Lbw response below 1100 Mhz

12feb07 Measuring the SEFD around 1100 Mhz
Looking at the rfi

Measuring the SEFD around 1100 Mhz:

The response of the lbw system was tested on 12feb07 to see if it could be used for galaxies below 1100 Mhz.

The setup:

The standard configuration for lbw has:

dewar
bandpass filter low frequency falloff around 1130 Mhz
pol Box with amplifier
filter bank

To test below 1130 Mhz the setup was:

PolB - from the dewar go to a 1200 Mhz low pass filter and then a 26 db amp (2 db noise figure). Then into the if/lo system.
PolA - from the dewar go to to 25 Mhz bandwidth tuneable filter and then a 26 db amp. Then into the if/lo system. The tuneable filter was tuned to 1100 Mhz.

The wave guide starts to cutoff around 1040 Mhz. A 500 Mhz bandpass centered on 1100 Mhz (via the spectrum analyzer) showed a 6 db difference between 1100 Mhz and the noise floor at 1000 Mhz. The noise floor was being set by the 26 db amp (2 db noise figure) rather than the sky/omt.

Taking the data:

On, off position switching using the source B0038+32 (3C19) was used to measure the system performance. It has a flux of 4.09 Jy at 1100 Mhz. The telescope was at za=15.1 degrees when the measurements were made. 2 1 min on/offs were done. The interim correlator was used with 25 Mhz bands centered at: 1040, 1060, 1080, and 1100 Mhz. The sefd was then computed as:

sefd=srcFlux/srcDeflectionTsysUnits.

This becomes Tsys/Gain with units of Janskies.

The plots show the system performance (.ps) (.pdf) :

Page 1 Bandpass shape: The first 4 plots show the bandpass shape for each of the 25 Mhz bands. Black is polB, red is polA. Only band 3 has appreciable power in polA (because of the tunable filter). Each of the 25 Mhz bands have different attenuation. The bottom plot places the 4 bands on the same power scale so you can see the power drop off with frequency. The power starts to rise around 1090 Mhz probably from an increase in the system temperature. It turns over and starts coming down around 1070. This may be the omt starting to cutoff the frequency. The power drops by 7 db between 1070 and 1040 Mhz. The actual drop may be greater since Tsys is probably increasing over this range.
Page 2 SEFD below 1100 Mhz: This shows the sefd for the system. 4.09 Jy was used for the source strength. Both on/offs are over plotted. Black is polB, red is polA. Only the 1100 Mhz band for
PolA (red) is higher than polB (black). This probably because the tuneable filter has a larger noise figure than the 1200 Mhz low pass filter.

The SEFD is 4-4.5 till around 1080. It is 5 and 1060 and then shoots up rapidly. There is probably not much sky getting through the omt below 1050 Mhz.

Conclusion:

The sefd looks pretty good down to 1080 Mhz (4. vs 3.5 at 1175).

Things to think about:

Filter edges can be less stable than the flat portions. We haven't looked very hard to see if there are any omt resonances in this frequency range.
The rfi needs to be looked at closer.
This setup will probably not be the final configuration for use (since it required manual intervention to set it up).

Looking at the rfi around 1100 Mhz with lbw.

rfi near 1100 Mhz was monitored with LBW and the interim correlator on 12feb07. A 1200 Mhz low pass filter was used instead of the standard 1120 to 1750 Mhz bandpass filter. 500 1 second dumps were taken with each data set. The frequency channel resolution was 50 Khz (after hanning smoothing). The data sets were:

60 Mhz of data centered at 1100 Mhz. The 60 Mhz narrow band filter was placed after the 1st mixer to cut down on IF harmonics of the aerostat radar. The telescope was sitting at az=270,za=18 degrees.
80 Mhz centered at 1070 Mhz. Unfortunately this data was taken while tracking the 4 Jy source above (so Tsys was about doubled).

The plot shows the rms/mean (by frequency channel) (.ps) (.pdf) :

The black line is the 80 Mhz centered at 1070. The red line is the 60 Mhz centered at 1100 (it has been offset for display).

Dynamic spectrum for each 25 Mhz band show the time variation of the rfi for the 500 seconds.
The first 4 images are less sensitive than the last 3 since they were taken while tracking a 4 Jy source (doubling Tsys) (you can see this by comparing the two 1080 Mhz bands).

1040 Mhz cfr trking src(.gif) There is probably not very much sky signal getting into this band so it looks better than it should.
1060 Mhz cfr trking src(.gif)
1080 Mhz cfr trking src(.gif)
1100 Mhz cfr trking src(.gif)
1080 Mhz cfr sitting(.gif) the birdies at 1082 and 1084 have a time variability of 45 seconds.
1100 Mhz cfr sitting(.gif)
1120 Mhz cfr sitting(.gif)

The worst rfi is around 1090 Mhz. It gets up to 1.5*Tsys in a 1 second integration and covers 4-5 Mhz.

processing: x101/070112/lbwrfi.pro

 home_~phil