Intro:

• 4 azimuth swings were done (2 in the morning, 2 in the afternoon) to see if there were any tv station intermods with the filter removed.
• Sitting at a single azimuth, 50millisec spectra were taken to look at the time dependence of some of the rfi.
  

Setup and processing

 The setup was:

• az swings from 270 to 630 and back.
• 2 swings were done around 11am
• 2 more swings were done around 14:00 hours. The sun would move between these sets (so it would show up at different azimuths).
• The dome was az 18 degrees za.
• the azimuth moved at .35 deg/sec
• The iflo was setup at:
• cfr 430 MHz, 750nb if.
• The mock spectrometers were used to record the data:
• cfr 430 MHz, 34 MHz bw, 8192 channels, 1 second sampling.

Processing the data:

• For each swing
  
• a median bandpass was computed and then a robust fit was made (linear fit and 9th order polynomial).
• The spectra were divided by the fit and then the median of each spectra was removed.
• The x dimension was compressed by a factor of 5 , the yscale was compressed by 2.
  

Dynamic spectra were then made for each az swing.

• the horizontal axis is frequency
• the vertical scale is azimuth position
• The graph at the bottom of each image is the peak value for each channel.
  

Notes on the images:

• az 120-130 swings 1-3 show wide rfi (> 5MHz)  centered near 430 MHz.
  
• It is much stronger in polA (with no filter). You can still see it a little in polB (with the filter).
  
• This is close to the azimuth's that that the analog tv's showed intermods
• the digital tv is now spread over 6 MHz so the intermods will be a lot wider than the analog tv intermods.
• -30 -> 0 swings 3,4. this may have been the sun.
• sw1,2 442.5 birdie drifting down in freq.
• sw 1 442.5 -> 441.9
• sw2 441.9 -> 441.5 (this went 270 -> -90)
• sw1 large, wide rfi at az=240. This occurs at different azimuths in different swings.
• Probably some out of band signal

Tv station intermods in polA.

    PolA had the filter in front of the dewar removed. In the dynamic spectra you can see >5MHz interference between 120 and 130degs az (swings 1-3). Swing 4 had the interference at az=110. This rfi was seen (but much weaker in polB).
       PolA intermod (spectra) for azimuths 123-128 (.ps) (.pdf)

• Processing:
• spectra taken between az=123.13 and 128.73 were plotted.
  
• the spectra were first smoothed by 1 MHz.
• the plots:
• The solid line is polA, the dashed lines are polB
• Page 1-4 show the 4 swings
• for swing 4, the intermods occurred around az=110 (not shown)
  
• The peak is shifting in frequency with azimuth (looks periodic).
  
• the 3 swings have the highest freq near az=126.

• Old analog stations had intermods between:
• chn14 - 471.25 and chan 22 519.25 (videos carriers).
• Currently the strongest digital channel is chan 15 (479 MHz center of band).
• This is currently making an intermod with channel 23 (527 MHz)
• 2*479 - 427 = 431 MHz.
• Unfortunately, the notches in the cavity filter do not cut off very much of these signals.

Average, peak hold spectra for swings:

• PolA and B have been averaged.
• Top: average spectra for each az swing.
• 2nd: blowup of average spectra
• 3rd: peak hold by freq channel for each az swing
• bottom: blowup of peak hold.

Plotting the azimuth dependence of individual birdies

• Each az swing is color coded: CW1: black,CCW2:red, CW3:green, CCW4:blue
• for a constant rfi at a particular az, the signal will appear at the same az for all 4 swings.
• The table shows the az dependence of the various birdies:

• freq	azPeaks	strength Tsys	notes
  423.7	180	1.5	only swings 1,4
  424.14	none	2.5	in the dome?
  425.0	10,160	.8,.2	10deg swings 2,4 160 all 4.. maybe 2 birdies?
  425.98	-35	.4	all
  427.63	150	.4	all
  431.3	-15	2.4	all
  431.87	135	4.6	swings 1,2
  432.01	111	.2	all, some glitches 180
  433.55	195	34	all, 1 second period angel says this is ham slow scan tv.
  433.78	random	>1
  433.92	random	>2
  434.0	random	>2
  434.17	random	2
  434.79	52, 180,200,220	40, 30	also has time periodicity
  435.01	160,-20	.5,.3	peaks 180 deg apart
  436.41	random	.3
  437.27	random	1	only swing 1
  437.3	randome	>5	mainly swing 1
  437.5	-10	2	all
  438.73	random	>.4
  439.42	random	>1
  439.48	random	>2
  440.0	random	>10
  440.6	random	>1000	probably ham
  442.37	-4	15
  442.40	110	4	power variation through 4 swings (with same az dep).

Over plotting birdies with similar az dependence (.ps) (.pdf)

• Page 1:rfi peaks near az=-20
• the control room is probably centered at -35
• Some of the birdies have weaker peaks near az=160. This is 180 Degrees from the main peaks.
• Page 2:
• top: 423.7, 432.01.
  
• These had similar peaks near az=180 during 1 single swing.
• Bottom: birdies with time variation
• 433.55 has a 2 second period. 434.79 has a longer period
• They don't peak at the same azimuths.

Birdies with time dependence:

• Telescope sitting at az=260,za=18.
• PolA had the filter removed from in front of the dewar.
  
• the mocks took 34 MHz Bw centered at 433 MHz using 8192 channels and 50 millisecond integrations.
• The two pols were averaged together.

• there is large rfi between 80 and 120 seconds.
• periodic data is seen at:
• 433.54
• 434.00
• 434.17
  
• 434.67, 439.78
  

• page		freq	period	notes
  page-1	top	433.55	1.00	az dependent with peak at az=195,..30Tsys
  	2nd	433.55	1.00	blowup, duty cycle=12%
  	bot	434.00	57.76	no repeatable az dependence seen. Probably do to the short duty cycle.
  page-2	top	434.71	-	starts at 16.26 secs but then period varies no repeatable az dependence.
  	2nd	434.68	10
  	bot	434.78	6.35	434.78 and 434.79 vary together. az dependence: 30,50,70 deg peaks az:180,200,220
  page-3	top	434.79	6.35

• 
  

Summary:

• digital channel 15 (479 MHz), chan 23 (527 Mhz) cause intermods in the 430 system
• When the filter in front of the dewar is removed, the intermods are strong.
• When the filter is installed (polB) there is still a hint of the intermods.
• The strongest tv stations changed freq with the move to digital tv. The strongest channels were/are:
  
• analog: 14,22
• digital: 15,23
• The filter notches no longer cut out the strong digital channels completely.
• a number of the birdies have a repeatable azimuth dependence
• 5 birdies also have a periodicity.
• 440.6:  4000Tsys. we need to look into this birdie. it is a  ham.
• 424.16: Not much change with azimuth. We  should look in the dome for this.
• 430 carriage house.
• when the digital tv stations  turned on, we started to see intermods in  then 430ch even with the single pole notched filter in front of the dewar.
  
• For PolA, we replaced the single pole notch filter with a 4 pole bandpass filter.
• The problem was that the notches were no longer doing a good job cutting off the strong tv stations (they moved by 1 tv channel).
• the 4 pole filter has more loss than the single pole notch filter, but it does a better job of cutting off all of the tv stations.
• back then, we didn't see strong intermods in the 430 dome with the notch filter in place. This is probably because the carriage house has larger sidelobes than the dome.
• We only changed polA, since the aeronomers only use polA.