Dynamic Range Tests on Cass A

October 25, 2003
F. Ghigo, NRAO-Green Bank

The feasibility of mapping Cass A and its surroundings was tested by doing scans across Cass A and across weaker sources to see how linear the response of the system is to a wide range of power levels.

System Components

Table 1. Journal of Observations: 410 MHz, October 19, 2003

Scan# Source type of scan levels
4-7 3C286 peak/point normal on background
9 Cass A Scan 5 deg in RA set for peak of Cass A
10 Cass A scan 5 deg in RA set for peak of Cass A
11 3C286 scan 5 deg in RA set for peak of Cass A
12 3C286 scan 5 deg in RA set for peak of Cass A
Levels: "normal" levels means power in Prime focus receiver and IF Rack is set to nominal 1v and Spectral Processor to nominal -11db inputs when off source.
"set for peak of Cass A" means power meters in Prime focus receiver and IF rack set at about 8v (near high end of allowable range) and Spectral Processor inputs set to about -5 db.

Setup: Scans 9-12 used 20 MHz bandwidth centered on 410 MHz for both the continuum(DCR) and Spectral Processor backends. Used 0.1 second noise cal switching; Spectral Processor integration time 1 sec; scanning at 200'/min.

Results and discussion summarized at the bottom of this page.

Peak Scans on 3C286

Peak Scans on 3C286 were done with normal level settings on off-source background. Used normal Tcals (2.73 and 2.80 K). These were done with the continuum backend for a 20 MHz band centered at 410 MHz.
Scan 4 scans across the source in the +RA direction; scan 5 in -RA; scan 6 in +DEC; scan 7 in -DEC.
The data were calibrated using the Tcals and are presented in units of Kelvins.
In the plots, the observed data are shown in red and the fitted gaussians in green.
Plot 4A shows peaks on Xpol data.
Plot 4B shows peaks on Ypol data.

Table 2. Summary of peak data.
Xpol Ypol
Scan# Tsys(K) Peak(K) Tsys(K) Peak(K)
4 62.1 53.5 60.3 48.3
5 62.4 65.0 59.6 57.2
6 63.6 52.4 58.2 51.3
7 63.8 59.3 58.3 58.2

Discussion
The flux density of 3C286, extrapolated to 410 MHz using the data from Ott et al (1994), is 22.89 Jy. The mean of the peaks is 55.7 K, which means the gain is 2.4 K/Jy. This is too high -- the aperture efficiency by design is 70%, which works out to a gain of about 2.0 K/Jy, So the higher value of 2.4 is probably wrong. The Tcals may be error, and should be closer to 2.3K rather than the values we are using.
The variation of the peak values from one scan to the next is large, but this may be explained by the fact that variable RFI is in the 20 MHz band. One can see that the peak are ragged -- rfi spikes are occurring in every peak. This may be increasing the peaks and may explain the high gains.

Spectral Processor Scans

The GBT scanned across Cass A for a 5 degree track centered on the source at a rate of 200'/minute. The spectrum was dumped once per second. Thus there were 90 spectra in each scan, and a 3.3' spacing between spectra. Since the HPBW is about 30', there are 9 spectra per beam. All spectra have 1024 channels.
The high Tcal was used (30.99, 32.98 K) and was switched at a 10 Hz rate.
Plot 9A shows the first integration of Scan 9 across Cass A. This is uncalibrated, in arbitrary units.
Plot 9B shows integration number 45, on Cass A. Also uncalibrated. Any RFI is overwhelmed by Cass A.

Processing the Spectral Processor Data:
After looking at many of the individual integrations, we determined the rfi-free channels to be the following: channel numbers 40-230, 280-470, 600-660, and 780-920.
For each of the 90 integrations, we averaged the rfi-free channels, separately for the cal-on and cal-off phases.
We used the difference between the cal-on and cal-off phases averaged over integrations number 5-30 and 60-88 (the integrations without the source) and divided by the Tcal to get a scaling factor to convert to units of temperature.
The result, for scan # 9 is shown in the next plot.
Plot 9C shows the calibrated continuum data on Cass A. The higher peak (red/green) is X polarization, and the lower (blue) is the Y polarization..

Similar processing for the remaining scans yields the following plots:
Plot 10A A second calibrated scan across Cass A, done just the same way as for scan 9.

Plot 11A Shows the scan across 3C286, processed the same way. The red and green lines are the Cal-off and Cal-on phases, respectively, for X polarization, and dark and light blue are the same for Y-pol.

Plot 12A Shows the second scan across 3C286.

Summary of results

A summary of results from scans 9-12 is shown in the next table.
Tsys means the average baseline level for the cal-off phase.
Peak is the difference between the peak of the plot and Tsys.
Gain is Peak divided by the source flux density.
For the flux density of Cass A at 408 MHz in the year 2003, we use 4223 Jy, based on Baars et al (1977).
For 3C286, we use 22.89 Jy, as described above.

Table 3
Scan 9: Cass A Scan 10: Cass A Scan 11: 3C286 Scan 12: 3C286
Xpol Ypol Xpol Ypol Xpol Ypol Xpol Ypol
Tsys (K) 62 73 62 77 23 41 23 40
Peak (K) 9340 8840 10046 9480 58 53 55 50
Gain (K/Jy) 2.21 2.09 2.38 2.24 2.53 2.32 2.40 2.18

Discussion


Let's also compare with the peak data (scans 4-7) done on 3C286. Here the gain is about 2.4 K/Jy, which agrees pretty well with what we have in Table 3. So things are linear to 5% for power levels ranging from normal to the low end of the range.

But some worries remain:

  • Why do we see variations of the order 10% from one scan to the next?
  • and why is the Tsys for the 3C286 scans much lower than for Cass A?