|
20061118 report
HARP pointing
The first large dataset of pointing taken with HARP was acquired
on UT20061118. Opacity was decent (tau225GHz ~0.10) and stable.
The data are displayed below
after the removal of just collimation offsets:
Click for better view
Adjusting only the collimation terms (CA,IE) - by (+5.3", +50.6"),
in addition to the
pointing_offset CA="18.0" IE="28.0"
already installed - still left RMS residuals on the sky
(dS,dZ) of (10.5",4.8"). These are, as expected, minimally improved
( to (10.2",4.0")) by allowing optimiziation of all 7 parameters in the
model.
The image rotation angles, which are twice the K-mirror position, are
available in the file (eg)
/jac_logs/20061116/TCS_20061115-180718/JCMT_LOG.log
where the entries run into 2 lines and contain the following (reorganized
for viewing):
UT mm ss.ss seconds MJD az-demand (deg) az_actual diff(")
04 08 13.762 14893.762 54055.172381 345.024328 345.027199 10.34
el_demand (deg) el_actual diff(")
72.109739 72.112507 9.96
? ? ? ? ? ? ? img_rotn_demand _actual
-0.1 17187 -0.117187 0.000000 0 1 -0.238486 -17.540827 -17.443102
I extracted these angles for the above dataset and optimized (minimized residuals)
using the formulation of Per Friberg of Dec 2005:
-dAz = x0*sin(2K) + y0*(1+cos(2K)) - x0'*cos(E-2K) - y0'*sin(E-2K)
-dEl = x0*(1-cos(2K)) + y0*sin(2K) + x0'*sin(E-2K) - y0'*cos(E-2K)
where E is elevation and 2K the image rotation angle.
x0, y0 are the (constant) offsets of the K-mirror axis from the elevation axis,
and x' y' are the offsets of the HARP array center from the K-mirror axis.
Optimizing the residuals and rejecting 7 of the 37 original
data as lying >3sig from the means gave optimal
(x0,y0,x',y') of (-8", -10", 4", -16")
with rms in (uaz,uel) of (3.1", 4.0").
20061123
27 data logged, all with recorded HARP-array offsets of (17.37, -15,60);
presumably this is a tweak on H10's default (15,-15).
There were many quick repeats; probably leaving 18 useable independents.
One datum (L1551) was deviant and removed prior to this basic analysis.
The data need/suggest collimation changes of (CA,IE) = (+10.6",+49.1"),
apparently in addition to the (18.0",28.0) already installed in
/jac_sw/itsroot/src/harpb/XML/instrument_harpb.ent
The 18 data then have rms's in TPOINT's (dS,dZ) (like az,el !) of
(16.3", 6.6"), cf (10.5", 4.8") measured from ~30 sources on 20061118.
This supports Jane's observation that some degradation has occurred.
Bear in mind that many parameters are still missing from the TPOINT model:
who's to say at this stage that these new data wouldn't be fit better by
the full model than those of 20061118 !
20061124
16 (independent) data logged, all with recorded HARP-array offsets of
(17.38, -15,60). The data need/suggest (additional) collimation changes
of (CA,IE) = (-3.7",+52.7") over and above the (18.0",28.0) already
installed in
/jac_sw/itsroot/src/harpb/XML/instrument_harpb.ent
The 16 data then have rms's in (dS,dZ) of (6.6", 3.9"), but
with two measures substantially deviant (both of N7538irs1 -
which might indicate a morphological problem with using this as a
pointing source). The remaining 14 measures are surprisingly well
fit by the current model of 7 parameters : by changing the
parameters as follows:
coeff change value sigma
1 IA +21.590 -240.37 16.519
2 IE +43.836 +24.49 3.299
3 NPAE +50.206 +72.47 20.921
4 CA -51.693 -209.65 25.189
5 AN +0.902 +9.26 1.497
6 AW +0.870 -14.28 1.392
7 TF -13.633 -17.89 5.480
and leaving residuals in (dS,dZ) of (2.7",2.4") :
Click for better view
The small suggested changes in AW,AN are as expected, and the
correlation between IA, CA and NPAE often results in
suggested changes to all three terms despite a lack of
physical justification. IE & TF are also often correlated
if there is poor elevation coverage (not particularly
the case here - see plot).
However, we should also not expect that the RxA/Cassegrain
model should remain perfectly intact when tested at Nasmyth,
so it would be interesting to test the HARP-specific,
7-parameter pointing model as quantified above.
|
