sbw cal values measured 15aug03
10cot03
Links to PLOTS:
hcorcal
value versus freq from sky and absorber
spectra
of hcorcal on/off-1 on sky
power
levels during hcorcal measurements
Ratio
of other cals to hcorcal tracking blank sky.
Compare
tsys using old and new cal values.
Links to SECTIONS:
Measuring the high correlated
cal using blank sky and absorber
Measuring the other cals on blank
sky relative to the high correlated cal.
Measuring
the high correlated cal using sky and absorber: (top)
The high correlated cal value (diode 1 going to
polA and polB) for sband wide was measured 15aug03 on the telescope using
the sky/absorber
technique. The telescope tracked blank sky and used 3 second
cal on/offs.
The cals values had also been measured at the receiver
test shack in sep01. These measurements included the measurement of Treceiver
across the band. The Treceiver measured was used in the 15aug03 reduction.
The temperatures used in the computation were:
Tabsorber |
303K |
Tsky+Tscattered |
20 K |
Treceiver |
from test shack |
The sbw receiver has some frequency ranges that are
unusable because of rfi. The cal measurements were done with the rfi filter
bank filters in. The gaps in frequency of the data are where the largest
rfi occurs.
Each frequency band was measured 6 separate times
on the absorber and 6 separate times on the sky. For each measurement the
calon/off-1 spectrum was computed. The 2nd order polynomial
was fit to the spectra for the 6 measurements of each bandpass (25 Mhz).
All points with fit residuals above 3 sigma were thrown out. The total
power for each 25 Mhz band was then computed masking out these bad points.
A 4th order polynomial was fit to this data to generate
the cal values (throwing out any data points greater than 2 sigma on the
first iteration of the fit). The calValue
versus frequency plots show the results.
-
Fig 1 is the measured cal value in deg K versus frequency. The colors are:
Blue is measured using the sky and absorber ratio. Black is using just
the absorber and red is using just the sky. The Top plot is polA and the
bottom plot is polB. The dashed blue line is the old cal values.
The dash green line is the receiver temperature used (from the antenna
test range measurement). The Tsky+Tscattered was adjusted so that the Cal
from the sky agreed with the cal from the sky absorber ratio.
-
Fig 2 is a 4th order polynomial fit to the calAbsorber and calRatio measured
values (it will be used to generate the cal value table). The green crosses
were used for the fit. The dashed blue line is the cal values measured
on the antenna test range. The purple line is from cumfiltering the 6 measurements.
-
Fig 3 and 4 show the cal values for each measurement (polA, polB). The
top plot has the 6 measurements on the absorber. The center plot shows
the 6 measurements on the sky. The bottom plot is the absorber, sky ratio.
The absorber measurements have more scatter then the sky measurements even
though the sky measurements had more rfi. The time bandwidth product should
have had an rms error of about .05 K.
The spectra
of calOn/Caloff-1 for the 6 sky measurements shows the spectra
for the on sky measurements. 100 Mhz junks at a time were taken. The dotted
vertical lines are the edges of the 25 Mhz bandpasses.
-
Fig 1. This has the entire range of 1.7 to 3.1 Ghz. Top is
polA, bottom is polB. Each pass through the frequency range is offset from
the previous one for plotting purposes.
-
Fig 2-3 has the same data broken up into 500 Mhz sections.
There are ripple at the low end of the spectra. The
dips at the high end are resonances
in the omt.
The final plot
has the power levels during the measurements.
-
Fig 1 top is the power at the upstairs fiber optic transmitter for all
of the measurements. There were 66*6= 396 records on the sky, and
then 396 records on the absorber. The edges of each 120 records (high at
low freq:8000 and low at 10Ghz) have a higher spectral density since we
are looking at total power and the 1 Ghz IF bandwidth is getting cut off
at the edges of the rf band. So the spectral density for the -37 total
power was probably ok.
-
Fig 1 center is the power at the downstairs if power meter.
-
Fig 2 The power counters. They record the 50 Mhz as seen by the digitizer
(before the digital filters).
-
Fig 3 plots the 25 Mhz 9 level total power measured via the 0 lag of the
correlator. A 25 MHz digital filter sits between the power counter of figure
2 and this measurement.
processing: x101/030815/hcorcal/sbwinp.pro,sbwcal.pro,sbwspec.pro,sbwdiag.pro
Measuring the other
cals using sky and the high correlated cal from above. (top)
The high correlated cal was measured above using sky
and absorber as the hot and cold load. The other cals were measured
relative to the high correlated cal. This went from 9am to 10am (after
the sky tracking of the hcorcal). Two passes were done through the
frequency range. The sky was clear the entire time. The cal
sequence was:
-
hcorcal(on,off)
-
hcal(on,off),hxcal(on,off),h90cal(on,off)
-
hcorcal(on,off)
-
lcorcal(on,off),lcal(on,off),lxcal(on,off),l90cal(on,off)
-
hcorcal(on,off)
100 Mhz at a time was measured (4 by 25Mhz) going from
1700 to 3100 Mhz. The cal was cycled on/off for 3 secs at each step.
The entire frequency range was repeated 3 times.
The calOn/calOff -1 spectra was computed for
each measurement (giving cal in units of calOff or Tsys). Each of
these 25 Mhz band passes was cumfiltered. This was done on the 2 loops
of the same freq band (256 channels *2 numbers). The total power was computed
and then the ratio of each cal value relative to the hcorcal was computed
(via interpolation of the 3 hcorcal measurements). A 5th order polynomial
was then fit to the data vs frequency. These values were used to generate
the cals table (after multiplying by the hcorcal in kelvins). The figures
show the other cals relative to Tsys and the hcorcal:
-
Fig 1. This is the cal value as a fraction of Tsys for each of the cal
types. The top figure has the high cals while the bottom figure shows the
low cals. The lines with * are polB. The cal types where the same diode
feeds the same polarization (polA horcal,hcal.. polB hxcal,hcorcal, etc)
are in agreement. Since they were measured at different times, the weather
was not having large affect on the data.
-
Fig 2. The calType/hcorcal. The top has the high cals while the lower plot
is the low cals. The * are an 5th order fit to the data.
-
Fig 3. The hcorcal was measured 3 times in each set of 10 measurements.
This graph plots the ratio of tsys for the 2ndHcorcal/1stHcorcal (red)
and 3rdHcorcal/1stHcorcal (blue). The * have the cal on while the + have
the cal off. Each frame is a different pass through the frequency range.
These are the raw tsys values. There has been no cumfiltering done yet.
If the + and * jump together, then it was Tsys that was changing. If the
+ and * jump differently then it could have been the cal value. The ratios
have interpolated between the 3 hcorcal measurements. Most jumps are within
.5%
Tsys is measured daily at 2290 Mhz for sbw.The final plot shows the
tsys values using the old and new cal values.
-
The top plot uses the old calvalues (only the high cals are plotted). Pol
B is the dashed lines. The measurement that drift up are all coming from
diode 1.
-
The bottom plot uses the new cal values. Day 223 was the date that
the new cal values were measured. The drift is in diode 1.
The cal diode values were recalibrated. The 15aug03 they probably give
the correct values. Diode 1 is drifting at about 4% per month. These diodes
need to be replaced if people are to use them for calibration.
processing:x101/030815/sbwcal.pro,chksbwcal.pro
home_~phil