Trouble with lbw polB 1230-1470 filter (filter number 3)

        New diodes were installed in the lbw receiver in nov05. On 24nov05 they were measured using the absorber. The setup was:

• Place the absorber over the mouth of the horn
• Step thru 1120 - 1720 in 100 Mhz junks.
• At each step do 3 one second integrations with cal off and then 3 one second integrations with cal on.
• At each step setup the lband wide filter bank for that frequency range. The frequencies and filters used were:
• 1120-1220 filter 2 (1120-1220)
• 1220-1320 filter 5 (1100-1800)
• 1320-1420 filter 3 (1230-1470)
• 1420-1520 filter 5 (1100-1800)
• 1520-1620 filter 5 (1100-1800)
• 1620-1720 filter 4 (1550-1820)
• Repeat stepping through the entire band 9 times.

• Fig 1: CalDeflection/TsysOnAbsorber spectra. The 3 one second samples at each band have been averaged together. The top plot is polA and the bottom plot is polB. The vertical black lines show the 6 100 Mhz steps (bands) that were taken. There are 9 colored spectra (one for each pass through the 600 Mhz band). Each pass took about 72 seconds. You can see jumps in polB band 3. This is the 1320-1420 Mhz band with filter 3 in. The narrow spikes at 1240 and 1260 MHz are the aerostat radar leaking through the absorber.
• Fig 2: This is fig 1's data with offset placed between each strip for display. You can see the discontinuities between the adjacent bands for polB band 3. Strips 4, 5, 6, and 8 seem to be ok.
• Fig 3: total power CalDeflection/TsysOnAbsorber. The total power has been computed over each 100 Mhz band. The 3 1 second samples have not been averaged. The colored lines are:
• black: 1320-1420 polA (filter 3)
• red : 1320-1420 polB (filter 3)
• blue: 1420-1520 polA (filter 5)
• purple: 1420-1520 polB  (filter 5)
There are 3*9= 27 samples for each freq band. The dashed green is a 72 second break in time. It is the start of each 3 second caldefl/tsys. You can see two things happening with filter 3 polB:
• On a 1 second time scale CalDefl/Tsys it is drifting down. This means that the attenuation through the filter is changing (a loose connection?).
• Each dashed line is a 72 second delay (the next pass through the band). The filter switches have also been deselected and the reselected (since the adjacent frequency bands used different filters). It may be that the switch itself is also not making good contact???

 home_~phil