Polarization test during VLBI experiment Europe46

Observations

Experiment Europe46 had special schedule. 3 additional scans were added to the end of the experiment. We chose two the strongest sources visible at the sky at that time: 3C345 and 3C454.3 . All stations except EFLSBERG and DSS65 which were unavailable for extended experiment made one scan 3C454.3 (349-114100) and 3C345 (349-115800). They observed maximal time: 752 seconds. Then station MEDICINA and CRIMEA switched off polarization from right circular polarization (RCP) which is standard polarization setup for the geo-VLBI experiments to left circular polarization (LCP). It was done automatically in MEDICINA and manually in CRIMEA. Operators at the station CRIMEA climbed up (at height 25 meters under strong wind and heavy rain!) to the receiver box and made necessary manipulations. Then all stations made again two scans: 3C345 (349-130000) and 3C454.3 (349-131700). Antenna motor failed in CRIMEA during the first scan, but operators succeeded in repairing the motor (for 5 minutes!!) and antenna made the second scan. WETTZELL ended observations too early and there is no data with WETTZELL for the last scan.

The purpose of the experiment was to evaluate the level of instrumental polarization. Cross correlation of RCP and LCP should be zero and if we had had perfect systems we would not have found fringes between antennas observing opposite polarization. Real antenna systems record power not only of the targeted circular polarization but also some power of opposite circular polarization. Therefore having observed the strongest sources the maximal scanning time we could expect obtaining week fringes. Having compared fringe amplitudes between scan RCP/RCP and RCP/LCP we are able to made judgment concerning the level of instrumental polarization.

Raw results

Many people afraid to make raw results of the experiment publically available before the final publication. Nevertheless we take the risk.

We had a sure fringe detection for all scans at all baselines except one observation. It is good (we have information :-) ) and at the same time is bad (antennas systems are not perfect :-( ). Observed SNR is presented in this table. Time here is a reference time tag. These data are gathered in another table. Time tag is the time of the start of the scan.

We see that there is strong signal in all stations which observed opposite polarization. SNR in cross-polarization scans is 0.07-0.20 of the SNR which we get in normal polarization setup.

Quality factor of many of cross-polarization scans is low: 2, especially at X band. FOURFIT disqualified these observations because residual phases had large scattering and fringe amplitudes in different channels had strong deviations.

Plots of residual phases for last 2 hours:
MATERA/MEDICINA X-band and S-band

MEDICINA/NOTO X-band and S-band

MEDICINA/TIGOWTZL X-band and S-band

CRIMEA/MATERA X-band and S-band

CRIMEA/MEDICINA X-band and S-band

Residual phases in different channel are shown by different colors. Phases are in radians, time is in hours elapsed from the start of the session.

Those who understands magic of fringe plots may find interesting to look at some fringe plots of the cross-polarization scans. (Warning: fringe plot is an ASCII file 128 symbols width and 112 symbols height).

plot_01 plot_02
plot_03 plot_04
plot_05 plot_06
plot_07 plot_08
plot_09 plot_10
plot_11 plot_12
plot_13 plot_14
plot_15 plot_16
plot_17 plot_18
plot_19 plot_20

Discussion

All stations have polarization leakage at the level 10-20% at both bands. Situation is especially bad with S-band of MATERA where SNR in cross-polarization scan was only 20-30% less than in normal polarization setup. It means that when antennas observe the same polarization: RCP/RCP power of cross correlated signal from LCP/LCP is at the level of 1-2%.

Change of polarization during experiment caused a jump in phase calibration phases what was predictable since signal traveled another path. But it caused also a substantial increase (in tens times!!) in residual phase scattering what was not predicted. Phase scattering was larger at X-band than at S-band. Fringe amplitude between adjacent channels were also not aligned.

We have too few data to judge about systematic shifts in values of group and phase delay for cross-polarized scans.

Comments

Any comments are as always welcome.

Back to Europe46 page.
Back to Web page "Study of systematic errors in VLBI caused by polarization impurity".
This report was prepared by Leonid Petrov petrov@kuestner.geod.uni-bonn.de

Last update: 08-MAR-99 12:52:15