20010314 report
Empirical correction updated
Grade 4 weather was used to test the empirical correction
/jcmt_sw/mtroot/src/tcs/params/thi_empirical.dat
for the first time since
26 Dec 2000. This time the correction was
disabled prior to the use of the
map16 method.
3c273 was tracked through transit and, after repointing, down
into the west :
- from (az,el) = (166,62) at HST 01:05 to (224,66) at 02:29
- from (az,el) = (230,64) at HST 02:40 to (265,20) at 06:00
Transit occurred at HST 01:22 at elevation 72 degrees. The seeing was
between 0.4" and 0.5" throughout, although some instability in the sky is
indicated by the reported variation in tau_cso between 0.12 and 0.16.
The pointing residuals in these two trackings are displayed below :
Residuals through transit
Through transit (left, above) the azimuth performance is marred by
the 'lump' peaking at azimuth=194, and the discontinuity at azimuth=218.
Each is of amplitude 3" and has no elevation counterpart.
There are track joints at 193, 195, 196, 198, 217, 220, 222 and 225,
which do appear to correlate with these defects - although the track
model was updated only 2 days ago.
Overall the rms scatter in azimuth is 1.2", which is still not bad
for 1.5 hours tracking.
The elevation residuals show the positive step as expected, but the
amplitude is 7", twice what it was previously, the location is at
azimuth=185, cf azimuth=183 in the current version of
the empirical correction, and the width,
previously estimated at +-10 degrees, is now +-15 degrees.
The green (sine) curve represents this new parameterization.
The new version of thi_empirical.dat was installed at 2pm on 14 Mar
2001,
with the assumption that the northern 'node' of the defect occurs at
azimuth=355.
Tracking through the South-West
The residuals for the second, longer, part of this tracking experiment
are shown above right. They are decidedly not constant through these 3.3
hours, but that they correlate so precisely with each other is quite
puzzling.
Atmospheric noise, especially at the lower elevations,
would be expected to affect both - perhaps elevation more than azimuth -
but not in a correlated way like this. Analysis is on-going.
Conclusions
-
The defect being corrected by the 'empirical correction' has changed its
character, becoming stronger both in amplitude and (azimuthal) width.
-
The empirical correction lookup table has been rewritten to this new
specification.
-
Pointing should be monitored closely over the next couple of days to
gauge the efficacy of the new correction. Specifically a transit-tracking
experiment such as that described above should be performed with the
correction left enabled.
-
The new correction and the old pointing model will not be compatible.
In particular, the TPOINT parameter AW, which
describes the E-W misalignment of the azimuth axis, is likely to be in
error. Pointing errors may be expected until the pointing model is
updated next, although, as usual, the use of local pointing will overcome
this effect.
15 March 2001
Last night's data showed a distinct curvature in daz-vs-el
(below left), which has been flattened by adjusting the mix of
TPOINT parameters IA, NPAE and CA - see below right.
The rms scatter in daz falls from 1.9" to 1.1".
Iain Coulson
Latest Update : 15 Mar 2001
|
