cross scans with 430 MHz yagi
Note: 29may06 updated results with corrected cal value: 124.9
K instead of 42.1 K
The 430 Mhz yagi was lowered on 23dec05 by .41 m
to put the feed 3 meters below the paraxial surface. The system was recalibrated
on 12feb06 using 3C48 (37.5 Jy at 430Mhz). 10 cross scans were done on
the 430 yagi and two cross cans on the linefeed. The previous calibration
(before the move) can be found at 06nov05
calibration. The setup was:
The plots show the
system performance (.ps) (.pdf):
1.5 degree strips (10 crosses on the 430 yagi and 2 crosses with the linefeed)
The yagi feed adds the polarizations together. The lrr1 polarization of
the linefeed was used for the 430 linefeed measurements.
The signal was centered at 430 Mhz and was passed through a 1 Mhz butterworth
IF filter. Each strip took 90 seconds. The signal was detected with a .02
sec time constant and then sampled at 100 hz.
At the beginning of a set of crosses, the cal was fired for two seconds
to measure Tsys. The cal values used were: yagi cal:124.9K, linefeed cal:
2d gaussians were fit to each cross. The fit included a rotation angle
for the ellipse as well as a linear term in za (for changes in tsys). The
first sidelobes were excluded from the fit.
The flux used for the source was 37.5 Jy.
There was a problem with the yagi. There were glitches in the tsys that
lasted for a few seconds. The glitches were random. They were fixed by
cleaning out the water in the hybrid that converted linears to circulars.
Fig 1 shows the individual strips:
Top: yagi strips on 3C48. Black are azimuth strips, red are za strips.
You can see the jumps in the values that occurred randomly. There were
a total of 10 crosses taken in the sequence: cal, 4 crosses, cal, 4 crosses,
cal 2 crosses. I've plotted the first cross after each cal (the other 3
in each set were the same as the first.). The increase in Tsys is probably
real and not a problem with the cal (since the sefd also sees this change).
Bottom: line feed strips.
Fig 2: The average of the yagi strips after removing the system temperature.
Each strip was rotated by the measured pointing error before averaging.
The az strip is black and the za strip is red. The top plot is linear (in
degK). The bottom plot is db below the peak. The za beam width is
a little larger than the az beam width. The az sidelobes are 10 db and
28 Amin from the peak.
Fig 3 has the Gain and Tsys. The yagi Tsys measurements jump every
4 points (about 5%). They probably correspond to errors in the cal measurement
since the cal was fired every 4 points. The sefd should not show
this jump since it does not use the cal.
Fig 4 : sefd and beam widths. The top two plots are the sefd (source flux
needed to equal the system temperature). The bottom two plots are the beam
widths from the fits. * is the azimuth beam width and + is the za beam
width. The yagi has a beam width of 18 amin and the linefeed has a beam
width a little less than 10 amin.
Fig 5: The pointing error measured from the fits. The values are in great
circle arcminutes. The top plot is the yagi and the bottom plot is the
linefeed. The ch model was used for pointing with great circle offsets
of: az: .4833 deg, za: -1.540 deg. These pointing errors say the offsets
should have been: az: .500, za: -1.560
Note that the gain, tsys, sefd are at za=15 while the
nov05 data were at za=13. The ratio of sefd went from 40 to 30 so the move
improved the yagi G/T by about 1 db.
|az pnt err (avg)
zapnt err (avg)
|az pnt offset used:
ch pnt offset used
| 0.4833 deg
|az pnt offset measured
za pnt offset measured