cband hi (6 to 8 ghz) calibration
feb04
The cband hi receiver covers 5.9 to 8.1
Ghz (maybe a little more on the edges). It is a dual beam receiver but
only one of the beams is currently installed. cbh or cbh_1 refers to the
first (currently installed horn/amp). cbh_2 is the 2nd horn/amp.
Sections:
History
Recent
system performance measurements
Daily
monitoring of Tsys
Dewar temperatures
Calibration measurements
Miscellaneous
History:
14feb04 surveyed horn into position.
11feb04 installed in dome.
Calibration measurements:
29feb04: fit Gain Curves to feb04 calib data.
20feb04: looking at 6.7 Ghz masers:
IRAS srcs: 19410+2336, 20110+331
16feb04: Tsys vs frequency
15feb04: on/off position switching on
3C454.3 covering the entire band
15feb04: first look at the cband hi
performance.
Miscellaneous
08aug05: check system performance for Thankins.
18may04: aas plots
for jagadheep
15feb04: RFI: rms/mean vs frequency
for 60 second blank sky.
08aug05 check system performance
for Thankins. (top)
Tim hankins had been looking at giant pulses from the
crab with cbandhi (as well as cb,xb,sbh..). He had not seen any pulses
for a few days and asked us to check the system performance. A calibration
run was done on 08aug05 using B1622+238 (3C336) .62 Jy at 7000 MHz and
B1829+290 .91 Jy at 7000 Mhz. The calibration runs make measurements at
6600, 6900, 7200, and 7400 Mhz The plots show the
measured system performance (.ps) (.pdf):
-
Page 1: plotted vs za is: gain, tsys, SEFD, and average beam width. The
source B1622+238 is extended (~24 asecs) so its gain/sefd/beamwidth should
be worse.
-
Page 2: This is the pointing error in za (top two plots) and azimuth (bottom
2 plots). The source B1645+174 is included. It was taken on 05aug05 using
sband high (3-4 GHz).
The system performance and gain do not look like they should be causing
problems with the cband high measurements. Tim's lack of pulses must have
been some other problem (un-cooperative pulsar???).
processing: x101/050808/doit.pro
feb04. fit GAIN
CURVES to calib data. (top)
link to
gain curve plot
link to system
performance for the data used to fit.
Gain curves were fit to the cband hi gain data taken
during feb04..The plots
show the gain data (black) and the fits (red) for 6600, 6900, 7200,
and 7400 Mhz.
-
Fig 1 shows the az,za distribution for the data. The fit used a 3rd order
polynomial in (za-10) and 1az, 2az, and 3az sin, cos terms.
-
Fig 2 plots the gain data and the fit to za. The fit equation is
plotted with the sigma for the fit (in K/Jy).
-
Fig 3 plots the fit residuals (data-fit) vs za.
Plots of the
system performance (gain,tsys, sefd, beamwidth,...) for
this dataset is also available.
The routine gainget() or corhgainget() will now return
the cband hi gain for data after 14feb05 (horn installation).
The coefficients can be found in the ascii file data/gain.datR10
(this is provided in the AO idl distribution for correlator
routines). You can also find a copy of it at AO in /home/phil/idl/data/gain.datR10.
processing: x101/cbh/feb04/dogainfit.pro
20feb04: looking
at 6.7 Ghz masers: IRAS srcs: 19410+2336, 20110+331
Two masers were tracked on 20feb04 with the cband hi
receiver. The correlator was set to: .781 Mhz with 2048 channels, and .3905
Mhz with 2048 channels. This gave frequency resolutions of .034 km/sec
and .017 km/sec (after hanning smoothing). The plots
show the masers.
-
Fig 1 was a 60 second on, off position switching. The plot is in degrees
K. There has been no gain correction (since we do not yet have a model
).The bottom figure is a blowup showing the baseline. The rms noise is
what is expected.
-
Fig 2 is IRAS source 20110+221. There were 9 60 second ons each followed
by a cal on, off. Each scan was scaled to kelvins and then a linear baseline
was removed (there was no bandpass correction). The top two plots are the
wider resolution (.034 km/sec). The bottom two plots are the higher resolution
(.017 km/sec). There is a birdie in the center of the band that comes from
the correlator digital filters. The measured rms's are a bit higher than
the expected. This is probably do to contamination from the source, and
the fact that this was not a position switched measurement so the bandpass
was no divided out. Since no off position was taken, there is no information
about the continuum from the source.
The receiver saw the two masers. The noise levels measured were close to
the expected values. The source flux is not known since there is not yet
a gain curve. It also turns out that the individual components from these
masers tends to be variable.
You can see previous measurements of these sources
in A
survey of the 6.7 Ghz methanol maser emission from IRAS sources
(M. Szymcazk. Astron.Astrophys. Suppl. Ser. 143,269-301).
processing:x101/040220/cbh.pro
15feb04: tsys vs frequency.
(top)
On 15feb04 on off position switching was done on the
source 3C454.3. The correlator was setup for 4 by 50Mhz bands. 60 1 second
records were taken at the on and then the off position. The cal was then
fired after the off. The band from 5750 through 8150 was covered two complete
times.
The cal on, off was used to compute the system temperature.
10 percent of the bandpass along the edges as well as any outliers were
not used in the computation. The average Tsys was then computed for each
of the 50Mhz bands. The plot
shows Tsys vs frequency.
-
Tsys PolA is black.
-
Tsys polB is red.
-
The green line is pola (srcOn-off)/off for 3C454.3. It has been scaled
and offset for plotting. It shows where the resonances in the feed occurs.
-
The blue line is polB (srcOn-off)/off for 3C454.3.
The resonance at 5845 and 7336 cause Tsys polA to increase. Part of this
will be a true Tsys increase while part will be an error in the measurement
of the cal values at these frequencies (since they were injected after
the omt and did not go through the resonance). Tsys for polB jumps around
a lot more than pola fore freq less than 7200 Mhz. This same pol is the
one that has a large number of resonances in the region below 7200 Mhz.
processing: x101/040215/cmptsys.pro
15feb04: on/off position
switching on 3C454.3 covering the entire band. (top)
On off position switching was done on the continuum
source 3C454.3. The correlator was setup to do 4 50 Mhz frequency junks
simultaneously. A one minute on,off was done and then the frequency band
was moved up by 200 Mhz. The frequency range 5750 through 8150 Mhz was
sampled two complete times. The first pass covered za 15 through
5 degrees za. The second pass covered the za range 3 degrees rising to
10 degrees setting.
(ON-OFF)/OFF was computed for each pair. The data
is plotted in units of Tsys with the full frequency resolution (24Khz but
100 channels on each each of a 50Mhz band is not plotted).
The
plots shows the source strength in units of Tsys vs frequency.
-
Pol A is in black and polB is in purple. There are two complete traces
for each pol since there were two complete passes.
-
The green horizontal lines are the boundaries of each 50Mhz subcorrelator.
-
The red vertical lines are spaced by 200 Mhz. Data between these lines
was taken at the same time.
-
The jumps in power levels for each 200 Mhz integration is probably from
the pointing.
-
There are two large resonances at 5840 Mhz and 7330 Mhz. The are negative
going since the resonance in the on is negative going while the resonance
in the off is positive going.
-
The ripples in polB cover most of the band. The ripples in polA start
after the 2nd resonance and then follow what pol B is doing.
The omt has a Y with polB going off at the angle
and pol A going straight thru (at least i think this is the correct order).
The positive going resonances in the off were a bit of a surprise at
first. If the omt is cooled to 20 kelvins, then it should have close to
the same temperature as the sky radiation. In this case you should not
see much of a bump. This is what happens with our other cooled mts. On
the other hand, there is a thermal transition with a long 50 K stage, and
then another transition to the 300 K. If the trapped mode is reflecting
back toward these higher temperature areas, then you would probably expect
a bump in the resonance while on the sky.
We need to put an absorber in front of the horn
and do this same experiment. It will give us a cleaner picture of what
is going on without any jumping around because of the pointing.
processing: x101/040215/onoff3c454.3_15feb04.pro
15feb04: first look
at the cbhi performance. (top)
Calibration scans were done on 4 sources on 14feb04
(after the horn survey) and 15feb04. Data was taken during the daytime
(13:00 hours to 15:00 hours). The frequencies measured were: 6600,
6900, 7200, and 7400 Mhz. The source J2253+161 (3C454.3) was used just
for the pointing offsets since it is a variable source. The plots
show the system performance. The plotted parameter is the average of
polA and polB (stokes I/2). The cal values measured on the hilltop test
range were used in the calibration. Color is used to separate the frequency
measurements while symbols separate the sources.
-
Fig 1 has the gain (K/Jy), Tsys (degK), Sefd (Jy/Tsys), and beam width
(asecs).
-
Fig 2 plots the coma parameter, first sidelobe height, main beam
efficiency, and the main beam plus first sidelobe.
-
Fig 3 has the zenith angle pointing error (vs za and azimuth ) followed
by the azimuth pointing error (vs za and azimuth).
-
Fig 4 shows the az,za coverage of these sources on the dish.
The gain (which depends on the
validity of the cals and the source fluxes) varies from 2.5 to 7.5 K/Jy
for the frequencies measured. The source B0134+329 (3C48) has the largest
gain (and also the best known flux). The sefd goes from 4 to 8 Jy/Tsys.
The Tsys measurement is the average of polA and polB.
processing: x101/040215/doit.pro
15feb04: RFI: rms/mean vs
frequency for 60 second blank sky. (top)
plots
of the rms/mean for the entire band.
plots of
the spectral density function for the bands with rfi.
On 15feb04 60 second on/off position switching was
done on 3C454.3. The off position scans were used search for rfi in the
band. For each off position the rms/mean was computed by frequency channel
for the 60 one second integrations. This was then plotted versus frequency.
The correlator was setup to cover 200 Mhz in 4 50Mhz bands using 3 level
sampling. It took 12 integrations to cover the entire 2.4 Ghz band. The
entire band was measured 2 times. The
plots show the rms/mean for the entire band.
-
PolA is black, polB is purple.
-
The green lines are the edges of the 50 Mhz bands
-
The red lines are the 200 Mhz integrations that were done at one time.
Frequencies with large rms/Mean are rfi (since it is usually not stationary
in time).
The second plot shows the
spectral density function for the bands with rfi.
-
The plot is a peak hold on each channel for the 60 seconds.
-
There are two complete passes through the band. The second pass had the
frequencies offset by 5 Mhz from the first pass.
Note: This data was taken from 12:45 to 14:00 AST . Some of
the rfi could be reflections from the sun.
processing:x101/040215/onoff3c454.3_15feb04.pro
home_~phil