Pointing errors oct,nov10
22nov10
The U2-U2L3
beam repair took place between 11sep10 and 13oct10. This included
adding weight to each corner of the triangle (where the new beams were
installed). Calibration data from 13oct10 thru 16nov10 were used to
check the pointing error after the U2-U2L3 repair. This data was mainly
sband narrow (sbn) and some cband. On 18nov10 the repair of
U3-L3 began. This will add more weight to the corners and may change
these results.
Notes:
- These change were not installed in the pointing model
- On 14oct10 the td12 encoder was changed. This could have
generated a 1az error term.
The plots show the pointing errors and
fits for the 1,2,3 azimuth error terms (.ps) (.pdf):
- Page 1: az,za coverage of data
- top: all the data
- middle sbn coverage
- bottom: cband
- The telescope was limited to za<16 deg za for this data.
- Page 2: pointing errors, sband,cband over plotted
- black is sband, red is cband
- 1,2: za error vs za,az
- 3,4: az error vs za,az
- You can see the 1az term in the za and az errors vs az.
- errors:
- za:mean:1.7 rms:4.26
- az:mean:-.89 rms:6.15
- Page 3: pointing errors sbn by source
- Same as page 2 but limited to sbn. Plotted by source
- errors:
- za:mean:1.39 rms:4.86
- az:mean:1.33 rms:5.83
- Page 4: pointing errors cband by source
- errors:
- za:mean:1.86 rms:3.90
- az:mean:-2.08 rms:5.99
- Page 5: 1,2,3 az fits to errors (all the data)
- top: za errors vs as with fit (red)
- 2nd: az errors vs az with fit (red)
- 3rd: az (black), za(green) residuals from fit vs az.
- 3th: az(black), za(green) residuals from fit vs za
- Page 5: pointing errors all data vs za with fit removed (by
source)
- 1,2: za errors vs za,az
- 3,4: az errors vs za,az
The az,za errors were separately fit to all the data using:
- y=C0 + C1*sin(az - C2) + C3*sin(2*az-C4) + C5*sin(3*az-C6).
Pointing errors
|
Azimuth
errors (asecs)
|
za
Errors(asecs)
|
points
|
|
mean
|
rms
|
mean
|
rms
|
|
All Data
|
-0.89
|
6.15
|
1.70
|
4.26
|
285
|
sbn
|
1.33
|
5.83
|
1.39
|
4.86
|
194
|
cband
|
-2.08
|
5.99
|
1.86
|
3.90
|
91
|
AllData - Fit
|
0.00
|
3.47
|
0.00
|
2.68
|
285
|
The results of the data fitting were:
|
AMPLITUDE (asecs)
|
PHASE (deg)
|
constant
|
|
|
sin(az-ph)
|
|
|
sin(2az-ph)
|
|
|
sin(3az-ph)
|
|
|
The 1az term for az errors and za errors both have
close to the same amplitude and their phase difference is about 90
degrees. This is a tilt in the platform.
The model uses cos() sin() rather than
amplitude,phase. The table below contains the values that should be
subtracted from the model to correct for these errors:
-
Values to SUBTRACT
from current model
Model Term
|
az Err
Asecs
|
za Err
Asecs
|
constant
|
0.649
|
0.911
|
cos(az)
|
-1.377
|
6.055
|
sin(az)
|
-5.440
|
-0.076
|
cos(2az)
|
0.349
|
0.967
|
sin(2az)
|
0.172
|
-0.135
|
cos(3az)
|
2.263
|
-0.075
|
sin(3az)
|
-0.048
|
-0.248
|
Summary:
- The pointing errors after the U2-U2L3 beam repair were
- az: mean:-.89 , rms:6.15 asecs
- za: mean: 1.70, rms:4.26 asecs
- Fitting 1,2,3 az terms showed that the largest error was a 1az
term which is a platform tilt.
- On 14oct10 the tiedown 12 encoder was changed.
- If the new encoder was not replaced with exactly the same
value as the old setting, then this could have caused a platform tilt.
- This tilt should have had a phase near 180,270 (az,za) which
is pretty close to what was measured.
- The individual sbn,cband errors differed probably because they
didn't sample a complete 1az cycle.
- The errors after fitting for 1,2,3 az were:
- az: rms: 3.47 asecs
- za: rms: 2.68
- This is probably better then expected since we did not have a
complete az,za coverage.
- The
pointing model was not corrected for these errors.
- The u3-l4 reinforcement will probably change these values.
- If the error was from the td12 encoder replacement then we
probably should install the changes.
processing: x101/101114/chkpntx101.pro
page up
home ~phil