430 dome PolA Tsys jumping around. sep12

19sep12 data: (top)

• the setup was:
• 422-442 filter installed
• tracking blank sky za 7-10 deg
• 25Hz winking cal used
• mock spectrometer measured 21 MHz bw spectra centered at 432 MHz.
• 2 millisecond dump time to give 10 samples on calOn,calOff
• 1024 channels were taken across the spectrum.
  
• Data was taken for 15 minutes.
• Data processing:
• The calOn and caloffs were separated
• calOn-calOff was computed and then the spectra were averaged in time to 1 second.. 
• For each 1 sec of data, tsys was computed as:
• CalOff/(calOn-caloff) * Tcal
• Tcal was taken as a constant value across the 20 MHz band (we only have 1 measured point)
• TcalA=95.5, TcalB=102.55 K

• Page 1: AvgCalon,calOff, calON=calOff
• The vertical units are spectrometer counts
• Black is polA, red  is polB.
  
• The spike at the center is DC (as well as the lousy response of the mixer amps at DC).
  
• Top: calOn,calOff
• The higher set of lines is the calOn
• You can see more rfi in polA than polB
  
• Bottom: calOn - calOff
• Page 2: Rms/Mean by channel for CalOn, calOff records
• rmsByChan/Mean computes deltaT/Tsys.
• The value is determined by the radiometer equation: sqrt(2)/sqrt(bw*time)
• The value for the measurement were: sqrt(2)/sqrt(21.5e6/1024. * .5) = .0138 The measured value was about .013
• PolA has more interference that polB
• Page 3: Tsys vs  freq and time
• Top:CalOff/calDefl * TsysCal.
  
• The median spectra over the 15 minutes was computed.
• We used a constant value for Tcal (at 430).
• TcalA=95.5
• TcalB=102.55
• There is a Tsys slope across both pols (but more in polB)
• This could be real, or the cal value may not be constant for the 20 MHz (cal coupler?).
• You can see the difference polA, polB (about 10K)..
  
• The daily Tsys measurements are made at 431 +/- .5 MHz.
• Bottom: Tsys Vs time
• The median was computed for each 1 second 21 MHz spectra and then plotted.
• The system temperature is not jumping around during these 15 minutes.

• Top polA
• bottom: polB
• A median bandpass was used to flatten each image.
• PolA has more instabilities (especially at 422=428 MHz .. (the filter?))
• You also see horizontal lines in polA. Must be broad band interference?

TsysA,B TsysA-B for 2012 (top)

• The 430 Tsys is measured at +/-.5 MHz about 431 MHz.
• black is polA, red is polB
• Top: Tsys vs date
• the dates with no data are usually world day transmitter runs.
• Middle: TsysA - TsysB
• Around march Tsys polA decreased by about 5 K.
•  It stayed there till sometime in july.
• during july TsyspolA increased to equal TsyspolB.
  
• Since mar12 there have been jumps where TsysPolA increaes by 10-15 K and then drops back down
• Bottom: TsysA- tsysB blowup jun12-sep12.
• You can see on 18,19sep12 there were a number of Tsys measurements with TsysA 10K higher than TsysB (it then jumped back down).

Summary:  (top)

• Tsys PolA  has been changing -5K and the back since mar12
• TsyspolA also jumps by about 10-15k, stays there awhile, and then returns to original value.
• Monitoring TsysA for 15 minutes showed no jumps (while it was 10 K higher than tsyspolB).
• There is a slope of 7K in TsysA and 17K in tsysB across 422-442 MHz.
• This may be real or it could be that the calValue is not constant across the band.
• PolA has more Rfi than polB
• Dynamic spectra show that the bandpass of polA is not as stable as polB
• For 2011-2012 we have also had problems with the stability of the polA bandpass.