Introduction (top)

• 27jan16:
•  felix and i looked for rfi using the anritsu spectrum analyzer
• we also did a visual inspection of the cabinet and it's rfi protection.
  
• 28jan16:
  
• morning: 327 receiver, mock spectrometer took data while the the sband console was powered on ,off
• afternoon: victor and crew worked on fixing the rfi shielding
• afternoon: 327 receiver, mock spectrometer took data while sband console powered on,off

27jan16:anritsu spectrum analyzer test, visual inspection.

• The back door is closest to the transmitter room entry door.
• The front door of the cabinet points to the center of the azimuth arm.

Visual inspection:

• I took some pictures of the front of the cabinet before any changes were made:
• Bottom of cabinet before opening the door (.jpg)
• This is what we saw before opening the front door.
• The door had been closed on a piece of finger stock that had broken off inside.
  
• first test showed that there was lots of rfi coming out before we removed this piece
• The bottom of the cabinet used to have an rfi mesh to let cooling air enter/leave the cabinet.
• When new amps were added, they needed more room in the cabinet, so the amps were mounted at the bottom.
• This "extension" was added. I've been told it was only to pass the air into the console. I hope people aren't relying on duck tape for shielding!!
• bottom of cabinet, door open (.jpg)
• opening the door showed the broken piece of finger stock
• There doesn't appear to be any fingers along the bottom portion of the cabinet.
• the inner vertical fingers also seem to be coming off.
  
• top of the cabinet, door open (.jpg)
• When we opened the door, the finger stock along the top edge fell down.
• the copper fingers on the upper left are also corroded (see next image)
• 
  
• blowup of top front  showing corrosion (.jpg)
• the metal rail and the finger stock in the upper left is corroded.
  
• When we tested for rfi with the door closed, we saw strong rfi coming out of this area.
• I'm a little surprised that the transmitter group let the rfi shielding deteriorate to this level.

Some data taken with the anritsu spectrum analyzer

    We used the anritsu spectrum analyzer and the large loop antenna to snoop around the cabinet when the doors were closed

•    the plots show some of the birdies we found (.ps) (.pdf)
• the spectrum analyzer setting was:
• preamp on, rbw=3hz,avg 3, and peak hold
• the -160dbm baseline is about 2415 Kelvins (about 24 times higher than the 327 receiver)
• The birdies were found strongest at the front door.
• As we moved away from the door, the birdies got weaker.
• I didn't check the anritsu drift (it wasn't locked).. previous checks showed in within 10's of hz at 327..
  

160128: sbconsole on/off (before repairs) (top)

    On 160128 (in the morning) we took data with the 327 receiver and the mock spectrometers. While taking data we cycled the power on the sband console a few times. We wanted to see which birdies went away when the rack was off.

Mock setup:

• Same as 160125 az swing.
  
• 7bands each with 8 Mhz bw, 8192 channels, 1 second sampling.
• This has a 1 Khz resolution bandwidth.
  

The dynamic spectra:

    The dynamic spectra were made by:

• Computing median bandpass over when cabinet was off
• fitting a 20 order harmonic to this median bandpass
• flatten each of the spectra with this.
• remove tsys from each spectra using the median
• average the two pols.
• I placed horizontal lines where the cabinet was powered on ,off
• powering the cabinet on,off would take up to 30 seconds..
• i flagged where the cabinet was completely off..
  

• The strongest rfi occurred before the first turn off.
• After turning it back on, it looks like it takes awhile for the strength of the rfi to increase.
• The solid horizontal lines are usually when one set of the breakers were thrown.. probably an arc.
  
• a ups outside the rack was usually the last thing to be turned off
• this drove a linear power supply (outside the rack) that powers some drivers inside the cabinet
• (is this power cable filtered?)
• You can see a birdie above 304Mhz that has a delayed turnoff. this is probably coming from what the ups drives.
• Not sure where the sweeping rfi at 306 Mhz is coming from ..
• it's not the sband console..
• looks like some of the sweeping rfi we had at lband long ago, when the receiver cal drivers  were oscillating.
• I forgot and left the alfa motor and the cameras on during this test..
  
• some of the solid vertical lines will be coming from them..
• Some combs can be seen at:
• 303.8,309.8,314.25,320.25.... it continues all the way up..
  

Comparing cabinet on with cabinet off:

    For each band i averaged when the power was all the way on, and when it was all the way off.. i tried to leave out the transition periods.

The plots show the rfi when the cabinet was on and off (.ps) (.pdf)

• Each page shows 1 8MHz band.
• Top frame: vertical scale is 0 to 1.5 Tsys
• Bottom frame blows up the vertical scale to 0 to .2 Tsys
• The on time avg is 394  seconds. the off time avg is 238 seconds.
  

• Black is with the cabinet off, red is with the cabinet powered on.
• I plotted the red (on) and then over plotted the black (off).
• If you see a red line, with a black line over it, then that rfi did not come from the sband console
• Red lines sticking up by themselves are coming from the sband console.

Summary (before repairs)

• in the 327 Mhz receiver we  see 100's of spikes coming from the sband console.
• 1kHz resolution, 400 second integration
• The strength of these birdies was probably stronger than we measured.
• The broken piece of finger stock that was sticking out of the cabinet was radiating... prior to this test we removed it.
• Some rfi remains when the cabinet breakers are thrown, but the external ups was still on.
• The rfi was strongest after the cabinet had been powered on for awhile.
• victor and crew spend 28sep16 afternoon working on the shielding...
• Notes: i also took some lbw data to see if it was getting into lband..
  

processing: x101/160129/rfi327_before.pro

160128: sbconsole on/off (after cabinet repairs) (top)

    Victor and his crew worked on repairing the sband console shielding on 160128. After the repairs, we repeated the sband on/off console tests using the 327 mhz rcvr and the mock spectrometer (see above for setup). The external ups,power supply was still present for these tests.

Dynamic spectra while cabinet turned on/off.

• these are the dynamic spectra made after the cabinet was repaired.
• I placed horizontal lines where the cabinet was powered on ,off
• powering the cabinet on,off would take up to 30 seconds..
• i flagged where the cabinet was completely off..

Comparing cabinet on with cabinet off:

    For each band i averaged when the power was all the way on, and when it was all the way off.. i tried to leave out the transition periods.

The plots show the rfi when the cabinet was on and off (.ps) (.pdf)

• Each page shows 1 8MHz band.
• Top frame: vertical scale is 0 to 1.5 Tsys
• Bottom frame blows up the vertical scale to 0 to .2 Tsys
• The on time avg is 364  seconds. the off time avg is 224 seconds.
  

• Black is with the cabinet off, red is with the cabinet powered on.
• I plotted the red (on) and then over plotted the black (off).
• If you see a red line, with a black line over it, then that rfi did not come from the sband console
• Red lines sticking up by themselves are coming from the sband console.
• The vertical axis is in units of Tsys.. the plots are clipped to 1.5 Tsys

the final plots show the peak value for each rfi (.ps) (.pdf)

• The top frame shows the full scale.
• There are 27 rfi peaks that were found (there may be more).
• The frequency and pkValue (in units of tsys 1khz rbw) are printed.
• 328.0205 is the strongest at 22.8*Tsys.
  
• the bottom frame is clipped to .8*Tsys.
• There is a comb with spacing of 4.0002Mhz
• it starts at 300.0186 (so it is probably not locked to the station clock).
  

Summary (After repairs):

• The repairs made a big difference
• Prior to the repairs there were 100's of rfi spikes coming from the cabinet (see top of page)
  
• after the repairs, there are about 27 rfi spikes (in a 1khz rbw and 364 sec integration).
  
• Part of the remaining rfi is  a strong 4 mhz comb coming from the cabinet.
  
• The strongest element is at 328.0205 Mhz at  about 23*tsys
  
• We will try to cleanup the rest of these rfi's
• A question for the astronomy/radar group is:
• While we are trying to cleanup the rest of these birdies,  should the sband console be turned of when not in use?

processing: x101/160128/rfi327_after.p

160203: 4mhz comb coming from ups (top)

   On 03feb16  Felix went looking for the 4 mhz comb coming from the sband cabinet.
He used the anritsu spectrum analyzer (3 hz rbw) and the loop antenna.
    The strongest component of the comb (328.0205 Mhz) was found to come from a ups that was sitting outside of the cabinet.
When the exciter amplifiers failed, they were replaced with another set. This set was install in the bottom of the rack. I was told that there was no room to put the power supply inside the rack, so it (along with a ups) were placed outside the rack, and then at twisted pair power cable was brought into the rack (with no filter?).
    The power supply was  6 volts, 3 amps, linear supply..

    Victor and crew will try to place this power supply   in the cabinet tomorrow (04feb16). This will eliminate the need for the ups outside of the cabinet.
We can then recheck  things with the 327 receiver to see what is left.

The plot shows a screen grab (from the anritsu spectrum analyzer) of the 328 Mhz birdie when the antenna was close to the power supply/ups (.jpg)

• Before the measurement felix checked and found that the anritsu was off by about 70 hz at 328 Mhz when free running.

160204: sband console on/off, external ups removed. (top)

    The sband console had a power supply and ups outside of the rfi cabinet. On 04feb16 victor and his crew placed the power supply inside of the rfi cabinet, and connected it to the cabinet power. The external ups was then removed.

    We then did a test to check for any rfi coming from the cabinet:

• Victor went to the sband power supply building  to turn the power on/off to the sband cabinet in the dome.
• data was taken with the 327 receiver and the mock spectrometer while the power was cycled:
• 300 to 350 Mhz
  
• 7 mock spectrometers used to cover 300 to 350 Mhz. 8192 channels/box, 976 channel width, 1 second integrations.
• two on/off cycles were performed.
  

Dynamic spectra while cabinet turned on/off.

• these are the dynamic spectra made after the cabinet was repaired and the ups removed.
  
• I placed horizontal lines where the cabinet was powered on ,off
• powering the cabinet on,off would take up to 30 seconds..
• i flagged where victor told me the cabinet was on or off..

Dynamic spectra notes:

• There was some broadband interference that was occurring during this test.. not sure where it was coming from.. It may have been some other people working on the platform.

Comparing cabinet on with cabinet off:

    For each band i averaged when the power was all the way on, and when it was all the way off.. i tried to leave out the transition periods.

The plots show the rfi when the cabinet was on and off (.ps) (.pdf)

• Each page shows 1 8MHz band.
• Top frame: vertical scale is 0 to 1.5 Tsys
• Bottom frame blows up the vertical scale to 0 to .2 Tsys
• The shortest on time avg is 90  seconds. the shortest off time avg is 70 seconds.
  

• Black is with the cabinet off, red is with the cabinet powered on.
• I plotted the red (on) and then over plotted the black (off).
• If you see a red line, with a black line over it, then that rfi did not come from the sband console
• Red lines sticking up by themselves are coming from the sband console (or from time variable rfi that only occurred during the off time
  
• The vertical axis is in units of Tsys.. the plots are clipped to 1.5 Tsys

the final plots show the peak value for each rfi (.ps) (.pdf)

• The top frame shows the full scale.
• There are 7 rfi peaks that were found.
• The frequency and pkValue (in units of tsys 1khz rbw) are printed.
• 337.8887 is the strongest at .146*Tsys.
  
• The comb with spacing of 4.0002Mhz coming from the ups is gone.. (although 344.0322 falls on one of the harmonics)
  
• it starts at 300.0186 (so it is probably not locked to the station clock).

Summary (after ups removed):

• the 4 mhz comb is not seen (although 342.0322 was a comb harmonic and it is still visible)
• there are now 7 frequencies (between 300 and 350 Mhz) that we can see coming from the cabinet (976 rbw, 1 sec integration).
• Removing the external ups , external power supply, and the power cable running into the sband cabinet
  
• reduces the number of rfi birdies from 27 to 7 frequency spikes.
• In general the spikes that remain have gotten weaker (they may have been getting out through the power cable that came from the external power supply.

processing: x101/160204/rfi327_sbcon.pro

home_~phil