Intro

    On 17dec21 the pulsars B0329+54 and B0833-45 (vela) were observed with the 12meter using the xband receiver. These are the two strongest known pulsars (vela 1000 mJy, B0329+54 200mJ at lband). Ben detected Vela but did not detect B0329+54 (although detection was expected). I looked at the Tsys during the observation to see it there was a reason for this.

Setup:

•  mock spectrometer used to take the data.
• 7 bands of 172.032 MHz/band recorded
• Cfr of bands:
• 8219.84 8360.96 8502.08 8643.20 8784.32 8925.44 9066.56 MHz.
• bands spaced by 141.2  MHz
• nchan: 128 (1.344 MHz per chan)
•  B0329+54
• dump time : 500 usecs
• secsOfdata: 3600 
• B0833-45:
• dump time 200 usecs
• secsOfData: 300 secs.
• Prior to each pulsar acquisition, a 60 second hardware winking cal (25hz) scan was taken.
• It was used to scale the data to cal Units (we have not yet measured the cal value in deg K).
  

---

Standard pulsar processing.

Ben processed both pulsars using the standard psr reduction routines. He detected vela but did not detect B0329+54.
Ben's notes:

---

Looking at the total  power vs time

    I went back and looked at the total power vs time for the pulsar data. Was there a reason why B0329+54 was not detected?

Processing:

    For each pulsar:

• Input the 60 second winking cal spectra and compute the conversion factor from mock counts to cal Units.
• exclude 6% of the band pass on each edge when doing this.
• This data was taken while on the pulsar.
  
• input the psr data.
  
• compute the total power for each band (excluding 6% of the band on each edge)
• average the total power to 1 second.
• This will give Tsys in mock units (since the psr's are so weak).
  
• multiply each band by the cal scale factor.
• This puts the total power (tsys) in cal units.
• We have not yet measured the cal values in deg K (we're having a hard time trying to place the absorber in front of the feed).
• The cal  is temperature stabilized so we are able to compare the results from the 2 pulsars.
• For B0329+54 also compute the total power at the full time resolution (.5 millisecs) for seconds 3000 to 3400 (where ben's plots looked ugly).

The plots show the results of the processing.

Average bandpasses of the data (.ps) (.pdf)

•  This shows the band pass shapes and overlap.

Plotting the total power vs time (.ps) (.pdf)

• Page 1: Tsys vs time,freq
• Black is PolA
• Red is PolB
• Top: B0329+54 Tsys vs time
  
• There is 3600 seconds of data.
  
• The data have been averaged to 1 second.
  
• Tsys is in units of the cal value.
  
• The 7 freq bands for each pol are over plotted.
• The spread for the 7 bands is because the cal value varies across the 1 GHz of data.
• You see some bumps around 2000 seconds.
• Middle: B0833-45
• There is 300 seconds of data.
• Bottom:  TsysRatio:  (B0329+54)/(B0833-45) for each band vs freq.
• The Tsys for vela (B0833-45) is about 3.5% higher than the Tsys for B0329+54.
• This can be explained by the zenith angle differences for the observations
• B0329+54: za=40.3 deg
• B0833=34: za=65.0 deg
• Using the fits of tsys vs elevation (more info) Tsys for za=65 deg should be 4% higher than Tsys at za=40.3 deg
• So the problem in not detecting B0329+54 is not because of the average Tsys.
• Page 2: relative change in tsys vs time
• Black is polA, red is polB
• the 7 freq bands are over plotted.
• The 1 second tsys samples have been divided by the median value for each pol and band
• Top: B0329+54
• there is 3600 seconds of data.
  
• you can see increases in Tsys (about 1%) that lasted for about 100 seconds starting at 2000 seconds.
• Pol B Starts trending down after the first jump. PolA is more stable
• around 3000 seconds you start to see more junk (this was seen in ben's plots).
  
• Bottom: B0833-45
• there is 300 seconds of data
• Tsys is relatively stable for this time period
• there is 1 freq band that wanders a bit. This may be rfi.
• Page 3: hi time res blowup of B0329+54 from 3000 to 3400 seconds.
• Top: 1 second resolution blowup
• You an see downward going glitches that may last for a few seconds.
• It happens to all 7 freq bands to the rf is probably getting compressed.
• Middle: blowup 3240 to 3260 seconds at .5 millisecond resolution.
• You can see the negative going spikes
• Bottom: blowup 3255 to 3258 seconds.
• the dropouts are lasting for 100 to 200 milliseconds.

---

Summary

• Psr B0329+54 and B0833-45 were observed at xband
• 300 seconds for B0833-45 and 3600 seconds for B0329+54
• B0833-45 was detected
• B0329+54  was not detected.
• The ratio of the average Tsys values for the 2 observations matched what was expected from the 2 zenith angles.
• B0329+54
• showed positive going tsys bumps  of about 1%  starting at 2000 seconds.
• after the first bump, polB tsys started to drift down by about 1.5%.
• If the positive going bumps were weather related, then the negative drift may be caused by the match at the amp input deteriorating.
• Around 3000 seconds into the scan compression dips were seen in all bands.
• During this observation, the sband rf signal from the 12meter was being passed down the rf fiber with the xband signal.
• It was probably causing the compression (even with the 20db pad in from of the rf chassis in the pedestal).
• B0833-45 (vela0
• Pol A was relatively stable for the 300 seconds.
  
• Pol B had a bit more drift.
• one freq band had more drift (rfi?)
• B0833-45 looked a lot more stable with no compressional dips.. probably because the data only lasted for 5 minutes (vs 1hour)
• We'll put the 50ohm terminators in the sband rf signal back in so it won't affect the xb data  (until we can get a new sband filter).
• This should get rid of the compression.