Radiated Emissions From Ethernet Data Traffic Signals June 13, 2010.ppt
- 1. 1 Radiated Emissions from Ethernet Data Traffic Signals June 13, 2010
- 2. 2 Measuring The Common-Mode Current On the Ethernet Cable • The test setup shown below was used to measure the common-mode current for the various equipment under test (EUT). • A 27 turn coil was wrapped around the CAT5 UTP Ethernet cable to couple any common-mode current on the Ethernet cable coming from the EUT to the spectrum analyzer. • The EUT is powered from the POE Injector at the other end of the Ethernet cable. • The EUT makes a Ethernet communication link with the Ethernet Switch at the other end of the Ethernet cable, causing Ethernet traffic to flow on the CAT5 UTP cable. Earth Ground Ethernet Switch Netgear FS605 v3 POE Injector Phihong PSA16U-480 120VAC Power Source Spectrum Analyzer HP 8591EM EUT 50 Ohm Coax Cable 27 Turn Coil CAT5 UTP Cable CAT5 UTP Cable CAT5 UTP Cable 4.5 feet 20.5 feet
- 3. 3 Separating Switching Power Supply Common-Mode Current Noise From Ethernet Data Common-Mode Current Earth Ground Ethernet Switch Netgear FS605 v3 POE Injector Phihong PSA16U-480 120VAC Power Source Spectrum Analyzer HP 8591EM EUT 50 Ohm Coax Cable 27 Turn Coil CAT5 UTP Cable CAT5 UTP Cable CAT5 UTP Cable 4.5 feet 20.5 feet • To determine which components of the common-mode current on the CAT5 UTP cable is due to Ethernet Data Traffic the following technique is used: • Step 1: Remove the CAT5 UTP link between the POE Injector and the Ethernet Switch. Power up the EUT from the POE Injector and after the EUT is fully powered observe the noise floor on the spectrum analyzer in the frequency band from 30 MHz to 100 MHz. • Step 2: Connect the CAT5 UTP cable between the POE Injector and the Ethernet and observe the link LED on the Ethernet Switch to ensure that the two Ethernet PHYs (EUT and Ethernet Switch) have linked up and are transmitting data streams in both directions. Again observe the signal level on the spectrum analyzer in the frequency band from 30 MHz to 100 MHz. • On the spectrum analyzer any regions along the frequency axis where the signal level drastically increases between Step 1) and Step 2) is an indication that we are observing power levels of common-mode current where the signal source is the Ethernet data traffic.
- 4. 4 Separating Switching Power Supply Common-Mode Current Noise From Ethernet Data Common-Mode Current • In any of Pelco’s Class B IP cameras the level of common-mode current due to switching power supply noise is going to be much lower than the level of the common-mode current due to Ethernet data traffic. If the switching power supply common-mode noise current was higher than Ethernet data traffic common-mode current then the EUT would not be able to pass Class B radiated emissions as the switching power supply noise would exceed the Class B radiated emissions limit. Therefore it is easy to measure the level of the common-mode current due to Ethernet data traffic. • Past experiments performed in the compliance chamber have shown that the radiated emission power from Ethernet data traffic signals is a constant value and does not vary with the amount of data being sent across the Ethernet link per unit time. The radiated emission power is the same whether the data payload is only housekeeping data, or if the data payload is at the maximum capacity of the Ethernet link. Therefore for this common-mode current measurement we only require the two Ethernet PHYs to be linked. We do not need to be sending any video streaming data across the Ethernet link as this would not change the measured results. Earth Ground Ethernet Switch Netgear FS605 v3 POE Injector Phihong PSA16U-480 120VAC Power Source Spectrum Analyzer HP 8591EM EUT 50 Ohm Coax Cable 27 Turn Coil CAT5 UTP Cable CAT5 UTP Cable CAT5 UTP Cable 4.5 feet 20.5 feet
- 5. 5 Ethernet Data Traffic Radiation from Ethernet Cables • In a system with a perfectly balanced transmitter and receiver at each end of the Ethernet cable and a perfectly balanced Ethernet CAT5 (unshielded twisted pair) UTP cable there would be no common-mode current on the Ethernet cable due to Ethernet data traffic because the Ethernet communication uses differential signaling with no common-mode signaling. In perfectly balanced differential signaling systems the impedance to all metal reference planes, including the earth ground reference plane, at all points in the signal path from the positive signal conductor or negative signal conductor is equal. In a balanced system both the positive signal conductor and the negative signal conductor need to see the same complex impedance to every metal reference plane within the extent of the near field energy bubble, at every point along the signal path from transmitter to receiver. • In a practical system the transmitter and receiver at each end of the Ethernet cable have some level of imbalance. All practical Ethernet cables have some amount of imbalance. Both of these imbalances convert some portion of the differential Ethernet data traffic signal energy to common-mode current in the Ethernet cable. • The near field from the differential Ethernet data traffic is small, mostly contained inside the CAT5 UTP cable structure. Therefore the amount of radiated emission energy radiated to the far field from the differential Ethernet data traffic signal is negligible.
- 6. 6 Ethernet Data Traffic Radiation from Ethernet Cables Earth Ground Ethernet Switch Netgear FS605 v3 POE Injector Phihong PSA16U-480 120VAC Power Source Spectrum Analyzer HP 8591EM EUT Common-Mode Ethernet Data Current Mirror Image Common-Mode Ethernet Data Current Common-Mode Ethernet Data Displacement Current (Virtual Photons) Large common-mode current loop creating efficient low frequency antenna (30 MHz to 100 MHz) • Ethernet transmitter, receiver, and cable imbalances convert some portion of the differential Ethernet data traffic signal energy to common-mode current. This common-mode current flowing in the CAT5 Ethernet cable has a mirror image current flowing in the earth ground reference plane. The current loop between the CAT5 Ethernet cable and the earth ground reference plane is very large, resulting in an extremely large near field. • The Ethernet data traffic common-mode current has an extremely large near field, creating a very large antenna structure. This large antenna structure is very efficient at radiating the common-mode Ethernet data traffic energy in the frequency band from 30 MHz to 100 MHz.
- 7. 7 Measured Ethernet Data Traffic Common-Mode Current • The table below enumerates the results of measured Ethernet traffic common-mode current on various Pelco IP camera products as measured by the test setup described in this document. • The Iliad had the lowest measured level, 4.6 dB below the average level for all measured products. • The P617 Beta dome with the A1 Backbox had the highest measured level, 4.4 dB above the average of all measured products. • As seen in the results below the differential in measured common-mode currents between the Iliad and the P617 Beta dome with the A1 Backbox is 9.0 dB. • In the report generated by Dustin and Skip, based on comparative radiated emission scans performed in the compliance chamber, they reported a measured difference in radiated emissions between the Iliad and the P617 Beta dome (with the A1 Backbox) of approximately 10 dB in the low end wideband energy (30 MHz to 100 MHz). • The 10dB radiated emission differential measured in the compliance chamber and the 9 dB differential in measured common- mode current levels confirm the expected correlation between radiated emission results and cable common-mode current measurements. EUT Description Measured Peak Ethernet Data Common-Mode Current (dBm) Deviation From Average (dB) Iliad -73 -4.6 Spectra Mini IP -70 -1.6 Beta P617 Dome with X0 Backbox and External In-Line Filter Dongle Added -70 -1.6 Atlas IP110 -69 -0.6 Homer -67 1.4 Odyssey -67 1.4 Troy -67 1.4 Beta P617 Dome with A1 Backbox -64 4.4 Average Level of Above Products -68.375
- 8. 8 Converting Common-Mode Current Readings to Predicted Radiated Emission Performance • It makes sense that the common-mode current measurements correlate with the chamber radiated emission measurements since the measured common-mode current is the drive signal to the Ethernet cable antenna which is transmitting the energy being measured in the chamber radiated emissions test. In the 30 MHz to 100 MHz frequency band the antenna has to very large in size to be efficient, and the only transmitting antenna in the equipment under test (EUT) large enough to be efficient in the 30 MHz to 100 MHz band is the Ethernet cable antenna. • What we require is a scaling factor to translate between radiated emission readings and the common-mode current readings. When the P617 beta dome with the A1 Backbox was tested in the recently developed maximized resonant (no damping) radiated emission test setup the maximum quasi-peak reading due to Ethernet data traffic radiation was approximately +4.0 dB above the Class B radiated emission limit. This data point provides a scaling factor to translate the measured common-mode current readings to the expected maximum quasi-peak radiated emission readings from Ethernet data traffic. • The table below shows the results of the conversion from measured common-mode current to expected (predicted) radiation emission results, based on the P617 data point scaling factor.
- 9. 9 Converting Common-Mode Current Readings to Predicted Radiated Emission Performance • Based on the radiated emission estimates shown below only the Iliad is expected to pass the Class B radiated emission test with Pelco’s 3dB of margin using the maximized resonant (non-damped) radiated emission test setup. • Based on the radiated emission estimates shown below only the Iliad, Spectra Mini IP, and IP110 are expected to pass the Class B radiated emission test using the maximized resonant (non-damped) radiated emission test setup. • A hypothetical “average” EUT is expected to pass the Class B limit with a margin of 0.4 dB. In practical terms this margin is too small. Under repeated testing the “average” EUT will alternate between barely passing and failing. • It would be informative to open a research project to investigate how the Iliad has achieved 5 dB of passing margin, where none of the other products come even close to this result. A possible explanation is that the Iliad is the only product in this group to use the Akros AS1601 active EMI suppressor IC, which actively balances the Ethernet lines. A useful experiment would be to test the Iliad with the AS1601 IC removed. • On average the current Pelco designs do not have enough margin to pass Class B radiated emissions. Something needs to changed in the designs to achieve more margin.
- 10. 10 Conclusions • On average Pelco IP camera products do not have enough margin to pass Class B radiated emission with acceptable margin. The “average” EUT shown below is only expected to pass Class B radiated emissions with 0.4 dB margin. • Research is required to determine how to increase the average radiated emission margin. Two possible areas to investigate are: a) whether the Akros AS1601 Active EMI Suppressor IC (active balancer) is providing extra margin in the Iliad, and if so can this Akros AS1601 Active EMI Suppressor IC part provide extra margin in the other designs; b) determine through experimentation how much extra margin can be achieved by changing from unshielded twisted pair (UTP) Ethernet cables to shielded twisted pair (STP) Ethernet cables. Exploration of any other possible design change options is also recommended.