Intro:

• use the wapp spectrometer with 100 MHz bw/band, 3 level sampling, 2048 channels, full stokes. The data was sampled at 1 second intervals.
• Do a 15 second cal on , 15 second cal off 12 seconds of time from the source then track the source for 600 seconds. The cycle took about 667 seconds.

• Input the cal scan, data scan and convert them from cross correlations to uncorrected stokes.
• Create an rfi mask for the cal using calOn/caloff (this removes the bandpass). Use  corblauto to excluse channels that are more than 3 sigma from the mean.
  
• Call the  corstokes to scale to kelvins with the cals and then phase correct the signal path using the cal data (hi correlated cal). There was no mueller matrix correction since the matrix we have is out of date.
• Convert the data from Kelvins to Janskys using the gain curve for cband.
  
• Computing the total power.
• We have and array of 2048 channels by 600 seconds. Use corrms()  (rms/mean along each chan) to make a mask where there is no rfi. Average the good chanels to get the total power.
  

• First set of plots. IQUV, vs data sample and frequency band.
  
• The horizontal axis is data sample (1 second). I have left out the 67 second gaps between each 10 minute dataset. The vertical axis is milliJy. 
• There are 8 frequency bands with 4 per page.
• The data is plotted I:page1,2 Q:page3,4,U:page5,6, V:page7,8.
• Red vertical lines are drawn at the start of each 10 minute dataset. 
  
• I data
• The median value for I has been removed.
  
• The variation is the za dependence of the gain (the SEFD is higher at large za).
• U,V data
• The 5125 MHz band had a problem with the correlations. There was a  high frequency variation (probably the interleaving for the 100 MHz). You probably should not believe what you see in this band.
  
• 2nd set of plots. Stokes V vs Time at the Bursts.
  
• Each frequency band has been color coded.
• Each frequency is offset for display. The lowest frequency is at the bottom, the highest frequency is at the top.
  
• The horizontal axis is seconds from the first dataset of the nite. The xtitle shows the utc start data,time and seconds for midnite for the origin of the x axis (start of first data set for the night).
  
• A blue dotted vertical line is where the bursts (and the opposite  sense burst) should occur. The phase and period was taken from the 4725 MHz data on 17may08 and 18may08 (ast). The label pulseSecs: xxx,xxx tells when the two pulses occurred from the start of the nites data.
  

A2327: may08 data  (top)

• IQUV vs data sample and frequency band (.ps) (.pdf)
• STOKES V total power vs time at the bursts (.ps) (.pdf)
• The positive burst is in 3 frequency bands (about 300 MHz).
• The Negative burst occurs in maybe two bands. It occurs in bands were the positive burst was not seen.
  

• IQUV vs data sample and frequency band (.ps) (.pdf)
• STOKES V total power vs time at the bursts (.ps) (.pdf)
• The 1st Burst:
• The positive burst lasted for 650 MHz in the data. We don't see the start frequency..
• We see the negative burst start and stop in freq. about 600 MHz.
• The 2nd burst:
•  The positive burst started during the cal off.
  
•  There may be some of the negative burst but it's hard to tell.
  

• IQUV vs data sample and frequency band (.ps) (.pdf)
• The time for the first burst had a barely detectable signal
• A second burst occurred about half way between the 1st and 2nd expected bursts with only the positive pulse.
• A 3rd burst came a little after the expected second burst. It only came in the negative sense.
  
• STOKES V total power vs time at the bursts (.ps) (.pdf)

Summary may08

• Periods:
• 7051.3 seconds or 1.958703 Hrs. (using 17may,18may pulse at 4725 MHz.)
• Time between burst and opposite sense burst: 497 seconds.
• Pulse Start Times.
• the file shows the AST start times for the pulses using the 17may08 and 18may08 positive pulse.
  

A2414 - jun09  (top)

    Looked at tvlm513 7,8,9,10,11,12,13,14,15 using the wapps. See setup for a2327 above.

The type of plots were similar to those of a2327. Some newer plots were:

• When plotting the entire days i,q,u,v:
• For some day a 2nd polynomial was also fit to each 10 minute scan. this helped finding pulses and  interference. If any pulses last for a long time it could corrupt them (so the pulse should be compared to the unfit data).
• When plotting the pulses, the expected pulse arrival time flagged.
• This was computed for the entire set of days using  the first pulse of 08,12jun09 (and knowing the approximate period from 2008).
  

• 07jun09
• IQUV vs data sample and frequency band (.ps) (.pdf)
• STOKES V total power vs time at the bursts (.ps) (.pdf)
• 08jun09
• IQUV vs data sample and frequency band (.ps) (.pdf)
• IQUV vs data sample and frequency band with fit by scan (.ps) (.pdf)
• STOKES V total power vs time at the bursts (.ps) (.pdf)
• 09jun09
• IQUV vs data sample and frequency band with fit by scan (.ps) (.pdf)
• STOKES V total power vs time at the bursts (.ps) (.pdf)
• 10jun09
• IQUV vs data sample and frequency band (.ps) (.pdf)
• STOKES V total power vs time at the bursts (.ps) (.pdf)
• 11jun09
• STOKES V total power vs time at the bursts (.ps) (.pdf)
• 12jun09
• IQUV vs data sample and frequency band (.ps) (.pdf)
• STOKES V total power vs time at the bursts (.ps) (.pdf)
• 13jun09
• IQUV vs data sample and frequency band (.ps) (.pdf)
• IQUV vs data sample and frequency band with fit by scan (.ps) (.pdf)
• This  use a linear rather than  2nd order fit for each scan
  
• STOKES V total power vs time at the bursts (.ps) (.pdf)
• 14jun09
• IQUV vs data sample and frequency band with fit by scan (.ps) (.pdf)
• This  use a linear rather than  2nd order fit for each scan
• STOKES V total power vs time at the bursts (.ps) (.pdf)
• A linear  fit was removed around the pulse
  
• 15jun09
• IQUV vs data sample and frequency band with fit by scan (.ps) (.pdf)
• This  use a linear rather than  2nd order fit for each scan
• STOKES V total power vs time at the bursts (.ps) (.pdf)
• A linear  fit was removed around the pulse

a2430 - dec09  (top)

   a2430 continued a2327 measurements. The mock setup was:

• 7 x172Mhzx1024 channel Full stokes data
  
• 1 millisecond sampling. 10 minutes on source, then calon,caloff (off source).
  

• Src1: 2MASS0036.. No bursts were seen
  
• Src2: 2MASS0746.. bursts were seen on multiple days. Somedays we were unlucky and the burst occurred during he calOn,off
  

Plotting the data by day:

• 02dec09
  
• Src1:
  
• IQUV vs data sample and frequency band (.ps) (.pdf)
• IQUV vs data sample and frequency band with fit by scan (.ps) (.pdf)
• Src2:
• IQUV vs data sample and frequency band (.ps) (.pdf)
• IQUV vs data sample and frequency band with fit by scan (.ps) (.pdf)
• STOKES V total power vs time at the bursts (.ps) (.pdf)
• 03dec09
• Src1
• IQUV vs data sample and frequency band with fit by scan (.ps) (.pdf)
• Src2:
• IQUV vs data sample and frequency band with fit by scan (.ps) (.pdf)
• STOKES V total power vs time at the bursts (.ps) (.pdf)
• 04dec09:
• Src1:
• IQUV vs data sample and frequency band (.ps) (.pdf)
• IQUV vs data sample and frequency band with fit by scan (.ps) (.pdf)
• Src2:
• IQUV vs data sample and frequency band (.ps) (.pdf)
• IQUV vs data sample and frequency band with fit by scan (.ps) (.pdf)
• STOKES V total power vs time at the bursts (.ps) (.pdf)
• 05dec09
• Src1:
• IQUV vs data sample and frequency band with fit by scan (.ps) (.pdf)
• Src2:
• IQUV vs data sample and frequency band with fit by scan (.ps) (.pdf)
• STOKES V total power vs time at the bursts (.ps) (.pdf)
• 06dec09:
• Src1:
• IQUV vs data sample and frequency band with fit by scan (.ps) (.pdf)
• Src2:
• IQUV vs data sample and frequency band with fit by scan (.ps) (.pdf)
• STOKES V total power vs time at the bursts (.ps) (.pdf)

 home_~phil