23jan16: cal spectra come out rotated.

• 90 sec winking cal (40ms) scan. folded at cal period  to  give 10 records
• long scan of 8 bit search data..
• repeat the  90 sec winking cal scan

The winking cal used the hardware 40 ms cal.

Plotting the data showed that the cal scans were wrong (.ps) (.pdf)

• page 1: frames 1-4. stokes i,q,u,v spectra averaged over the 90 seconds.
• This included the cal on and cal off
• the spectra are not centered at 430 mhz.  This is independent of the cal  on,or  off.
• Bottom frame: the total power for each stokes I spectra vs  time
• You can see the cal turning on,off with a 40 ms period.
• the red lines show the cal period..
• These are the 10 folded spectra.
• Page 2: rotate the cal spectra by 512 channels.
• the data acquisition took 2048 channels, and then recorded 1024 of these channels.
• Top Frame: I  rotated the averaged i,q spectra by 512  channels and then over plotted them. You can now see the receiver bandpass.
• Bottom frame: the  u,v spectra were rotated by 512 channels and then over plotted.
• Page 3: 1.6 sec average spectra from the search data.
• frames 1-4 show the i,q,u,v averaged spectra. They are aligned correctly.

Summary:

• both of the 90 sec  cal scans returned spectra that were not registered correctly in frequency.
  
• This must be a problem with the puppi setup (the if/lo  places the center of the  band at 750 Mhz)..
• The search data were recorded correctly.
• I  checked the cima _server log for this date and did not see any errors reported during the configuration.
• During test time we should repeat this setup and check that the cals are registering the spectra  correctly.