data for these effects does not improve the elevation residual rms.

The data were run through the TPOINT program and the 7-parameter pointing model updated. This new model was installed at 14:20 HST on 26 Dec 2000. RMS scatters in each coordinate are expected to be 2.0" (not optimal), and caveats regarding use of the new model include :

• the unknown performance in the 5^th quadrant, and
• the extrapolation of the new model to elevations above 70^o.

29 Dec 2000
Data the following night showed a deficiency at high elevations and a further correction was made - an overcorrection, as it happened - and on the basis of 18 data taken on UT20001229, a second adjustment was made that ought to provide good performance (rms of 1.7" in each coordinate) from now on.