sband narrow calibration.
On/off position switching on continuum sources was done at 2380 Mhz with
the sband narrow receiver and the correlator. The setup was:
Pol A, pol B are native circular.
One minute integration's on and then off source were followed by
a 10 second cal on/off in the off source position.
Try to track sources rise to set.
Record 1 25 Mhz band centered at RF frequency 2380 Mhz.
Use the total power information, cal values, and source flux to compute
tsys, gain, sefd.
Fit the data to a function of (az,za).
Compare the measured gain with the expected gain. Include losses
from pitch, roll, and focus.
The dates for the observations and the sources observed are:
22aug00 : J2250+143, B0116+082
24aug00 : J2316+040, J0129+236,J0318+164 (CTA21)
26aug00 : B0116+082, J0431+206
02sep00 : J1347+122
03sep00 : J1330+251
Cal values used during calibration runs. This is a single noise diode.
The plots have all the data over plotted with different symbols for each
source. Dotted lines are polA and dashed lines are polB. The figures
cal polA [K]
cal polB [K]
The average gain, Tsys, and avgGain/Tsys are plotted in a 2-D plot vs az,za.
Up is feed in the north and to the left is the feed in the west (when the
source is rising). The length of the arrow is an absolute measure of the
value. The angle the arrow makes with the vertical is a relative measure
of the strength. The rotation angle is an easy way to visualize relative
strength. Be careful though since it is calculated relative to the maximum
measured value. If the flux for a particular source is wrong, and it contains
the maximum gain, then the angles would have a bias.
SEFD vs za.
Tsys vs za. T
Gain vs za in K/Jy.
Gain (average) vs za. The average is (gainA+gainB)*.5. Since this receiver
has native circular polarization, linearly polarized sources should not
affect the polA, polB gains.
Cal scale factor vs za. This is (cal In Kelvins)/(Cal deflection in correlator
units). This should be a flat line for each run since the attenuators were
not changed (within a source). Variations are probably do to gain changes
in the if/lo.
Fractional gain difference defined as (gaina-gainb)/(gaina+gainb).
The table at the lower left averages the value over all the dish, za
0-5, 5-10, 10-15, and 15-20 degrees. The figures are:
Average gain [K/Jy]. 1 tick= 4 K/Jy. Horizontal is maximum and 180 deg
(down) is 1/2 of maximum.
Tsys [Kelvins]. 1 tick=10K. Up is minimum temperature and 180 deg
(down) is 2*TsysMin.
Gain/Tsys [Tsys/Jy]. 1 tick= .1 Tsys/Jy. up is maximum, down is 1/2
Fitting the data to f(az,za):
The data is fit to:
f(az,za)=c0 + c1*(za-10)+c2*(za-10)^2+c3*(za-10)^3+
c4*cos( az)+c5*sin( az)+c6*cos(2az)+c7*sin(2az)+
Tsys-fit These plots show the data - fit vs za for sefd
and Tsys (gain is done separately below). The (za-10) terms are for za
0 through 20 degrees (unlike some of the other fits that used (za-14) for
za > 14).
with fit, (data-fit) by:sample,za,az, PRFloss, and prfAzTerms.
These plots take a closer look at the average gain fit and how the azimuth
terms of the fit compare to the azimuth terms of the gain loss computed
from pitch,roll, and focus. The figures are:
SEFD - fit. Units are Jy/Tsys.
Tsys - fit. Units are deg K.
of SEFD,Tsys,Gain, and avgGain . These plots show the azimuth components
(1az,2az,3az) for the sefd, tsys, gain, and average gain. . PolA is solid
and polB is the dotted line. The coefficients for the fits are also printed
on the plot. PolA on the top and polB on the lower portion.
AvgGain versus sample with the fit over plotted.
AvgGain - fit by sample
AvgGain - fit by za
AvgGain - fit by az
GainLoss computed from pitch,roll, and focus. The gain is down by
15% at high za.
Fit to avgGain. The upper coefficients are the fit to the measured data.
The bottom coefficients are the gain computed from the pitch, roll, and
focus errors, and then scaled to a maximum gain of 7.3 Jy (this was the
measured maximum). The plots show the azimuth terms for each fit (solid
measured data, dots - pitch,roll,focus). Comparing the azimuth components
should give you an idea whether the gain degradation is coming from pitch,roll,
focus or from some other source (say pointing or dish irregularities).
The 3 az terms line up in phase and match in amplitude. This tells us that
the 3az gain loss is coming from the pitch, roll, focus and that we have
parameterized the pitch, roll, focus losses correctly (at least for sband).
The 2az terms are in phase but the amplitudes differ by about 50%. The
1 az term in the data does not correspond to the 1az term in pitch roll
Warnings on using the fits:
The data used for the fit does not cover the entire dish. The sources are
biased for southern sources ( 6 south, 3 north) and there are no northern
sources above 25 deg dec. For sources close to a dec that was included,
the fit should be as good as its sigma's. Other measurements have
shown an azimuth dependence for high dec sources (>30 deg) . For these
sources the gain fit should be used with caution.
Coefficients from the fits:
Below are listed the coefficients from the various fits. The first column
in each row specifies the data that was fit. For gainAvg there is just
one fit while gain,tsys,sefd have a separate fit for polA and polB.
> last computed: Sat Sep 23 14:32:07 2000
> fit Ids
> 1 GainAvg
> 2 Gain
> 3 tsys
> 4 sefd