cband stability (shutter closed)

 The setup was:

• cband receiver centered at 4745.
  
• shutter on cband receiver closed (receiver looks at absorber load).
• 7 iflo bands selected at: 4325,4466,4608,4750,4892,5034, and 5176
• mock spectrometer:
• 172 MHz bw/band with 8192 channels/band.
  
• 1 second dumps
• full stokes
• Two "scans taken"
• scan1 3600 seconds
• scan2 2878 seconds
• 8 seconds between scans.

• The total power was computed for each sample,pol, and band by averaging the 8192 channels of each spectra.
• Each set of 6478 total power numbers was normalized to the median value (this defines Tsys).
• Using the first 3600 seconds of data (contiguous in time) compute the spectrum of the total power to see if there are any periodic signals.
  
• From the  6478 normalized total power dataset, remove a 2nd order polynomial.

• Pages 1-2: total power vs time. No polynomial removed.
  
• black is polA, red is polB.
• Each frame is a separate frequency band.
• The vertical lines mark every 600 seconds (to help align things in adjacent bands).
• total total power varies 0 to 2% of tsys
• Pages 3-4: total power vs time with 2nd order polynomial removed.
  
• Same as pages 1-2 but the 2nd order polynomial was removed. This allows us to blow up the vertical scale.
• Page 5: Spectra of the total power.
• the spectra of the total power was computed for the first 3600 seconds.
• Top frame: polA, bottom frame polB. Each colored plot is a totalPower spectra from a different frequency band.
• The black (4325 band) has a set of harmonics with .13 Hz  (7.69sec period). This may be coming from rfi leaking in.
  

DISCUSSION:

• The total power varies by up to 2% of Tsys over the 6478 seconds.
• Tsys is probably 300 Kelvins with the shutter closed.
• The variation includes the receiver and the iflo.
• Some of the spikes in the total power spectrum may come from rfi leaking through the shutter. There is occasional strong rfi in the 4325 band.
  
• Comparing this variation with that seen on the sky.
• Tsys on the sky is about 30K. Tsys on the shutter is probably around 300K.
  
• If there is a gain variation, then having Tsysshould not make any difference.
• If variations are "extra noise power" added near the frontend, then the Tsys difference may make a difference.
• The dome did not move during the test.
•  If there is a loose connection in the dome, then tracking a source may cause the gain through this connection to vary  more than when the dome sits still.
  

cband page
home_~phil