20mar14: resolution:

• the drifting birdie was coming from the sbn postamp chassis. it was brought down to the lab 20mar14 15:00 to be worked on.
• The problem is the same problem that we had back in jun 2004 with drifting birdies in lband (more info)
• the cal control board in the postamp chassis used some hi speed chips.
  
• Some of the gates in the chips were unused and were left open. They then oscillated (many volts).
• We supposedly  fixed all of these cards back in 2005 by putting terminations on the unused gates.
  
• The problem may be that when some cards went bad, they were replaced with new cards that did not have the terminations in place.
• We need to make sure that there are no other postamp chassis's that have these cards with no terminators in them.
• We also to need ensure that this doesn't happen a third time.
• We spent 1 month of telescope down time using the 327 receiver for pulsar search.. with this birdie drifting all over the place.
  

---

•  p1693 took 327 MHz pulsar search data during the M8-4 cable repair.
•  A birdie was seen to slowly drift from 319 -> 325 MHz over 24 hours. (see dynamic spectra of drifting birdie).
• The 24 hours of data from 23Feb14 shows a complete cycle in 24 hours (.gif)
• the drifting is continuous. The image was made using the first two minutes of each 1 hour scan. There are 129 rows in the image (2 minutes) and then a jump in time of  1 hour).
• The birdie started near 325 MHz at 7:30 am on 23Feb14.
  
• it went down to 319 around 17:00
• it then drifted back to 325 MHz near 6am on 24Feb14.
  
•  The birdie strength would get up to 20 times Tsys in the 57KHz channels of p1693.
• An examination of the p1693 data on 20Feb14  looked like the rfi was compressing the total power every 1.5 seconds (more info)
• high frequency resolution data taken on 27Feb14 showed that the band was not compressing.

On 27Feb14 voltage sampled data was taken using the mock spectrometer. The setup was:

• 327 receiver.
• mock spectrometer with 2 bands both centered at 325.73440 MHz
  
• band1 : 1.6 MHz bandwidth
• band2 : .160 MHz bandwidth

Dynamic spectra from 8 seconds of the 1.6 MHz band (.gif)

• The frequency resolution is 195Hz, about 300 KHz is displayed
• The birdie is about 30 kHz wide
  
• You can see the birdie jumping every 1.2 to 1.5 seconds.
• It jumps by about 20 KHz.
  
• The total power of the rfi was computed between the dashed lines.

The plots show the birdie strength and time variation (.ps) (.pdf)

• page 1:average spectra
•  top - 1.6 MHz average spectra
• bottom: .16 MHz average spectra
• Page 2:
  
• top: time variation of the birdie.
  
• This integrates the spectral channels between the dashed lines of the dynamic spectra (see above)
• The dips occur when the rfi jumps in frequency
• middle: the total power in the birdies was fft'd and the squared to show how the birdie changes with time
• each spectra lasted for 5 milliseconds, so the max frequency was about 100 Hz.
• bottom: blowup in frequency
• the largest spike was around 1.22 Hz. this is close to the crosshead frequency.
  
• It may be that the jumps are not evenly spaced in time, so the cross head frequency power may no be related.

summary

• There is a strong drifting birdie seen in the 327 receiver during the p1693 pulsar search observations.
• It was seen 28jan14. I don't have date before this online.
• over 24 hours the birdie drifts from 325 MHz down to 319 MHz and then back to 325 MHz.
• The birdie is about 30 kHz wide. It jumps by 20kHz
  
• The jumps are spaced by 1.4 and 1.6 seconds.

processing:x101/140227/rfi327.pro