Short duration rfi (see by p2558)

---

Introduction (top)

• 327 MHz rcvr with 57.3 MHz bw centered at 327 MHz. (50 MHz filter in rf)
  
• mock spectrometer used:
• polarization adding
• a/d blanking enabled
• 172.032/3=57.3 MHz bandwidth
• 285.71429 usecs dump time
• Recorded 60 2gb files (each about 67 secsonds long).

Processing:

• Only used the first file to examine the data
• Computed the total power for each dumped spectra (285 usec resolution).
• Looked at all  spectra whose total power peaks were > 1.2 Tsys in the 1st and 93 records (about 60 seconds apart).
• For the entire file, averaged all spectra whose total power peaks were > 1.2 Tsys,
  

• Page 1: total power vs time for the 1st (top) and 93rd (bottom) records.
• each record (row) holds 2555  285usec spectra (about .7 seconds total).
  
• The total power is computed over the 57 MHz spectra.
  
• the total power has been normalized to the median value (making the units Tsys).
• The red * show all the samples > 1.5 Tsy.
  
• The green * show the samples 1.2 to 1.5 Tsys
• Discussion:
• The rfi is impulsive. It usually lasts for less than 1 sample (285 usecs). Looking at the entire file, occasionally rfi > 1.2Tsys can be seen in adjacent samples.
• Page 2:plot spectra  with totPwr > 1.5Tsys in records 1 and 93.
• Top is record 1, bottom in record 93 (spaced by about 60 seconds).
• The colored spectra are:
• red: the spectra with Tsys > 1.5 Tsys
  
• black: spectra before the peak
• green: spectra after the peak
• An offset has been added to each set of 3 for display purposes.
• Discussion:
• The spectra before and after the peak spectra look normal
• The spectrum of the rfi look relatively stable for the rfi peaks in 1 record and for records spaced by 60 seconds.
• Page 3: Average all spectra with totpwr > 1.2 Tsys for the 67 second file.
• 458 285 usec spectra had totPwr > 1.2 Tsys.
• This is out of 260610 spectra or .18%
• The bandpass was removed from the averaged peak spectrum

Summary:

• broadband rfi with short time duration (< 285 usecs) is seen with the 327 rcvr
• For a 67 second sample it occurred .18% of the time.
• The frequency structure of the rfi looks relatively stable over the 67 second measurement.
• Whatever it is, it is not generating a random frequency spectrum. So this probably rules out something arcing.
  

home_~phil