'Transit' tracking errors
This problem was assigned the moniker 'Transit Error', but, more strictly, resulted from the change of direction of the elevation motion of the antenna no matter where the action occurred. The first indications were from transit tracking experiments in Dec 1999 and March 2000, collected together in the report of 08 Apr .

The problem revealed itself as a 4" change in elevation pointing during tracking through transit, although examination of all-sky pointing datasets reveals it to be pervasive. Analysis of all archived pointing data by Remo Tilanus, using the new database, reveals a steady deterioration over the past few years, with a similar degradation between 1989 and 1994 being corrected, possibly by the welding of the cone bars - see his report .

Other, subsequent measurements and input include :

• An FEA analysis by Ian Pain predicts 5" effects.

• Local pointing is enough to combat the effect (as it does for all known and unknown systematic pointing problems), although 'local' must now include 'on the same side of the meridian'.

• 24 Apr 2000 Problem also affects northern transits

• 16 May 2000
  The pointing correction file, tel_empirical.dat, is resurrected for the purpose of correcting for this effect in software. This 's-correction' is a sine function over an azimuth range of +8^o with a semi-amplitude of 1.85". The fit of this curve to the southern tracking data is shown here

• 30 Aug 2000
  Due to fairly poor observing conditions, subsequent tracking data has been inconclusive in assessing the efficacy of the 's-correction' - see the report of 25 Aug 2000, et op cit - and a definitive test, now per force with the new Telescope Control Software, is still pending.

• 14 March 2001 Amplitude now 7".

• 19 March 2001 - theory due to antenna imbalance and windup/hysteresis/friction in the elevation bearing. Note how the speed of the elevation is a function only of azimuth.

• 18/19 April 2001 Amplitude of effect now 9".
  Additionally, a 'spike' has appeared of amplitude ~4" and width 10" centred at az=185. Checks of the encoder mounting would be useful, but are thwarted by presence of SPIFI at the R.Nasmyth focus. The elevation encoder seems the likely culprit.

• 24 April 2001 - Elevation encoder replaced.
• 01 May 2001 - Transit step amplitude is elevation dependent
• 14 May 2001 - Summary so far - prepared for presentation to JCMT Board.
• 23 May 2001 corrective algorithm described
• 19 June 2001 corrective algorithm installed
• 21 July 2001 corrective algorithm works, at reduced amplitude.

• 03 Feb 2003
  Currently, the correction is in the form of a lookup table describing a +3" graduated step from east to west occurring within a half degree of azimuths 0, 180, and 360. The telescope control software multiplies this correction by cosine(elevation) before applying it to the elevation coordinate -- so it's v.small almost everywhere. It probably doesn't correct for all impacts of the problem but compensates for a general E-W tilt in the plane of the elevation pointing residuals.