[daip] Spillover and opacity corrections.
Craig Walker
cwalker at nrao.edu
Fri Sep 26 18:55:29 EDT 2014
In trying to understand the zero opacity results for my C band data, I
have noted that the spillover correction used in APCAL is probably not
correct at lower frequencies. The numbers in APCAL.FOR, subroutine
SPILL are those we use at 7mm in PTANAL, the pointing analysis program.
They are not what we use at lower frequencies. This is of some
importance because all of the gains distributed for the VLBA are for
opacity corrected data.
The simplest way to explain what is done in PTANAL, and what would
likely be better for APCAL, is to show the subroutine from PTANAL that
does it. It has the data statements, and the code that selects which to
use, and lots of comments about why it is doing things. It is attached.
Note that, for example, the 13cm spillover is very far from what it is
at other bands because of the dichroic plate.
With the higher spillover curve (the 7mm curve), the variation of Ts
with elevation in my C band data was entirely accounted for with
spillover and the fitted opacity was zero in many cases. That is not
correct as the atmosphere has a bit of opacity everywhere. See B
Butler's VLA Test Memo 232 for more about this sort of thing. He also
found that the VLBA 7mm spillover curve did not give good results at
lower bands.
Somewhat separate issue, but also with the opacity fits:
Something disturbing happened in the fits to my 6cm test data for the KP
site. The output from APCAL for that site is:
APCAL1: KP RCP 0/ 0h 0m Trec (K): 28.73 Zen. opac.: 0.003
APCAL1: KP LCP 0/ 0h 0m Trec (K): 24.96 Zen. opac.: 0.000
and the plot of the fits is in the attached postscript file. As can be
seen, the fit lines don't match the data at all. For R, the fit line is
above all of the points while for L it is below. It's not because it is
fitting a common curve because the R tsys values, let alone the model,
are above the L values.
APCAL was run within VLBACALA using the following commands on noatak:
default vlbacala
inn='TA035A_VLBA'; incl='MULTI'; ins=1; ind=3
dofit 1
vlbacala
Disk 3 is /home/noatak2/AIPS/DA01
Cheers,
Craig
--
---------------------------------------------------------------------
R. Craig Walker Array Operations Center
cwalker at nrao.edu National Radio Astronomy Observatory
Phone 575 835 7247 P. O. Box O
Fax 575 835 7027 Socorro NM 87801 USA
---------------------------------------------------------------------
-------------- next part --------------
REAL FUNCTION TSPILL( EL, FREQ )
C--------------------------------------------------------------------
C Function to calculate spillover.
C A full blown version should integrate the spillover pattern
C over the measured horizon and over the atmospheric emission.
C For now, just use a simple curve.
C
C These curves will have minima of 0. This ignores the fact
C that there is some spillover at all elevations. Just treat
C the minimum spillover contribution as part of the receiver
C temperature.
C
INTEGER MEL, MELD, MEL7, MEL13, MEL90, I, M
PARAMETER (MEL=10)
PARAMETER (MELD=5)
PARAMETER (MEL13=7)
PARAMETER (MEL90=2)
PARAMETER (MEL7=9)
REAL EL
REAL SELEV(MEL), SPILL(MEL)
REAL SELEVD(MELD), SPILLD(MELD)
REAL SELEV13(MEL13), SPILL13(MEL13)
REAL SELEV90(MEL90), SPILL90(MEL90)
REAL SELEV7(MEL7), SPILL7(MEL7)
REAL FREQ
C
C The default spillover is based on a Sept 11-12 1989 6cm Ts vs el
C curve that was especially clean so it seemed apparent what was
C spillover. It will not be quite right for other bands or
C other azimuths, but it should be better than nothing, I hope.
C
DATA SELEVD / 2., 14., 19., 30., 90. /
DATA SPILLD / 8., 7., 5., 0., 0. /
C
C The 13 cm spillover curve is based on early test data from PT
C with and without the dichroic. The plots are with the Jodrell
C meeting viewgraphs. A 2.7 deg. atmosphere was assumed and
C the 0 level was set by the "without" data (which match the
C default spillover reasonably well.
C
DATA SELEV13 / 0., 11., 14., 19., 30., 46., 90. /
DATA SPILL13 / 18., 19., 17., 15., 11., 7., 0. /
C
C Treat 90 and 50 cm as if there is no spillover (or rather
C that it is constant. Certainly the ground pickup around the
C subreflector at low elevation is not a problem at these
C bands.
C
DATA SELEV90 / 0., 90. /
DATA SPILL90 / 0., 0. /
C
C 7mm spillover from Vivek.
C
DATA SELEV7 / 2., 15., 20., 25., 30., 40., 50., 70., 90. /
DATA SPILL7 / 12.0, 11.0, 9.0, 6.5, 5.0, 2.0, 1.0, 0.0, 0.0 /
C
C--------------------------------------------------------------------
IF( FREQ .GT. 2000.0 .AND. FREQ .LT. 3000.0 ) THEN
C
C Have 13 cm.
C
M = MEL13
DO I = 1, M
SELEV(I) = SELEV13(I)
SPILL(I) = SPILL13(I)
END DO
ELSE IF( FREQ .LT. 1000.0 ) THEN
C
C 50 or 90 cm
C
M = MEL90
DO I = 1, M
SELEV(I) = SELEV90(I)
SPILL(I) = SPILL90(I)
END DO
ELSE IF ( FREQ .GT. 30000.0 .AND. FREQ. LT. 55000.0 ) THEN
C
C 7mm
C
M = MEL7
DO I = 1, M
SELEV(I) = SELEV7(I)
SPILL(I) = SPILL7(I)
END DO
ELSE
C
C Some other band.
C
M = MELD
DO I = 1, M
SELEV(I) = SELEVD(I)
SPILL(I) = SPILLD(I)
END DO
END IF
C
C
C Interpolate:
C
TSPILL = 0.0 ! default
DO I = 1, M-1
IF( EL .GT. SELEV(I) .AND. EL .LE. SELEV(I+1) ) THEN
TSPILL = ( SPILL(I) * ( SELEV(I+1) - EL ) +
1 SPILL(I+1) * ( EL - SELEV(I) ) ) /
2 ( SELEV(I+1) - SELEV(I) )
GO TO 100
END IF
END DO
100 CONTINUE
C
RETURN
END
-------------- next part --------------
A non-text attachment was scrubbed...
Name: TA035A_KP_OPACITY.PS
Type: application/postscript
Size: 25045 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/daip/attachments/20140926/de6a1b64/attachment-0001.ai>
More information about the Daip
mailing list