18-26sep11: 1408 rfi in a2010 data
29sep11: 1408 rfi coming from AC unit on top of building I

---

rfi Alfa data 18-26sep11:

   Project A2010 (alfalfa) reported rfi at 1408 MHz on 18sep11. The data setup was:

• alfa receiver.
• 600 second drift scans
• 1 second dumps of 100 MHz by 4096 channels (24KHz resolution) centered at 1385 MHz

Processing the data:

• I processed one nights data at a time. The observations went  from 20:00 to about 3:00 hours.
• For all files for a night  i kept 101 channels about 1408 MHz.
• For each second,  the power at 3 channels (centered at 1408 MHz) was averged  and then it was normalized by channels [0:29, 71:100] from the 101 channels saved about 1408 MHz.
  
• Plots were made of the total power at 1408 MHz (channels) vs time
• Dynamic spectra images were also made for the  101 channels about 1408 for the entire night.
• The median spectra was removed from each night so constant values will be removed.
  

The plots,images

date	1408Pwr VsTime	Dynamic  spectra	Notes
18_19sep11	(.ps) (.pdf)	polA   polB	3b strong
19_20sep11	(.ps) (.pdf)	polA   polB	3A strongest
20_21sep11	(.ps) (.pdf)	polA   polB	no periodicity
23_24sep11	(.ps) (.pdf)	polA   polB	no periodicity
24_25sep11	(.ps) (.pdf)	polA   polB	no periodicity rfi on for longer periods of time
25_26sep11	(.ps) (.pdf)	polA   polB	6B strongest 4b continuous toward the end

18sep11 50 second average spectra when 1408 birdie on (.ps) (.pdf)

Notes:

• beam 4a was turned off for these measurements.
• The 1408 pwr vs time shows periodicity in the rfi.
• the pulses are space by 32 minutes drifting down to 15-20 minutes during the 5 hour run.
• When pulsing the rfi stayed on for about 230-240 seconds.
  
• The dynamic spectra show that sometimes the signal remains on for long stretches of time. When this occurs, you do not see the 1408 tp moving up and down.
• The 1408 birdie covers 3 24 khz channels:
  
• 1408. to 1408.47 with a dip in the middle.
• The high frequency ripple (see 20_21    beam3b) is probably coming from the fiber optics. A 174 khz ripple is created by reflections from the two ends of the fiber optics. The dynamic images only show the ripple of it changes (we divide by the average bandpass).
  

Things to check:

• Monitor 1408 till you see it turn on.
• Then switch in the 1370-1470 MHz filter. If it goes away, it is probably an intermod in our system.
• Az swing and see if it changes with az direction. Problem is that it is also time variable.
  

processing: x101/110918/a2010proc,a2010plot, procdat,plotdat.pro

---

29sep11 Opening, closing AC doors on building I.

    An azimuth spin on 28sep11 showed that the 1408 MHz rfi peaked at az=338.1, za=15. After some snooping with the spectrum analyzer i found that the rfi was coming from the new AC unit on top of building I (by the  ladder).

    On 29sep11 I took some data (using lbw) with the doors to the AC unit closed, open, and then closed again.
The setup was:

• az=338.1, za=15,lband wide receiver in linear pol mode.
• Mock spectrometer with 8192 channels over 4.4 MHz (538 chan resolution), 1 second dumps.
• Data collection was started and then:
• went to the roof and opened the doors for about 240 seconds.
• I used my cell phone time to mark when the doors were opened and closed. Unfortunately it didn't display seconds so I only had the time to minutes (must do that when it is in airplane mode ???).

The image shows 590 seconds of data while the doors were opened and then closed (.gif):

• The horizontal axis is frequency (but it is flipped). The vertical scale is seconds.
  
• The door was opened around sec=250 (vertical scale).
• The door was closed around 480 secs.
• There was  a jump in the strength of the signal around 60 secs (maybe it turned on??). The large jump in strength came around 250 seconds (the image stretch makes the strongest birdie appear saturated).
• When the door is opened, the entire noise bandpass increases. This was also seen with the spectrum analyzer on the roof of blding I.
  

The plots show the strength of the birdies(.ps) (.pdf):

• Page 1: the average spectra while the doors were open. Top pola, bottom polB
• You can see the 50 KHz comb.
• The strongest peak is 1408.0143 then 1408.0638.
• Page 2:  1408.0143 and 1408.0638 birdies for the 590 seconds of data.
• Top PolA, bottom polB
• Green line shows where the door was opened.
• Blue line shows where the door was closed.
• Pol B is much stronger than polA at this az,za.
• the 1408.0143 channel gets up to 14 times Tsys. It was not resolved in a 538 Hz channel.

Summary:

• The 1408 MHz rfi is coming from the AC unit atop building I (the new one).
  
• I looked at the AC unit on top of blding II with the spectrum analyzer but didn't see anything at 1408. It may not have been running at the time.
• The rfi is strongest at 1408.0143 MHz.
• It is not resolved in 538 Hz. channels.
• Using the spectrum analyzer on the roof, the entire noise band increases when the AC unit doors are opened.
• The unit has electronic modules as well as printed circuit boards inside the enclosure on the roof.
• The 1408 MHz frequency is the one that pops up. I'm sure there are other birdies at different frequencies.
  

processing: x101/110929/rfi1408.pro