1、August 15, 2006 Copyright 2006 Rockwell Automation, Inc. All rights reserved. 1 EMI Modeling and Characterization of the Rhino SMPS Carson Baisden Ranga Tallam Gary Skibinski Presented by: Mark Christini - Ansoft2 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Outline Problem Goals Pr
2、ocedure Modeling Process Results Modeling benefits Reduction of conducted EMI techniques Summary3 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Problem Goal: Characterize and reduce EMI emissions from Flyback SMPS Model SMPS Determine critical components in system Determine noise att
3、enuating techniques Procedure Model: Actives and Passives (discrete components and interconnects) Use simulations to find critical paths Focus attention on these paths to mitigate noise4 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Modeling Process Model all passive components using
4、 impedance analyzer Model active components from datasheet Create geometric model in Q3D and solve Create and solve transient model in Simplorer Perform FFT in Matlab 0 5 10 15 20 25 30 35 -5 -4 -3 -2 -1 0 1 2 LISN Resistor Voltages Volts (V) Time ( sec)V Rp V Rn 0.15 0.3 1 3 10 30 -20 0 20 40 60 80
5、 100 120 Simulated Total EM I Spectrum Noise (dB V) Frequency (M Hz)Simul ated EMI CISPR A CISPR B5 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Flyback SMPS LISN Line impedance stabization network Used for testing LISN Line impedance stabization network Used for testing6 Copyright
6、2006 Rockwell Automation, Inc. All rights reserved. Passive Modeling Model all passive components with impedance analyzer To create frequency dependent equivalent network for device 10 2 10 3 10 4 10 5 10 6 10 7 -20 0 20 40 60 80 100 Magnitude (dB) 2x SMT DO3316P-473 Coilcraft Inductor Impedance10 2
7、 10 3 10 4 10 5 10 6 10 7 -90 -45 0 45 90 Frequency (Hz) Phase (deg)Measured Modeled7 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Active Modeling for Mosfet 0 10 20 30 40 50 0 400 800 1200 1600 2000 Capacitance for STW4N150 C (pF) V DS(V)C rss C oss C oss,Measured C iss Datasheet P
8、lot SimplorerModel SimplorerModel Datasheet Datasheet Datasheet8 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Active Modeling for Mosfet SimplorerModel Simplorer Simplorer Model Model Datasheet Plot Datasheet Datasheet Datasheet 0 10 20 30 0 1 2 3 4 5 6 Output Characteristics for ST
9、W4N150 V DS(V) I D(I)6V 7V 8V 9V9 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Q3D Modeling Create geometric model in Q3D Draw using layout Import using interface tool Define materials Define boundary conditions Conductors (Nets) to solve the capacitance between Terminals (Sources/S
10、ink) to solve resistance and inductance between Solve (C, R, L) Post Processing Export lumped parameters Matrix Reductions Field plots10 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Q3D Results Capacitance Matrix (30 min) 1 10pF 11 5-10pF 24 1-5pF 428 1pF Inductance Matrix: R,L,M (3
11、0 min) 63 Paths (23 Pri & 40 Sec) 31 Paths (16 Pri & 25 Sec)11 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Simplorer Model 0 5 10 15 20 25 30 35 -5 -4 -3 -2 -1 0 1 2 LISN Resistor Voltages Volts (V) Time ( sec)V Rp V Rn Blue elements are manually input from Q3D and are turned on/of
12、f 2 ns step size 4 ms total time 30-60 min Simulation time V DC =200V nominally 50% (resistive) load nominally 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Simulated CM EMI Spectrum Noise (dB V) 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Simulated DM EMI Spectrum DM Noise (dB V) Frequency (MHz)12
13、Copyright 2006 Rockwell Automation, Inc. All rights reserved. Measurement Procedure 125 MHz sampling 10 ms recorded V DC =200V nominally (up to 600V) 50% (DC fan) load nominally DC Power Supply SMPS LISN13 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Modeling Results 0.15 0.3 1 3 10
14、 30 -20 0 20 40 60 80 100 120 Simulated -Black vs Measured -Red CM EMI Spectrum CM Noise (dB V) 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Simulated -Black vs Measured -Red DM EMI Spectrum DM Noise (dB V) Frequency (MHz)Simulated Measured14 Copyright 2006 Rockwell Automation, Inc. All rights reser
15、ved. Modeling Benefits Show the importance of various aspect within the circuit Trace impedances Active components Passive components: values and type Critical paths Simple demonstrations of attenuations techniques Adding filters into the circuit Re-arrange the components Shielding the Transformer15
16、 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Modeling Results Q3D R/L 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Simulated CM EMI Spectrum Comparing R/L Q3D Results CM Noise (dB V) 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Simulated DM EMI Spectrum Comparing R/L Q3D Results DM
17、 Noise (dB V) Frequency (MHz)w/ Q3D R&L w/o Q3D R&L16 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Modeling Results - MOSFET 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Simulated CM EMI Spectrum Comparing MOSFET Models CM Noise (dB V) 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Si
18、mulated DM EMI Spectrum Comparing MOSFET Models DM Noise (dB V) Frequency (MHz)Ideal MOS w/ junct cap Detailed MOS17 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Modeling Results Q3D R/L Secondary 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Simulated CM EMI Spectrum Comparing Q3D o
19、n Secondary CM Noise (dB V) 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Simulated DM EMI Spectrum Comparing Q3D on Secondary DM Noise (dB V) Frequency (MHz)w/ Q3D R&L w/o Q3D R&L18 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Modeling Results Cs & Ds Secondary 0.15 0.3 1 3 10 30 -2
20、0 0 20 40 60 80 100 120 Simulated CM EMI Spectrum Comparing Actives and Passives on Secondary CM Noise (dB V) 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Simulated DM EMI Spectrum Comparing Actives and Passives on Secondary DM Noise (dB V) Frequency (MHz)Detailed Caps & Diodes Ideal Caps & Diodes19
21、 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Modeling Results DC Excitation 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Simulated CM EMI Spectrum Comparing DC Bus Voltage CM Noise (dB V) 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Simulated DM EMI Spectrum Comparing DC Bus Voltag
22、e DM Noise (dB V) Frequency (MHz)V DC =200V V DC =685V20 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Modeling Results L 2 , L 3 Increase 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Simulated CM EMI Spectrum Comparing L 2& L 3 Values CM Noise (dB V) 0.15 0.3 1 3 10 30 -20 0 20 40 6
23、0 80 100 120 Simulated DM EMI Spectrum Comparing L 2& L 3Values DM Noise (dB V) Frequency (MHz)L 2 =L 3 =45 H L 2 =L 3 =400 H21 Copyright 2006 Rockwell Automation, Inc. All rights reserved. XFMR Shielding - Theory Where/how to physically place shield Where to electrically connect shield Size and cos
24、t of shield Is it all worth it? Add Shield22 Copyright 2006 Rockwell Automation, Inc. All rights reserved. XFMR Shielding - Geometry Core Core +12V -12V 24V 15V P N Pri 2 P B P A +12V -12V 24V 15V P N P B P A Shield ? P C P C Pri 1&2 Pri 1 Core Core Present Winding Cross-Section Proposed (Shielded)
25、Winding Cross-Section23 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Modeling Results XFMR Shielding 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Simulated CM EMI Spectrum Comparing XFMR Shielding CM Noise (dB V) 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Simulated DM EMI Spectrum
26、 Comparing XFMR Shielding DM Noise (dB V) Frequency (MHz)Shield to Gnd No Shield Shield to -DC Shield to +DC24 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Measurement DC Excitation 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Measured CM EMI Spectrum Varying DC Bus Voltage CM Noise
27、 (dB V) 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Measured DM EMI Spectrum Varying DC Bus Voltage DM Noise (dB V) Frequency (MHz)600V 500V 400V 300V 200V25 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Measurement DC Bulk Caps 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Measured
28、CM EMI Spectrum Adding DC Bus Capacitors CM Noise (dB V) 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Measured DM EMI Spectrum Adding DC Bus Capacitors DM Noise (dB V) Frequency (MHz)w/o DC Caps w/ DC Caps26 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Measurement DC Y-Caps 0.15 0.3
29、 1 3 10 30 -20 0 20 40 60 80 100 120 Measured CM EMI Spectrum Adding DC Y-Capacitors CM Noise (dB V) 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Measured DM EMI Spectrum Adding DC Y-Capacitors DM Noise (dB V) Frequency (MHz)w/o DC Y-Caps w/ DC Y-Caps27 Copyright 2006 Rockwell Automation, Inc. All r
30、ights reserved. Measurement Bulk & Y-Caps 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Measured CM EMI Spectrum Adding DC Bus & Y-Capacitors CM Noise (dB V) 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Measured DM EMI Spectrum Adding DC Bus & Y-Capacitors DM Noise (dB V) Frequency (MHz)w/o DC Bus &
31、Y-Caps w/ DC Bus & Y-Caps28 Copyright 2006 Rockwell Automation, Inc. All rights reserved. Potential Reduction of Conducted EMI 0.15 0.3 1 3 10 30 -20 0 20 40 60 80 100 120 Simulated Total EMI Spectrum Noise (dB V) Frequency (MHz)Nominal EMI Filtered EMI Filtered/Shielded EMI CISPR A CISPR B29 Copyri
32、ght 2006 Rockwell Automation, Inc. All rights reserved. Conclusion Modeling process that predicts EMI noise within 10 dB Passive Active Interconnects Demonstrated the benefits of the modeling and characterized key parameters that affect the noise level Primary loop is dominant XFMR inter-winding capacitance Applied various attenuation techniques: simulated and measured Filtering Shielding
