---

links to sections
the problem
setup
processing

21apr18 polN comb and polA standing wave:

• Comb seen in polB lband and cband
• Large standing wave in polA.. found loose IF cables at puppi. tightened 30apr18.
  

30apr18 polB comb

• Only cband data taken. Comb seen in polB
• Plots:
  
• the spectra averaged over each file (.ps) (.pdf)
• Dynamic spectra for Cband data (.png)
• gpu power vs time (.ps) (.pdf)
• the rms/mean by freq channel (.ps) (.pdf)

---

The problem:

    P3054 (coherent search) data taken 21apr18 and 30apr18 showed an 8 channel comb in polB. It was present in  cband data (which uses the new puppi mixer) and in lband data (which does not use the new puppi mixer). The comb sits in the same frequency channels in both lband and cband

---

p3054 Experiment setup:

Puppi setup:

•  coherent search, 512 channels, 800 MHz bandwidth, 10.24 usecond sampling, dm=557,  polA,B,CrossReal,CrossImg
• coherent search will split the 512 channel into 8 sets of 64 channels and send each set to a gpu to do the coherent dedispersion.
• The blocks are later combined back into a 512 channel spectra.

Receiver setup:

• cband:
• 4101.56  to 4900 MHz
• Dana's up,down converter is used, so the entire  800 MHz bw is sent to puppi.
• lband:
• 918.562 to 1780 MHz
• Dana's up,down converter is not used.
• the upper edge of the rf filter and the lower edge of the 1-2 GHz IF filter define the bandwidth
• puppi's input is 800 1600 MHz IF (center at IF=1200MHz --> 1380MHz rf).
  
• the 1GHz IF filter cutoff will start to cut off the rf below (1380-200)=1180 MHz rf.
  
• So there should be very little power from rf freq 918 to 1150MHz  because the 1GHz IF filter will cut it out.
  

---

30apr18: p3054 cband data

    cband coherent search data was taken on 30apr18. This was after the IF cables to puppi were tightened.

I looked at the file:

• puppi_58238_B0540+23_0598_0001.fits (41GB, 214 seconds).
• coherent search, 10.24 usec sampling 512 x 4 (polA,B, crossR,crossI), dm=557, center freq 4500 MHz using puppi up/down converter.

Average spectra, acf

The first set of plots show the spectra averaged over each file (.ps) (.pdf)

• Page 1:cband spectra averaged over file (214 seconds)
• color usage:
• Black is polA, red is polB
• the blue dashed lines show the 64 channel boundaries of the gpu processing
• Top: cband average spectra
• There is a now no ripple in  PolA
• bottom:  ACF shows shows distance to reflections
  
• The fft of the spectral density function gives the acf.
  
• Any ripples in the spectra will become spikes in the ACF at the delay for the reflection.
• The x axis is labeled: C/2 *.7 *delay.  I'm using .7 for the velocity in the cables (could be different).
• The 7.8 meter reflection in polA is no longer present.
• Page 2: cband band pass corrected spectra vs freq shows comb in PolB.
  
• to look at weaker, narrow spikes, a bandpass correction was computed and then divided into the polB avg spectra.
• a 43 order harmonic  fit to the average band pass was used as the bandpass correction.
• This will  leave interference that is narrower than 67 MHz.
• You can still see the comb across the polB band.
  
• Page 3:cband polB  Bandpass corrected spectra vs puppi channel number
• The band pass corrected spectra are plotted vs puppi channel number (0..511).
• Green shows the comb spacing (8 channels 12MHz)
• Top Frame: channels 0 to 255 are plotted
• Bottom Frame: channels 256  to 511 are plotted.
• You can clearly see that the comb lines up at 8 channel spacing for  cband.

Dynamic spectra from .3 sec averaged spectra

• A median band pass was used to flatten each data set
• The PolB comb is mostly removed.. although some of it can still be seen (which implies that the amplitude to not perfectly stable in time.
• For each image, polA is shown on the left, and polB on the right
• Vertical red lines show the gpu boundaries
• the gpu sections are labeled at the bottom

Dynamic spectra for Cband data (.png)

• You can see a little of the 8 channel comb on the right side of polB
• dividing by the median, removes the stable portion of the comb.
  
• The total power changed a little with time.. going from dark at the bottom to lighter a the top.
• The telescope moved za 6.7 to 7.3 degs during the 214 seconds. So there shouldn't have been much change in spillover

Total power vs time    

The average power in each gpu band was computed (at the .3 second row time resolution) and then plotted vs time.
    
gpu power vs time (.ps) (.pdf)

• Each gpu band is plotted using a different color.
• The top Frames are polA, the bottom frames are PolB
  
• Page 1 Cband power vs time puppi units.
  
• the power is average over each .3 second row.
  
• each plotted power is labeled with the gpu name and the center frequency of the band
• gpu's 1,2,3 have values < 15 .. see the average spectrum bandpass shape.
• You see a slight drift in the power level over time.
  
• Page 2: total power normalized to each channels median value
• The the total power in each gpu was normalized to the median value over the 214 seconds.
• The power drifts +/- 2% during the 214 seconds.
• Gpu1 is a lot noisier that the other bands.
• This is probably to be expected since the average value of this band was only 6 counts.
• the bandpass shape is cutting off too much at this edge of the band.

rms/mean for a .3 second row.

    The first record of the file was used to compute the rms/mean by freq channel (.ps) (.pdf)

• there were 32256 10.24 usec spectra in the record (row).
• The rms/mean was computed for each freq channel.
  
• with 800MHz and 512 channels, each channel is 1.5625 MHz
• From the radiometer equation this should equal 1/(sqrt(bw*time)
• Black is polA, red is polB, blue is the expected rms/mean.
• the rms/mean is increased for the 8 channel comb.
• The expected rms/mean is close to the average value
• any change in the total power during the .3 secs will increase the rms/mean.
  

Summary:

• The polA standing wave is not present after tightening the puppi IF cable connectors
• The 8 channel polb comb is still present.
• there is a +/- 2% change in total power over the 214 second observation.
• gpu1 (4850MHz) is a lot noisier than the other gpu bands.
• looking at the average bandpass, this band only had around 6 counts in the spectra.
• the bandpass shape is cutting off too sharply on this edge of the band.
• Computing the rms/mean for 1 .33 second record:
• The average rms/mean is close to the expected value.
• the 8 channel comb has an elevated rms/mean.
  

processing: x101/180430/p3054rfi.pro