A2335 lband wide instabilities sep10

07oct10:

• lbw in linear pol mode
• the 1280 hi pass and 1230 to 1470 bandpass rf filters were used.
• The tiedowns were left tracking during this observation. This may have increased the 1 Mhz   ripple in the spectrum.
  
• 300 sec on,off position switching (1 240 second on, off was also done)
  
• freq around 1360-1380 MHz

• src	onoff #	band1	band2	band3
  10863	1-2	1383	1365	1385
  11132	3-4	1373-1376	1365	1385
  11727	5-7 5 is 240 recs	1376	1365	1385
  11754	8-9	1365	1365	1385
  11815	10	1354	1365	1385

  
  
• interim correlator, 9 level, 1 second dumps with 4 boards:
  
• brd 1  band 1 12.5 MHz polA
• brd 2  band 1 12.5 MHz polB
• brd 3  band 2 25 MHz polA,polB
• brd 4  band 3 25 MHz polA,polB

• total power vs time (band1) and attenuation vs time (.ps) (.pdf):
• Page 1: total power vs time band 1.
• Light blue: start of position on scan, purple start of position off scan
• The source name is shown at the start of each on (dark blue)
• top : polA pwr vs time
• 2nd: polB pwr vs time
• 3rd: za vs time
• 4th: az vs time
• All on/offs are 300 secs on,off except the 2nd 113211 which was only 240 seconds on,off.
• During the on or off the power should be stable (except for any tsys za dependence which should be smooth)
• Problems:
• 11754 and 11815 (last 3) show power variations during the on and off. There is a small change in the last 11727 (4th to last)
• the changes occur in pola and in polB.
• There is a large power jump between 11727 and 11754.
  
• the za change is minimal.
• band 1 polA,B had no attenuation changes between these scans (see page 2).
• Looking at nvss, there is no large continuum sources in the field.
  
• Page 2:  attenuation vs time.
• top: rf attenuators. Constant at 0 for the run. The upstairs if attenuator was also constant.
  
• 2nd: Downstairs if2gain. polA constant except for a 1db step in the last on,off. PolB constant for the night.
• 3rd,4th: polA,polB cor attenuators for the 3 bands.
• band 1 constant
• band 2: polB constant. polA 1db step last on,off
• band 3: polB constant. pola 1db step 3,2
• So band 1 gain is constant except for polA in the last on,off.
  
• total power on,off over plotted, spectra of on/off-1 for bands 1,2,3:
• Page 1. total power vs time.
  
• i excluded the 240 sec on,off.
• on source is plotted in black, off source in red.
• top: polA on,off, 2nd polB on,off.
• 3rd: polA,B on source, 4th polA,B off source
  
• band 1 (.ps) (.pdf)
• band 2 (.ps) (.pdf)
• band 3 (.ps) (.pdf)
• The variations start in the last onoff of 11727 (4th from last).
• The variations are the same in polA and polB
• The variations occur in all 3 bands.
• The change is up to 2% Tsys in 5 minutes
• Page 2: on/off-1 for the 5 sources. For sources with more than 1 onoff, they are over plotted.
• The vertical scale is degK.
• the level changes for multiple onoffs on the same source. Implies the gain changes between on,off and cals.

Summary:

• lbw output power is varying by up to  2% in 5 minutes.
• The on/off -1 showed some 1 MHz ripple. This could have come from tiedown motion (they were not disabled.).
  
• The variation occurs in all 3 bands. This implies
  
• It is not a correlator board since it is in all 4 boards.
• It is before the 2nd mixers since it happens in all 3  bands.
• It is the same in polA and polB. These are different paths through the if/lo except:
  
• the LO for the  first mixer is common to polA and polB.
• The program ran in linear polarization mode. If the linear to circular chassis was not switching this could also affect both pols.
• Some chassis that process polA and polB have common power supplies.
• It has not been reported in any other receivers (although i doubt if anyone has looked close enough recently).