17oct12:netgear gsm7238fs switch emissions in the screen room:

   Luis measured the emissions of the netgear prosafe gsm72398fs switch  in the screen room on 17oct12.

Luis' setup and results (.pdf)

The setup was:

• Agilent E4445A spectrum analyzer:
• average 20 sweeps
• 8192 points/spectra
• 6db attenuation , preamp on
• 100 MHz span, 12.5 Khz/rbw. covering 0 to 3000 MHz in 30 spectra
• 904 Efield probe with a 15 foot coax cable.
• activity
• copper port (lower left) was plugged into ao net
• linksys 1GbE SX Mini-GPIC SFP module installed, but no fiber connected to the module.

• Luis measured the emissions with the 904 Efield probe close to the connected copper Ethernet port.
• He also measured the emissions at the rear of the device where the power cable left the device.

The plots:

    The plots show the emissions from the devices.

• The first page shows plots of 0-1000 MHz, 1000-2000MHz and 2000-3000MHz
• The 2nd page shows some blowups around some of our receiver bands:
• 300-500 MHz (327,430)
• 1350-1550 MHz (lbw)
• 2000-3000 MHz (lbw)
  
• Emissions at copper Ethernet port (lower left) (.ps) (.pdf)
• The strongest birdie is around 800 MHz at -78dbm
• Two combs were found (probably more exist)
• 125 MHz (red dotted line)
• It goes from dc to near 3ghz.
• the strongest birdie is at 1875 MHz ( -74 dbm)
  
• 330 MHz (green dotted line)
• the documentation say the cpu clock is 666 MHz. My guess is that it is 660 MHz.14.2854 MHz . 330 MHz is have the frequency.
• There are also some spikes separated by 82.5 MHz. This is 330/4.
  
• Emissions from the power supply (.ps) (.pdf)
• These emissions are similar but much weaker than the port8 output. Nothing is seen about 800 MHz.
  

Summary:

• The emissions from the switch are much stronger than the power supply
• there are at least 2 combs
• 125 MHz
  
• 0-3000 MHz
• 330 MHz (1/2 the cpu frequency)
  
• 0 to 1700 mhz
• There may also be an 82.5 MHz comb in lband.
  
• The table below summarizes the strongest birdies
  

• freq MHz	amplitude dbm	Notes
  660	-72	330 comb strongest birdie. The cpu clock is 660 MHz.
  330.0	-90	330 comb srongest birdie in 327 rcvr.
  437.35	-86	strongest birdie in 430 rcvr
  1375.00	-85	125. comb. strongest element in lbw
  1402.503	-83	These 3 birdies look like a 82.5 MHz comb.
  1485.057	-80
  1567.56	-80
  2250.06	-85	125 MHz comb. strongest element in 1-2 ghz
  2625	-91	125 MHz comb

• Most of the birdies are about 1 MHz wide. I wonder if they are chirping the cpu clock?
• The baseline has lots of variation. reflections from probe?
  
• the 1402 MHz birdie could be a real problem for HI observing.
  
• We also need to test this switch with a fiber optic cable installed to check on the fiber transceiver.
• These switches need to be checked on the telescope where ever they are installed. The best thing would be to shield them.
  

processing: x101/rfi/miscdevice/netgearsw_gsm7328fs_121017/do_gsm7238fs.pro

---