Pointing/gain before and after tertiary platform installed.
A platform to hold the tertiary control electronics
was installed on the feed tower in the dome between the rotary floor and
the outside balcony. The floor weighs about 600 pounds. The control electronics
may weigh another 600 pounds when installed.
Two sources were tracked during the day before and
after the installation (18may01,24may01). This was to check that the added
weight did not wreck the pointing. The sources J0512+166 (3C138) and J0738+177
at 2380 Mhz using the sband narrow receiver and the pntmod pattern
(the same pattern used for the pointing model).
The figures (.ps)
the pointing errors, gain, and sefd. The black lines are before the platform
was installed while the red lines are after the installation. The dotted
lines show the source rising while the solid lines are for the source setting.
Figure 1 shows the zenith angle error versus zenith angle in arc seconds.
Figure 2 has the azimuth error versus zenith angle in arc seconds.
Figure 3 is the beam squint in the zenith angle direction. There is a slight
slope of about 1 arcsec for za 0 to 20 degrees for both sources.
Figure 4 is the beam squint in the azimuth direction. There is a constant
offset of about 1 arc second. According to lynn baker there is an inherent
beam squint in the azimuth direction because of the offset optics.
Figure 5 is the gain in Kelvins/Jy for 3C138 (top figure). The bottom figure
is the SEFD (system equivalent flux density) for 3C138. This measures
the strength of a source that is needed to equal the system temperature.
Figure 6 top shows the ratio of GainPolA/GainPolB. The 18may01 data (red)
is stable while the 24may01 (black) values move around by up to 5%. Both
days start at the same ratio (1.03). There is only one cal value
taken at the start of each source. It is then used for the entire track
to convert d/a counts to kelvins. We previously measured a variation in
the gain of 1% per deg F change in turret room temperature. If the ratio
variation is from temperature, then the effective temperature for polA,
polB electronics must have differed by up to 5 deg F. The 3% difference
in gain at the start of each run may be a 3% error in the relative values
of the pola/polb cals.
Figure 6 Bottom is the ratio SEFDpolA/SEFDpolB. It is stable for
both days. Since the SEFD is srcFlux/(srcDeflection/TsysDeflection) it
does not depend on the cal. Any gain variations will cancel out since they
occur in both the src and tsys deflections. If we assume the gain for the
two channels is the same, then the 5% difference for a 25 deg K Tsys means
a difference in system temperatures of 1.25 K between channel A and B.
This is about what we measure using the cal values for polA and polB so
the cal values we are using are pretty good.
The addition of the platform did not radically affect the pointing.
There are jumps of up to 15 arc seconds in azimuth but these were also
seen in mar01 when model12 was made (they may be related to the active
tiedown motion). The 3C138 measurements started around 12:50 pm so the
sun may have affected the gain/sefd results.
The gain of 9 K/Jy and a SEFD of 2.9 implies a Tsys of 26 K.
The pointing needs to be rechecked when the control electronics is
installed on the floor.
The variation in the gain ratio shows that cals need to be taken more
The sefd ratio matches pretty well with the Tsys values we get using
the cal values. So the cal values are correct.
processing: x101/010524/doit.pro, doplot.pro