• mock spectrometer. 5 x 1.6 MHz by 8192 channel bands centered at:
• 1217.6,1339.4,1461.2,1583.0,1704.7
• This is the comb seen in the spectrum analyzer data.
• 1 second dumps, 195 hz resoloution
  

• processing: flatten image with median off spectra. 1.6 MHz bw, 8192 channels.
  
• Top: 1217.6 MHz band
• Bottom: 1339.4 MHz band
• didn't see anything in the 1461,1583 and 1704 mhz bands
• The yellow lines show when the cpu was turned on, green lines show when the cpu was turned off.
• The tone drifts by 2 to 3 khz in about 1 minute (after turn on).
• There is a delay in the rfi appearing after turn on.
• The birdies are spaced by about 65 khz.
  

• Page 1: average on,off spectra
• red cpu On, black: cpu Off (off overplots on).
• Top: 1217.6 mhz band
• Strongest birdie: 1217.6307 MHz
  
•  bottom: 1339.4 mhz band
• Strongest birdie: 1339.3937 (near DC).
• Page 2: blowup of strongest birdies
• The width is caused by the frequency drift.
• The veritical scale  is Tsys (in 195 hz channels).
• Page 3:Plot birdie channels vs time
• top: 1217.6307 birdie (avg about 2khz about birdie center)
• bottom: 1339.3937 birdie (avg about 3khz about birdie center)
• The cpu on, off times are flagged. You can see the delay from cpu turn on to birdie appearance.

see also:globe rfi measured using the spectrum analyzer

• lbw with cfr of 1485.4 MHz
• This is the stronger tone seen in the spectrum analyzer data.
  
• mock spectrometer with:
• 4.3 mhz bw, 8192 channels, 1 sec dumps.
• Turn the cpu (controlling the large display) on and off.
  

• processing: flatten with the median of spectra
• The green lines are cpu off, the yellow lines are cpu on.
• There is a delay of about 45 secs from cpu on to the rfi appearing.
  

• page 1: average on,off spectra
• Top: on,off over plotted
• red: cpu on, black : cpu Off.
• the birdies around the band center are baseband birdies from the mock spectromter (is didn't use the digital mixer for this band).
• bottom: blowup of the birdie
• the birdie is at 1485.0156 MHz. We did not resolve it in the 525 hz channel width (which  means it didn't drift much in the 110 secs the signal was present)
• Page 2: birdie power vs time.
• i averaged 3 channels centered on the birdie
• The colored lines show when the cpu was turn off, on.

processing: x101/160509/vc_exhibitsrfi/rfvVc_atmrdrlbw.pro

• mock spectrometer. 53 MHz,cfr=327,.1 sec dumps,8192 channels
• felix plugged, unplugged the turntable motor while data was taken
  

• processing: flatten with median bandpass.
  
• green lines, things are off, yellow lines motor on
• The pointing motor was turned on,off at sec 420-460
• Nothing obvious jumps out of the image..
  

• Page1: average spectra with motor on, off
• black is the average spectra while motor off
• red spectra is with motor on.
• I plotted the black trace on top of the red trace.
  
• Each has about 300 seconds
• the baseline was removed using a 19th order harmonic fit to the average spectra. The ripples come from this fit.
  
• Page2: total power vs time:
• the power was averaged over 330.5 to 338 MHz
• The time resolution is .1 seconds
• I've flagged where the motor was turned on,off
• The noise doesn't change appreciably when the motor is on. 

• 327: turn kinect on,off
• lbw: turn kinect on, off (2 wide bands, 1 narrow band)
  

• .1 second integrations.
  
• The median spectra with kinect off was used to flatten the image
• The green lines show when the kinect was unplugged from the cpu
• the yellow lines how when the kinect was plugged into the cpu.
• There is a delay from plugging in, to it being active.
• Nothing obvious is seen in the on's vs the offs.

• Page 1: average spectra with kinect plugged, unplugged to cpu
• A 19th order harmonic was fit to the on and off averages and then used to remove the filter shape. This caused the ripples you see in the spectra.
  
• black: kinect unplugged, red kinect plugged in. The cpu was left on.
• The 53 MHz band is split into 3 sections.
• nothing obvious between on,off. The dynamic spectra shows that some of the on spikes are just random during the on.

• 1 sec, 160 MHz dumps taken at 1550, and 1650 MHz.
• The 1550 MHz had too much rfi to use
• the green, yellow lines show when the kinect was off,on.
• nothing sticks out of the on,offs.
  

• Page 1: average spectra with kinect plugged, unplugged to cpu
• I used the 160 MHz band centered at 1650 MHz (same as image)
  
• A 19th order harmonic was fit to the on and off averages and then used to remove the filter shape. This caused the ripples you see in the spectra.
  
• black: kinect unplugged, red kinect plugged in. The cpu was left on.
• The 160 MHz band is split into 3 sections.
• nothing obvious between on,off. The dynamic spectra shows that some of the on spikes are just random during the on.
• Page 2: total power average (1650.1 to 1670 MHz) vs time
• Data was sampled at 1 second time resolution
• The kinect on,off are flagged.

• 327 receiver
• mock spectrometer, 53MHz bw, 8192 chan,  1 sec dumps
• start and top moving the the azimuth motor
• data was taken near noon, lots of 1 MHz standing waves from sun.
  

• 1 second integrations
• green,yellow lines show motor stopped, moving.
• ripples across band are 1 MHz standing waves from the sun.
• Can't see any difference motor stopped,moving.
  

• Page 1: average spectra with az motor moving and stopped
  
• A 19th order harmonic was fit to the on and off averages and then used to remove the filter shape. This caused the ripples you see in the spectra.
  
• black: motor stopped, red motor moving.
  
• The 53 MHz band is split into 3 sections.
• nothing obvious between on,off. The dynamic spectra shows that some of the on spikes are probably from the standing waves changing.

• spectrum Analyzer:
• 121.763 MHz comb
• 1217.6307,1339.3937,1461.1567,1582.9197,1704.6827 MHz
• Telescope:
• 1217.6307 MHz, .3 Tsys in 195 Hz channel.
• weaker 65Khz sidebands about this
  
• 1339.3937 MHz, .03 Tsys in 195Hz channel
• birdies drift by 2-3 khz/minute after turn on.
  

• Spectrum analyzer
  
• 1039.8,1183.5,1485.0156,1643.6
• 100 MHz noise bump seen beneath the birdies (may be unresolved sidebands).
• Telescope:
• 1485.0156 MHz , .23 Tsys in 525 Hz channel (not resolved).
  

• spectrum analyzer
• 79.198 MHz comb
• 1483.6 birdie.
• same as atm radar display (since same monitor), and it is stronger.
• Telescope:
• Not measured, but will be visible since stronger than atm radar birdie
  

• spectrum analyzer.
• 73.272 MHz comb.
  
• 1485.4 birdie (not part of comb). close to atm Radar birdie.. but weaker.
• Telescope:
• not meaured. 1485.4 birdie may or may not be visible in lbw.
  

• Spectrum analyzer
• 0-1ghz, 3MHz rbw
  
• spikes covering entire band. Up to -30dbm
• 327MHz  3MHz rbw, time domain
• spikes not resolved in 18 usec sampling
• Telescope:
• nothing seen in 327 band
• .1 second sampling..
  

• Spectrum analyzer
• 0-1 Ghz
•   birdies present,  but no obvious ones in 327 or 430 bands
• 1-2 ghz
• Lots of birdies. When recording, it looked like the rfi might be swept in frequency.
• Telescope:
• 327 receiver. Nothing obvious in .1 sec: spectra, or total power
  
• lbw receiver. nothing seen in 1 sec: spectra, or total power
  

• Spectrum analyzer
• 1-2 Ghz  3MHz rbw
• spike seen across band. up to -62dbm.
• Telescope:
• 327 receiver. Nothing seen in 1 sec spectra