[evlatests] Troubles with amplitude scale and phase stability

[Chuck Kutz]

I checked the Tcal tables for antenna 21's X-band (sn03) and the Tcal values
are in the 2+ range. Was there an error reading the numbers into the data
base?

Chuck

-----Original Message-----
From: evlatests-bounces at nrao.edu [mailto:evlatests-bounces at nrao.edu] On
Behalf Of Rick Perley
Sent: Sunday, March 06, 2011 1:22 PM
To: evlatests at aoc.nrao.edu
Subject: [evlatests] Troubles with amplitude scale and phase stability

    The 'Herc A Demo Project' data from earlier this week is being 
reduced.  Two problems are immediately apparent from X-band calibration:
   
1) The implementation of the switched power system is intended to (and 
should!) convert our raw correlation products into physical units, in 
Janskys.  It's close, but there is an offset.  Vivek has reported 
privately that 3C84 test data taken at C and K bands were being 
overcorrected.  This is also the case at X-band, where I've used 3C286 
to establish that most antennas are being overcorrected by a factor 
between 25 and 30% (in power).  Nearly all antennas show the same factor 
on all subbands and both polarizations.    An overcorrection can occur 
from three different ways:
    a) A mis-scaling, caused perhaps by an errant coefficient, or a 
misunderstanding of the meaning of a monitor value
    b) An antenna efficiency better than that assumed (I think we're 
using 60% for X-band)
    c) A Tcal which is actually lower than those in the tabulated values. 

     Six antennas are notably different than all the others, and for 
these it's fairly clear that explanation (c) is the most likely origin:
    Antennas 9, 16, and 20 are all *undercorrected*.  Persual of the 
'CD' table shows these antennas have the lowest values of Tcal of all 
antennas.
    Antennas 21, 22, and 26 are all overcorrected by large factors -- at 
least twice the others.  Perusal of the 'CD' tables shows these have 
unexpectedly high values of Tcal (especially antenna 21, whose Tcals are 
listed as 9.9 K -- surely this can't be right!)
    All other antennas have nearly the same value of overcorrection, 
which makes expalanations (b) and (c) unlikely (unless there is a 
systematic error in measuring the Tcals). 

2) More serious is a phase 'stability' error seen on many antennas.  I 
put quotes around 'stability' because the phase stability *within any 
one scan* is fine.  What is not fine is a large step in phase from one 
scan to another.  For this run, I changed bands cyclically around X, C, 
S, and L, returning to the same band once every ~10 minutes.   When we 
returned to X-band, we find the phase has changed by one of three 
different amounts:  0, +/- 45, or +90 degrees.  Each antenna (with a 
single puzzling exception) lies in one of these three families.  The 
population divides this way:

    1) 'Zero' degree jumps:  antennas 1, 6, 8, 12, 18, 19, 22, 23, 26 
and 28.  (NB  Antenna 1 is the reference antenna)
    2) +/- 45 degrees:  3, 4, 7, 11, 14, 15, 16, 17, 20, 24, 25, and 
27.  All of these are +45 except for antenna 3, which is -45.
    3) +90 degrees:  Antennas 6 and 10. 

    The jump values are approximate, and may actually represent a 
temporal progression, since the calibrators were observed in a regular 
loop.  (Roughly, once every ten minutes). 

    The only exception is antenna 9, whose changes are not regular. 

    Antenans 2 and 13 were out of the array. 

    It is certain that these changes are not due to baselines -- all 
three calibrators (different RAs, different decs) show the same phase 
progression. 
    There is no relationship to antenna location. 

    Ken and I recall my noting a similar type of problem with the 'flux 
calibration run' last December.  I'll have to return to those data to 
establish whether the effect noted above was apparent then in the same 
way. 

    I have not yet been able to reduce the other bands of the Herc A 
data.   This will be done over the next few days.


_______________________________________________
evlatests mailing list
evlatests at listmgr.cv.nrao.edu
http://listmgr.cv.nrao.edu/mailman/listinfo/evlatests