LISN UP Application Note

Transcription

1 LISN UP Application Note What is the LISN UP? The LISN UP is a passive device that enables the EMC Engineer to easily distinguish between differential mode noise and common mode noise. This will enable efficient troubleshooting of power line EMI filters shortcomings and allow the EMC Engineer to focus on the dominant mode noise performance. Once the dominant mode is determined, the filter components associated with the dominant mode can be identified. Identification of the offending filter components will focus efforts rather than the expensive trial and error random component change and retest cycle. The LISN UP is connected to the output ports of either a dual line LISN or two identical single line LISNs. Care must be taken to make sure the interconnecting cables are identical in length and construction and have good RF shielding effectiveness. The output of the LISN UP can be switched between the common mode component and the differential mode component by a manual two position switch without changing any external RF connections. How Does It Work? First, we need to review some of the theory so we can start from a common understanding of noise voltage modes. In a typical circuit, conventional current flows from the positive voltage terminal to the load and it eventually returns to the negative terminal of the voltage source as shown in Figure 1. This is differential mode current. The current flowing into the load is equal to the current flowing out of the load and back to the source. The only difference is the phase between these two currents. They differ by 180 o and ideally they would cancel out. Common mode currents flow on each lead and return on the ground plane or green wire safety ground. The amplitude and phase of these currents are the same on both conductors. A diagram of this is shown in Figure 2. Putting these two concepts together and examining the current flows we find that the net current on any of the conductors is algebraic sum of the differential mode current and the common mode current. The current returning via the green wire ground is the sum of the common mode currents or twice the common mode current of any one of the conductors. This is illustrated in Figure 3. Knowledge of these characteristics will be used to our advantage when we need to determine whether the voltages are common mode or differential mode. 1

2 Since we know the equations to describe the current in each conductor we can perform mathematical operations to separate the differential mode and common mode components. The equations are listed below: I a = I cm + I dm I b = I cm - I dm These currents flow across the impedance of the LISN which give us a corresponding noise voltage to measure. The noise voltage for each conductor can be described as follows: V a = (I cm + I dm) * Z lisn V b = (I cm - I dm) * Z lisn If we sum V a and V b together, ideally we will obtain twice the common mode voltage component and the differential mode component will cancel. This means that if we had equal common voltage magnitudes on both conductors and they had the exact same phase we would get twice the voltage or 6 db higher than the magnitude of either line. This is illustrated in Figure 4. the differential mode component is added. We ideally get twice the differential mode component. This means that if we had equal differential mode voltage magnitudes on both conductors and they had the exact same phase we would get twice the voltage or 6 db higher than the magnitude of either line. This is How Do I Set It Up? The LISN UP is designed to be used with either a dual LISN or two individual LISNS. The unit under test is connected to the power supply and LISN the same way it has always connected. The differences are the connections between the 50 ohm measurement ports and the EMI Receiver or spectrum analyzer. Each RF measurement port on the LISN is attached to the LISN UP input ports. The LISN UP Output port is connected to the measurement receiver. The RF output is switched between common mode and differential mode. Typical connections are shown in Figure 6. The coaxial cables between the LISN RF output ports and the LISN UP Input ports must be identical to maintain the performance of the LISN UP. The cables must be of identical length, shielding quality and characteristic impedance. If adapters are used they too must be identical. To separate out the differential mode voltage an additional operation must be performed. One of the inputs to the summing junction must have the phase inverted 180 o prior to summing the two voltage signals together. When the two signals are summed together the common mode component is ideally canceled and 2 How Do I Use It? The LISN UP is relatively simple to use if you have performed conducted voltage emissions tests, and have an understanding of the fundamentals of common mode, differential mode, and filtering. Typically, in the past, conducted voltage emissions were performed and the equipment under test passed or failed. If it failed you ended up in a try a fix and repeat the measurement scenario. Depending on your experience and luck you may have fixed the problem in a few hours or weeks. The LISN UP will help you focus on the dominant noise mode for those out of specification conditions rather than blindly changing components in the hope it fixes the problem. You may have a common

3 mode problem yet you decide to change a differential mode filter component. You repeat the measurement on both lines and you find there is no difference in the performance. Had you known that the common mode component was dominant you would have chosen to add a common mode component or modify the existing common mode component. This would have saved you valuable time! A generic procedure for troubleshooting your input EMI filter out of compliance conditions is outlined below. GENERIC PROCEDURE The RF conducted voltage emissions for each power line is generally measured first to determine compliance to the applicable regulatory specification. If there are non-compliances, the LISN UP is used to determine the dominant modes of conducted noise voltage. The dominant mode at each frequency of noncompliance needs to be characterized to efficiently troubleshoot and modify the EMI filter. This is accomplished using the test procedure below: Step 1. Measure the conducted voltage emissions for each line under test using the required LISNs. Step 2. Note any non-compliant voltage emissions conditions Record each non-compliant frequency, amplitude and the offending line. Step 3. Connect the LISN UP to the dual LISN or (2) single LISNs. This is accomplished by connecting the LISN RF Output Ports to the RF Input Ports of the LISN UP. Connect the measurement receiver to the RF output of the LISN UP. See Figure 6. Step 4. Power up the equipment under test. Step 5. With the LISN-UP RF Output Switch in the Differential Mode position and measure the conducted voltage emissions at the first non-compliant frequency. Step 6. Record the frequency and amplitude for the Differential Mode measurement. Step 7. Position the LISN-UP RF Output Switch to the Com mon mode position and measure the conducted voltage emission at this frequency. Step 8. Compare the relative amplitudes of the Common Mode component and the Differential Mode component of the conducted noise voltage. If either component is signifcantly larger than the other, that is the dominant mode. There will be cases where the amplitudes are of similar magnitudes for both components. In this case, there is both common mode and differential mode noise present and no one mode dominates. Step 9. Repeat this for each non-compliant frequency until all have been identified. Step 10. Depending on which mode(s) are dominant the filter components that need to be changed can easily be identified. Step 11. After the EMI filter has been modified repeat steps 1 and 2. 3 Step 12. If the equipment under test is still out of compliance repeat steps 3 to 11 until the EMI Filter meets the applicable specification. Which Filter Components Are Differential Mode or Common Mode? Differential mode filter components are usually the components that are either: line to line capacitors, series inductors and sometimes leakage inductance from a common mode choke. Common mode components are line to chassis capacitors and common mode chokes. A generic EMI filter diagram is shown in Figure 7 and the components are labeled differential mode or common mode. The differential mode components are labeled with a dm subscript and the common mode components are labeled with the cm subscript. Example Problem A simple conducted noise source and EMI Filter were designed and constructed to demonstrate how the LISN UP is used. This will help the user better understand how the LISN UP works, what to expect and how to interpret the data. A diagram of the test setup is shown in Figure 8. We will use the generic procedure as a guide. Step 1. We measured the conducted voltage emissions of the +5Vdc line and the power Return line. This is shown in Figure 9. Step 2. We find that the conducted voltage emission levels are higher than 40 dbuvrms on both lines over much of the frequency range of 10 khz to 30 MHz. Step 3. Connect up the LISN UP as shown in Figure 10. Step 4. Equipment under test is powered ON. Step 5. Position the RF output switch to the differential mode position and measure output. Step 6. Record the differential mode voltage data. The differential mode data is shown in Figure 11. Step 7. Position the RF output switch to the common Mode position and measure the output. Step 8. Record the common mode data. The common mode data is shown in Figure 12. Compare the common mode trace to the differential mode trace. The different

4 tial mode component dominates over the frequency range of 10 khz to 6 MHz. The common mode component dominates above 6 MHz. Step 9. Included above. Step 10. We first add a differential mode capacitor of 100uF and repeat the common mode and differential mode measurements. We find a significant decrease in the differential mode component and small change to the common mode component. See Figure 13 for the differential mode component and Figure 14 for the common mode component. Step 11. The 100uF differential mode capacitor is removed and replaced with a 0.01 uf common mode capacitor between each line and chassis. The common mode component is reduced for frequencies greater than 5 MHz. The differential mode component increases in the 2 MHz to 4 MHz range with a resonance at 3.2 MHz. The common mode component is shown in Figure 15 and the differential mode component is shown in Figure 16. The next configuration is the parallel series combination of the 100uF differential mode capacitor, 0.01uF common mode capacitors and differential mode inductors in each line. The differential mode component shows a resonance at 1.8 MHz and the common mode component is significantly reduced over the entire frequency range except for the local AM radio station at 1070 khz. The differential mode component is shown in Figure 17 and common mode component is in Figure 18. The final configuration to be tested is the same configuration as above with the addition of a 3.24 mh common mode choke. The resonance at 1.8 Mhz is reduced by 6 db for the differential mode component and the common mode component has little visible change. The differential mode component is shown in Figure 19 and the common mode component is shown in Figure 20. Step 12. The LISN-UP is removed from the test setup and the configuration shown in Figure 8 is used to measure the resultant conducted voltage emissions for both lines. The emissions are generally lower than 35dBuV except for the resonance at 1.8 MHz which is 47dBuV. 4

5 5

6 6

7 7

8 8

9 9

10 10

11 Summary and Conclusions The LISN UP is a valuable troubleshooting tool for the EMC Engineer that can help save time by enabling rapid diagnosis of the dominant noise mode. Diagnosing the dominant noise mode allows for efficient modification of the EMI filter. If the differential mode component is dominant over a specific frequency range the EMC Engineer can focus on the differential mode filter components rather than the common mode components. This also works for the common mode case. Background Information The LISN UP is based upon the device first proposed by Paul and Hardin in [ 1] The device is able to separate the differential mode and common mode components of the conducted voltage emissions. This is accomplished with RF transformers and impedance matching resistors to sum voltages using the pertinent phase information to discern the predominant noise modes. There has been other work done to separate the differential mode and common mode components of conducted noise emissions. The first documented work was from Toppeto in Naves work appeared in The method proposed by Toppeto involved a current transformer, barbell test fixture and a high pass filter.[ 3] This approach is more suited to conducted current measurements than voltage measurements. Nave proposed a Differential Mode Rejection Network (DMRN) that would sum the output voltages from the LISN RF Output port and only pass the common mode component since the differential mode voltages cancelled.[ 2] In practice, the DMRN depends on a highly balanced network for each leg and precision resistors. A slight imbalance or difference in resistance values could easily degrade the performance of the DMRN. Nave also spoke of a selectable mode rejection network. Phase and Amplitude Effects on LISN UP Accuracy The summation of two or more voltages depends on the relative amplitude and phase of each. Maximum addition of two signals occurs when they are the same amplitude and there is zero phase difference between the two voltages. Maximum cancellation occurs when the amplitudes are identical and the phase difference between the two voltages is 180 degrees. A slight amplitude or phase imbalance can adversely effect the measurement accuracy. A graph depicting the amplitude difference between two signals to be added is shown in Figure 22. It shows that for maximum addition the amplitude difference must be less than 0.5 db. A second graph depicting relative phase for cancellation is shown in Figure 23. It shows that the phase and amplitude imbalances must be small for maximum signal addition and cancellation. The rho term is the relative amplitude expressed as a linear ratio of the two voltages. References: 1. C.R. Paul and K.B. Hardin, Diagnosis and Reduction of Conducted Noise Emissions, IEEE Transactions on Electromagnetic Compatibility, November 1988, Vol. 30, No. 4, pp M.J. Nave, A Novel Differential Mode Rejection Net work for LISNs. IEEE EMCS Symposium Record, Denver, CO., 1989, pp A.A. Toppeto, Test Method to Differentiate Common Mode and Differential Mode Noise, Third Symposium and Technical Exhibition on EMC, Rotterdam, pp , May 1-3,

12 Fischer Custom Communications, Inc Earl Street Torrance, CA Phone: Fax: