327 receiver calibration.
The 327 Mhz receiver is a dual linear receiver
with room temperature amplifiers and a 30 Mhz wide filter.
Sections:
History
Recent
system performance measurements
Daily monitoring of tsys.
Calibration measurements
rfi
miscellaneous
cal values
rfi measurements
all measurements in date order
haslam et al 408 Mhz all sky survey (.pdf) ..
18.9 Mb
History:
04jan10: Now cooling is on one of the alfa compressors.
13jun08: Cooled 327 receiver installed. New cal values also
installed.
13apr06: rf cable of channel A between the front end and the
filter
at the post amp was replaced because it was bad. The cal cable for
channel
B between the postamp chassis and the antenna was also replaced.
11apr06: 750 Khz comb appeared. 39.45 Khz comb was seen (but not sure
when
it appeared). Turned out to be the sband
cabinet in dome.
07jan04: 327 receiver moved to a new position on the turret floor (to
make
room for alfa).
jul02: room temperature receiver installed.
11jul02: cal values measured in lab.
Calibration
measurements: (top)
21jun08: SEFDA/SEFDB shows that tsys polB is
10% larger than tsysB.
feb05: gain curves for data
taken
mar03 to feb05 (GAIN CURVE)
feb05: system performance
for gain curve data mar03-feb05
08aug02: Calon/off sky and
load.
Tsys contributions receiver, sky
07aug02: calibration runs
3C138,B0640+233,
and B08224+294
09aug09:
rfi from
azimuth encoders
Cleaning
up the 327 Mhz band (last updated 03nov05):
05nov05:
azimuth dependence of birdies in the 327 rcvr.
21oct05:
output of 327Mhz rcvr after the 1st amp (and before the rf filter).
16oct05:
1 Mhz comb coming from dome AC units.
27aug05:
the 16Khz comb in the 327 rcvr comes from the alfa motor controller.
Miscellaneous (top)
07may09:
tsys vs sky position for jun2008 to may09
16jul08:
tsys vs date,hr,ra,dec with cold receiver
07may07:
Strong 60 Hz in p1693 pulsar data using the 327 receiver and the wapps.
07jan04: 327 receiver
moved to new turret position.
jul02
:Room temperature receiver installed
29sep00:Turret position
All measurements in date order
(top)
09feb10: p1693 drift scan 220 sec dynamic
spectra and spectral avgs.
09aug09:
rfi from
azimuth encoders
07may09: tsys vs sky position for jun2008 to may09
16jul08: tsys vs date,hr,ra,dec with cold receiver
21jun08:
SEFDA/SEFDB shows that tsys polB is 10% larger than tsysB.
08aug02:
Calon/off sky and
load.
Tsys contributions receiver, sky
07aug02:
calibration runs
3C138,B0640+233,
and B08224+294
07may07:
Strong 60 Hz in p1693 pulsar data using the 327 receiver and the wapps.
07jan04: 327
receiver
moved to new turret position.
05nov05:
azimuth dependence of birdies in the 327 rcvr.
Cleaning
up the 327 Mhz band (last updated 03nov05):
21oct05:
output of 327Mhz rcvr after the 1st amp (and before the rf filter).
16oct05:
1 Mhz comb coming from dome AC units.
27aug05:
the 16Khz comb in the 327 rcvr comes from the alfa motor controller.
feb05: gain curves
for data
taken
mar03 to feb05 (GAIN CURVE)
feb05: system
performance
for gain curve data mar03-feb05
jul02
:Room temperature receiver installed
29sep00:Turret
posit
09feb10:p1693 drift scan:
p1693 did a long drift scan using the wapps on
09feb10. The setup was:
- cfr: 327, bandwidth: 50Mhz, 3lvl sampling, 200Usec dumps, addpols
- az,za:275,15.49
The first file of data was looked at (about 2 minutes)
dynamic spectra for 2 minutes of data
(.gif):
- spectra were hanning smoothed and averged to .5 seconds.
plots of the average spectra and the
rms/mean (.ps) (.pdf):
- Top: the 220 second average spectra. The vertical scale is linear
in power.
- Bottom: the rms/mean by channel for the first File. The rms was
computed after averging to .5 seconds.
- The expected rms (from the radiometer equation) is show in red.
processing: x101/100209/p1693drift.pro
07may09: Tsys vs ra/dec and Tsys vs
galLong/galLat. (top)
Tsys data from 13jun08 (start of 327 cooled
receiver) thru 07may09 was plotted vs ra,dec and galactic l/b to show
the portion of Tsys that is coming from the sky. Only data with zenith
angle below 15 degrees was used (to not include the tsys za variation).
PolA plott
Tsys
vs sky position (.ps) (.pdf):
- PolA is plotted in black
- PolB is plotted in
red
- Page 1: tsys vs ra,dec
- Top: ra,dec positions where the measurements were made.
- We normally sit at az=270,or 285 and za=around 10.
- Center: Tsys vs Ra.
- Places with elevated Tsys
- ra=6 galactic anticenter
- ra=15-17 north polar spur
- ra=19 : galactic plane
- Bottom: Tsys vs Declination
- Not much info since most decs are 15 or 18 Deg.
- Page 2: tsys vs galactic longitude and
latitude.
- Top: galactic latitude and longitude for the measurements
- Center: Tsys vs gal Longitude
- Places with elevated Tsys
- long=35 North polar spur
- long=50 galactic plane
- long=190 anticenter
- Bottom: Tsys vs galactic latitude
- lat -5 anticenter
- lat=5 galactic plane
- lat 30-70 top set, north polar spur
- Page 3: TsysB - TsysA vs ra, galactic
latitude
- The top plot is vs Ra, the bottom plot is versus galactic
longitude.
- There is not a lot of structure vs ra or longitude.
Polarization is probably not playing a major part.
- The median value is 8.9 Kelvins with an rms of about 5K
Summary:
- The lowest Tsys for polA was about 85 K around ra=9-12 hrs
- Tsys can increase to > 160 K on the plane
- PolB is 8.9K warmer than polA.
Processing:
x101/090506/tsysplotradec.pro
16jul08: tsys vs date,hr,ra,dec with cold
receiver (top)
The tsys measurements from 14jun08 to 16jul08
were used to plot tsys vs various parameters to see why tsys was
sometimes high.
The plots show Tsys vs date, hr, ra, dec,
za (.ps) (.pdf): The black *
are polA while the red * are polB. Tsys for
polB is about 10K higher than polA.
- Top: tsys vs date.
- 2nd: tsys vs Hour of day.
- tsys before 10 am seems to be higher than than after
10am. Could it be sunrise??
- 3rd: tsys vs za. we usually do tsys around za=10.
- 4th: tsys vs ra.
- The high points at ra=15 are the north polar spur. At 408MHz it
is about 20K higher than the ra=10 data (327 will be higher).
- The points at 2-4 hours with tsys =100 do not have a lot of
continuum. These are the 7-10am points in plot 2.
- 5th: tsys vs dec. Most points are taken between decs of 15-20
degs..
- The high point at dec=2 corresponds to ra=15. The continuum
here is not that large (the north polar spur picks up at ra=15,
dec=10-20) So i'm not sure why this is so high.
SUMMARY:
- Tsys can vary by 20-30 kelvins with sky position
- 8am-10am seems to be higher than 12-19 pm.
21jun08: Ratio SEFDA/SEFDB shows Tsys
B > TsysA (top)
The tsys
measurements after the cooled receiver was installed on 13jun08
showed that Tsys PolB is about 10K higher than Tsys polA. To see if
this was a cal problem or a real Tsys difference, calibration
scans using the source B0316+162 (CTA21) were done and then the SEFD
ratio was computed. The setup and processing was:
- source CTA21 is unpolarized
- 03 hours is far enough away from the north polar spur that the
background should not be polarized.
- Bandwidths of 25, 1.5, .78 Mhz were measured all centered at 326
Mhz.
- SEFD=Tsys/Gain or SEFDA/SEFDB = (TsysA/TsysB)*(GainB/GainA)
- since both Tsys and Gain depend linearly on the cal, the cal is
cancelled by the ratio. The sefd is not affected by errors in the cal
The plots show the calibration measurement
results (.ps) (.pdf):
- Top: TsysA and TsysB vs za.
- TsysB is 10 to 15 Kelvins higher than TsysA
- The 25 Mhz band of polb is about 5K higher than the 1.5 and .78
Mhz bandwidths. This may partly come from using the cal value at the
center of the band for the entire band.
- Middle: Tsource A and T source B.
- PolA and polB both give close to the same source temperatures
so the cal values are not wrong by much.
- PolB gives the same source temperature for all of the
bandwidths so using the cal value at 326 MHz is not the reason why the
top plot showed a difference in the polB bandwidths.
- Bottom: SEFDA/SEFDB
- The ratio is 90 to 93% so TsysB is really higher than TsysA.
Conclusion:
- Tsys polB is about 10% higher than Tsys polA.
- Tsys (measured at 326Mhz) for polB 25 Mhz is about 5%
higher than Tsys measured with 1.5 or .78Mhz bandwidths
processing: x101/080621/chk327.pro
28feb05: gain
curve
for data taken mar03 to feb05 (top)
Calibration data taken march03 to feb05 was used to
fit a gain curve for the 327 receiver (the data was taken at 326
Mhz). For the data to be included in the fit, it had to satisfy the
following
constraints:
- 6 < gain < 11.5 K/Jy
- Sefd < 18
- average beam width between 800 and 950 arc seconds.
These constraints were used to remove some bad measurements. The cal
values
used were those measured in the lab in jul02 (see the system
performance
for gain curve data for a more complete description).
Gain curves were then computed fitting for gain
as a function of za. The fitting function (with za in degrees) is:
g(za)=c0 +c1*za + c2*(za-14)^2
+c3*(za-14)^3
where the ^2 and ^3 terms are only applied for za ge 14 degrees.
The plots show the gain
curve and the residuals (.ps) (.pdf):
- Fig 1: the azimuth , za coverage for the data.
- Fig 2 top: Red is the data, black is the fits. The fit
coefficients are
listed on the plot. The fit rms is .47 K/Jy.
- Fig 2 bottom: The fit residuals. The colors are different sources.
The source with low flux (red x ) i s B0320+053. Its flux may be a bit
off. At 1400 Mhz our flux differs from nvss by about 8%. The yellow
diamonds
that are high come from B0521+281. The cal values for the 327 receiver
have jumped around by up to 10 kelvins in the last two years (see
327
Mhz tsys difference 2004,2003). If the polA-polB jump
was
do to a shift in the cal, then it will affect any calibration data
taken
during these jumps (i guess i need to check this out..)
The fit coef can be found in the ao idl distribution
in the file data/gain.datR1. At AO it is found in the file
/share/rsi/local/libao/phil/data/gain.datR1.
processsing: x101/327/feb05/dogaintfit.pro
28feb05:
system
performance for gain curve data mar03-feb05 (top)
Calibration data taken mar03 to feb05 was used to
fit
gain curves for the 327 receiver. For the data to be included in the
fit,
it had to satisfy the following constraints:
- 6 < gain < 11.5 K/Jy
- Sefd < 18
- average beam width between 800 and 950 arc seconds.
These constraints were used to remove some bad measurements. The cal
values
used were those measured in the lab in jul02. The plots show the
system performance for the gain curve data (.ps) (.pdf)
:
- Fig 1 top: the az, za coverage for the data
- Fig 1 bottom: the time of day when the data was taken
- Fig 2: gain, tsys, sefd, and average beam width. B0640+233 had a
beam
width
that was larger than normal. It tended to make the gain a little less.
- Fig 3: coma, avg first sidelobe, eta main beam, eta main beam +
first
sidelobe.
- Fig 4: pointing error: za vs za, za vs az, az vs za, az vs az.
The
large
az error for B0017+154 was after the horn had been moved and before the
final turret offset had been computed. This is purely an az offset and
did not affect the gain. Some of the other az,za pointing offsets were
probably from tilts (rain etc..) that would also affect the gain.
processing: x101/327/feb05/doit.pro
07jan04: 327
receiver
moved to new turret position. (top)
The 327 receiver was moved on 07jan04 from turret
position
21.5 to turret position 340. This was done to make room for the alfa
receiver.
Calibration scan data was taken on 05,06jan04 before the move and on
07jan04
after the move. The goal is to make the average pointing error
before
and after the moves the same. The az error can be manipulated by
changing
the turret position, the za error can only be modified by moving the
feed.
The pointing
error before and after the move is shown in the plots.
- Fig 1 has the pointing error in great circle arc seconds vs
azimuth and
za. The black color is before the move with the turret at 21.5 degrees.
The green color is after the move with the turret at 340.2 degrees.
The za pointing error remained within 1 arc second. The azimuth
pointing
error for source 2 on the second day (green) varied with za (tiedown
losing
tension??). To compute the mean (actually median values) I only used
source
1.
|
medianAzErr
asecs |
medianZaErr
asecs |
| before move(tur=21.5) |
-11.73 |
-17.05 |
| after move (tur=340.2) |
-26.84 |
-16.39 |
The average az error with the turret at 340.2 degrees is
(-11.73-(-26.84))=15.1
asecs too small. We need to move the azimuth by +15.1 asecs. The turret
coordinate system is opposite to the azimuth and 1 turret degree equals
45 arc seconds on the sky. The turret needs to be moved to
340.2-(15.1/45)=339.9
This will be the new turret position.
08aug02:calOn/off
load and sky. Tsys contributions receiver, sky. (top)
30 cal on offs were done while the load switch was
on
, and 30 more calon/offs were done while tracking blank sky (ra:075055,
dec:240000). The telescope was centered at 327 Mhz and 25,12.5,6.25,
and
3.125 bands were used. 2 seconds cal on were followed by 2 seconds cal
off. The cal values used were: 73.2 pola, 76.1 polB.
The load is beneath the rotary floor close to the
antenna. The temperatures were: room:92, platform:85, tertiary
platform:92
(this was after an sband radar run). Remeasuring these values 20
minutes
later, the temperatures were all within 1 F of the original values.
Using
an ambient temperature of 90F gives a load temperature of 305K.
The plots show the
system temperature measured for the cal on/offs. The lines are
calOff,
the * are cal on. PolA is black while polB is red.
- Top: Cal on,off using the load.
- Bottom: Cal on,off tracking blank sky.
On load calOff the average temp was 375K. Using
305K
for the ambient load temperature gives a noise contribution from after
the load of 375K-305K=70K for a noise contribution after the load
switch.
This is what is expected for the room temperature amplifier.
On sky calOff the average temp was 135K. This
includes
the flange temperature, sky temperature, and any
ground/scattered/spillover
radiation. Using 40K for the sky contribution and 70K for the flange
temperature
gives 25K from the spillover/scattered ground radiation. Higher
frequency
receivers have shown scattered radiation ranging from 10 to 20 K so 25K
for the 327 receiver seems reasonable.
The sky temp was crosschecked with chris
salters
408 Mhz survey and it came out to 39.1K.
327 Tsys budget
| load temp |
305K |
| receiver temp |
70K |
| Tsys on sky |
135K |
| sky temp |
40K |
| spillover,scattered radiation |
25K |
processing: x101/020808/docal.pro
07aug02:
calibration
runs 3C138,B0640+233, and B0824+294 (top)
Calibration runs using the heiles scan were done on
3C138 (B0518+165), B0640+233, and B0824+294 during the day. The first
two
sources were done after transit, while the last (B0825+294) covered 2
strips
before transit until set. B0640+233 has a width of 72" by 17" at
lband while the other two sources are < 16". The field has
other
sources within the +/-3 beams that were on the order of 3% of the
source.
The plots
show
the calibration results:
Fig 1. has gain Kelvins/Jy, Tsys, Sefd, and average beam width. The
za
dependence of the gain and tsys look correct. The gain is about 11
K/Jy,
Tsys about 150 K, SEFD 14, average beam width: 14 Amin
Fig 2 shows the coma parameter, first sidelobe height: -14 db, main
beam
efficiency 65% (using 3C138), and main beam + first sidelobe efficiency
of 80% (using 3C138)
Fig 3 has the pointing errors for these runs. Most of the data was
taken
after transit so the pointing errors may be biased to one part of the
dish.
The mean za Error is -142 Asecs with an rms of 34 Asecs. The feed is
pointing
too far downhill. The mean azimuth error is -3 Asecs with an rms of 24
Asecs. This feed was not surveyed into position. The mean pointing
error
can be fixed by updating the model for 327. (Note: updated pointing
model
07aug02).
measured 327 performance
| gain |
11 K/Jy |
| Tsys |
150 K |
| sefd |
14 Jy/Tsys |
| avgBeamWidth |
14 Amin |
| First sidelobe |
-14 db below peak |
| mainBeamEffciency |
65% |
processing: x101/020807/doit.pro
jul02
Room temperature receiver installed
The 327 receiver had been removed in 2001 to be
worked
on. It was re-installed in jul02 with a room temperature receiver.
29sep00: Turret
Position:
The 327 system was installed in late september under the turret floor.
For a number of receivers we have found a local maximum in the system
temperature
when the turret is aligned with the optic axis. I did a number of
turret
scans on blank sky to find the turret position of this local maxima.
setup:
On 29sep00 the turret was driven in a sine wave about turret=22 degrees
with an 8 degree amplitude and 15 second period. The telescope was
parked
at az=270, za=10. 2 polarizations were recorded. Each was detected with
a .02 sec time constant.
The turret
strips figure shows the average of 8 minutes of data tsys versus
turret
position. The top plot is polA and the bottom plot is polB. The units
are
a/d counts with the average value as the system temperature. The
local maxima occurs at 24.5 deg (polA) and 17.5 deg (polB). The turret
scale factor is 45 arcsecs (great circle) per turret degree. If
the
difference is caused by a misalignment of the dipoles then it is about
5.25 arcminutes or 1/3 of a beam. We need to do some scans on a
continuum
source to verify that this is beam squint.
processing: x101/000929/chkpnt.pro
home_~phil