Presenters: Bill Spence Bob Teutsch Gentex EME Lab. IEEE SMPS for EMC

Transcription

1 Presenters: Bill Spence Bob Teutsch Gentex EME Lab 1

2 Part I: EMC Challenges Part II: Input Filtering Part III: Spread Spectrum 2

3 3

4 I. Conducted Emissions (Ignition & Ground): a) Voltage b) Current II. Radiated Emissions: a) Electric (E-Fields) b) Magnetic (H-Fields) 4

5 Common Impedance (Conductive) Coupling Capacitive/Inductive Coupling Electromagnetic Emissions 5

6 1. Input Filtering 2. Spread Spectrum (Dithering) 6

7 7

8 Buck switching regulators designs draw current in pulses. The peak amplitude of these current pulses is equal to the peak output current. There is significant high frequency EMI energy in these pulses. 8

9 Reduce the switching frequency EMI s amplitude by more than an order of magnitude And several orders of magnitude at higher harmonic frequencies. 9

10 1. To prevent electromagnetic interference generated by the switching source from conducting down the power line and affecting other equipment. 2. To prevent high-frequency voltage or currents on the power line from passing through the output of the power supply. 10

11 A SMPS should never be built without an EMI filter, even if you use very simple converters. Dr. Ray Ridley President Ridley Engineering Power Supply Expert and consultant to thousands of companies worldwide 11

12 Adding an input filter to a SMPS, as the U.S. Navy discovered, can make something go awry. Every SMPS built has the potential for input filter oscillation problems. 12

13 13

14 What is important to note about the input resistance of a SMPS is that this input resistance is negative. An increase in input voltage causes a decrease in input current. 14

15 When the SMPS is connected to the input filter, the negative input resistance of the converter combines with the positive damping resistance of the filter, and can result in complete elimination of any damping. The previously damped filter may now ring indefinitely with any perturbation. 15

16 If the input impedance of the power supply is much greater than the output impedance of the input filter, there will be no problems with stability of the power system. Dr. Ray Ridley 16

17 The practical measurement location of a power system The proper interaction location of a power system 17

18 The calculated negative input resistance? The true closed-loop input resistance? The open-loop input resistance? Which method? 18

19 We will use the first and simplest to calculate: The Negative Input Resistance. Why? Simple Calculation Not the optimum design but a safe design Component parasitic values are usually published Insensitive to changes in loop compensation 19

20 1. The SMPS s Input Resistance 2. The Input Filter s Output Resistance 20

21 21

22 The SMPS Negative Input Resistance is NOT constant and is Highly nonlinear. For modern SMPS it can be assumed that the efficiency is: Almost independent of the input voltage. High. 22

23 The input power, we can safely assume, will also be independent of the input voltage. 23

24 So, under the assumption of input power independent of input voltage: R in = -V min2 /P max Where: R in is the negative incremental input resistance V min is the minimum input voltage P max is the maximum output power 24

25 For a 5 Watt SMPS operating down to 8 Volts: R in = -8 2 /5 =

26 26

27 Parasitic values on all components are needed: Both series and shunt parasitic values on Inductors and ESR on Capacitors. 27

28 Capacitance vs. Bias, Package, Value 28

29 At a 5 Volts Bias, approximately 4 each 10µFd Capacitors would equal a single 22µFd Capacitor! 10µFd, 10v, 0805 X7R 22µFd, 10v, 1210 X7R 29

30 In this instance, Size does matter! At a 5 Volt Bias, 2 each 0805 Capacitors would equal a single 1206! 10uFd, 10v, 0805 X7R 10uFd, 10v, 1206 X7R 30

31 Same Package, Lower Capacitance Value has less Change. Check your Capacitors data sheets in your designs! 22uFd, 10v, 1206 X7R 10uFd, 10v, 1206 X7R 31

32 % Change in Capacitance vs. Rated Voltage 1µFd, 0805, X7R % Change in Capacitance Volts 16 Volts 25 Volts 35 Volts 50 Volts Bias Voltage 32

33 Ceramic Capacitors change capacitance with bias. Impact on change depends on both capacitance value and package. Know the actual capacitance value at designed bias point. A capacitor can be an X7R, as long as it meets the temperature coefficient specs, regardless of how bad the voltage coefficient is. 33

34 Include the parasitic elements for the inductor and capacitors. Short the input. Power the filter s output with an AC current source of 1.0 Amp. Set the Plot s amplitude scale to linear. 34

35 35

36 36

37 Comparison of the 33 H and 3.3 H Input Filter Designs: The 33 H Filter in Red has a resistance of approximately The 3.3 H Filter in Green has a resistance of approximately 1.1. We will see that the 3.3 H Filter will be the correct choice. 37

38 38

39 Our SMPS Input Impedance = Our Input Filter s Output Impedance: 11.5 for the 33µH inductor 1.1 for the 3.3µH inductor 39

40 The parallel combination of the Input Filter s Output Resistance (from LTSpice) and the calculated SMPS Input Resistance must be positive for stability. 40

41 A minimum 2 :1 SMPS Input Resistance to Filter Output Resistance margin is recommended. In the case of the example ( /2, the Input Filter s Output Resistance should be 6.4. From the LTSpice Simulations: The 33µH Filter s Output Resistance = 11.2 The 3.3µH Filter s Output Resistance =

42 11.5 > 6.4 Reject 42

43 1.1 < 6.4 Accept 43

44 5.00 mv/div, 5.00 us/div 1.00 V/div, 20.0 us/div 3.3µH Filter Design Oscillating 33µH Filter Design 44

45 45

46 The exact calculations are much more complex as the actual input resistance of the power supply is a function of Frequency However: Using this simplified design method will ensure the input filter will work with any reasonable loop-compensation selection for the power supply. 46

47 47

48 48

49 49

50 50

51 51

52 52

53 53

54 54

55 55

56 Slowing Switching Transitions Choosing a Switching Frequency to least disturb the system Snubbing High dv/dt nodes Minimizing high di/dt current loop areas 56

57 57

58 58

59 59

60 Miss Hedy Lamarr (November 9 th, 1913 January 19 th, 2000) The first practical conceptualization of the spread-spectrum technique is credited to the late silverscreen actress Hedy Lamarr, who, at the height of her Hollywood career in 1942, patented the Secret Communication System along with George Antheil (US Patent 2,292,387) 60

61 Typical Pulse Waveform Spectral envelope of waveform obtained via Fourier analysis 61

62 I use the relationship between risetime and bandwidth as given by: t rise =0.35/f 3dB. Where f 3dB is the bandwidth The amplitude of each harmonic of a given signal can be calculated Over 90% of the energy of a periodic trapezoidal signal is contained in the first ten harmonics Some EMC engineers define the minimum bandwidth of a trapezoidal waveform as ten times the fundamental frequency 62

63 Narrowband: [A]n electromagnetic disturbance, or spectral component thereof, which has a bandwidth less than or equal to that of a particular measuring apparatus, receiver, or susceptible device (IEV No ) Broadband: [A]n electromagnetic disturbance which has a bandwidth greater than that of a particular measuring apparatus, receiver or susceptible device (IEV No ). 63

64 Narrowband: Think of viewing a mouse through a windowpane. You see the whole mouse the whole signal. Broadband: Think of viewing an elephant through a windowpane. You see only a portion of the elephant only a portion of the signal. 64

65 65

66 Signal passes through an attenuator and a low-pass filter Signal mixes with a signal from a swept local oscillator Sums, differences, and harmonics signals pass though the IF filter are amplified, rectified and displayed. 66

67 The principle blocks related to harmonic-peak detection are the mixer, the IF filter and the envelope detector. If a low enough resolution bandwidth is selected, dithering the switching frequency must translate into a harmonic-peak reduction. Why: As the power spectral density decreases, the narrow band noise becomes broadband noise: sideband energy is introduced, the noise floor increases and the harmonic peaks associated with the characteristic wave are reduced. 67

68 The classification of a signal as narrowband or broadband is determined by the occupied frequency spectrum of the signal under test relative to the resolution bandwidth (RBW) of the instrument used for measurement. 68

69 69

70 Where: BT is the bandwidth of the modulated signal is the FM modulation-index (Δf/f m ) 70

71 FM will produce an infinite number of sidebands. Modulating the switching frequency within a limited range creates sidebands. These sidebands decrease the spectral density and lower the harmonic peaks. 71

72 The total power of a signal is unaffected by the frequency modulation. It is equal to the sum of the squares of each harmonic 98% of the total power of an FM signal is contained inside the bandwidth 2(df+f m ) 72

73 73

74 120 khz Res Bandwidth 50 khz Res Bandwidth 3 khz Res Bandwidth 74

75 120 khz Res Bandwidth 50 khz Res Bandwidth 3 khz Res Bandwidth 75

76 120 khz RBW 50 khz RBW 3 khz RBW *RBW: Resolution Bandwidth 120 khz RBW and 3 khz RBW are Wideband sweeps with Spread- Spectrum Modulation enabled. Note the peak reduction of 120 khz vs. 3 khz RBW. The 50 khz RBW is a Narrowband plot with Spread-Spectrum disabled. From this information, the 120 khz RBW plot could well be considered a Narrowband plot. 76

77 Reducing the Resolution Bandwidth attenuates the signal sidebands. Removing the sidebands reduces the power to the detector. The reduced power is measured as a reduced signal strength. 77

78 Low End: Modulation must be greater than the resolution bandwidth specified in the EMC standard. Spreading artifacts below 20 khz could manifest itself as audible noise due to magnetostriction and piezoelectric effects. High End: Modulation switching frequency should not be greater than the loop-crossover frequency of the feedback-control loop as this could impact loop stability and output-ripple voltage. 78

79 SMPS designs use: Linear (Triangle Wave) Sinusoidal Square Wave RC Filtered Square Wave (The energy in the spectrum of the modulated waveform will tend to concentrate at those frequencies corresponding to point in the modulated waveform where the time derivative is small. The triangular modulation waveform is the best compromise.) 79

80 Modulation Spreading can be: Up Spreading Center Spreading Down Spreading 80

81 81

82 10 khz Triangle Wave 10 khz Sine Wave 82

83 10 khz Square Wave 10 khz RC filtered Square Wave 83

84 10 khz Triangle Wave 20 khz Triangle Wave 84

85 30 khz Triangle Wave 40 khz Triangle Wave 85

86 To make the spectrum flatter, a special curve, named a Hershey Kiss TM, is used as the modulation waveform. 86

87 87

88 Impact on output voltage ripple Tendency to raise the noise floor Potential for audible noise Magnetics must be designed to operate over the frequency deviation introduced 88

89 Magnetic Emissions are reduced to below limits Residual emissions are less audible. 89

90 Spread-Spectrum dithering can be an effective tool for reducing the spectral-peak energy. The Modulation waveform modulation frequency, and depth of modulation can uniquely impact spectral-harmonic spreading and must be carefully selected in accordance with the application and well understood principles of FM modulation. 90

91 Effects of Input Filter Effects of Spread Spectrum 91

92 Understanding Noise-Spreading Techniques and their Effect in Switch-Mode Power Applications, John Rice, Dirk Gehrke, and Mike Segal, Unitrode Power-Supply Design Seminar SEM1800, 2008/2009, Available online at Impact of PCB Design on Switching noise and EMI of Synchronous DC-DC buck Converter, Kyoungchoul Koo, Jiseong Kim, Myunghor Kim, and Joungho Kim, EECS, Korea Advanced Institute of Science and Technology, (KAIST) Gusongdong, Yusong-gu, Daejeon, Korea Simple Switcher PCB Layout Guidelines, Sanjaya Maniktala, National Semiconductor Application Note 1229, July 2002 Investigation of Noise Coupling from Switching Power Supply to Signal Nets, Songping Wu, Keon Kam, David Pommerenke, Bill Cornelius, Hao Shi, Matthew Hendron, and Jun Fan, Electromagnetic Compatibility Laboratory, Department of Electrical and Computer, Missouri University of Science and Technology 92

93 Reducing Ringing Through PCB Layout Techniques, David Jauregui, Application Report SLPA005-June 2009, Texas Instruments Spread Spectrum Techniques to Reduce EMI in SMPS Devices, Matthew Majeika, Application Note AND8428/D, ON Semiconductor Understanding Noise-Spreading Techniques and their Effects in Switch-Mode Power Applications, John Rice, Kirk Gehrke, and Mike Segal Minimizing Input filter Requirements in Military Power Supply Design, Venable Technical Paper #4, Venable Industries 93

94 DC-DC Buck Converter EMI Reduction Using PCB Layout Modification, Ankit Bhargava, David Pommerenke, Keong W. Kam, Ferderico Centola, and Cheng Wei Lam, IEEE Transactions on Electromagnetic Compatibility, Vol. 53, No. 3, August 2011 EMI Specifics of Synchronous DC-DC Buck Converters,Zhe Li, David Pommerenke, Electromagnetic Compatibility Laboratory Department of Electrical and Computer Engineering, University of Missouri-Rolla, Rolla MO Modeling EMI Problems Associated with DC-DC Converters, Eleonora Darie, Costin Cepisca, Emanuel Darie, Electrotechnical Department, Technical University of Civil Engineering, Bucharest, Romania, Electrotechnical Department, University Pollitehnica of Bucharest, Bucharest, Romania, Engineering Department, Police Academy Bucharest, Bucharest, Romania 94

95 The Evolution of Power Electronics, Switching Power Magazine, Fall 2001 Analysis and Mitigation Techniques for Broadband EMI from Synchronous Buck Converter, Keong Kam, David Pommerenke, Ankit Bhargave, EMC Laboratory, Missouri University of Science and Technology, 2012 IEEE Electromagnetic Compatibility Magazine, Volume 1, Quarter 3 PCB Layout for Switchers, 2010, National Semiconductor Reducing Ground Bounce in DC-to-DC Converters Some Grounding Essentials, Jeff Barrow, Analog Dialogue 41-06, June (2007) 95

96 Switch-Mode Power Supplies, SPICE Simulations and Practical Designs, Christophe P. Basso, McGraw-Hill, 2008 Layout Guidelines for Switching Power Supplies, Clinton Jensen, National Semiconductor Application Note 1149, October 1999 Switcher Efficiency & Snubber Design, On Semiconductor Power Point Presentation Negative Input Resistance and RMS Input Currents, powerguru, 96

97 EMC Guideline for Synchronous Buck Converter Design, Keon Kam, David Pommerenke, Federico, Centola, Cheung-wei Lam, Robert Steinfeld, EMC Laboratory, Missouri University of Science and Technology, Rolla, MO Switching Power Supplies A to Z, Sanjaya Maniktala, Elsevier Inc Switch-Mode Power Supplies, SPICE Simulations and Practical Designs, Christophe P. Basso, McGraw Hill, 2008 Switchmode Power Supply Handbook, Keith Billings, McGraw Hill, Second Edition,

98 Switching Power Supply Design, Abraham I. Pressman, McGraw Hill, 1998 Practical Design Of Power Supplies, Ron Lenk, John Wiley and Sons, 2005 Switch-Mode Power Supply SPICE Cookbook, Christophe P. Basso, McGraw Hill, 2001 Input Filter Design for Switching Power Supplies, Michele Sclocchi, Application Engineer, National Semiconductor, 2010 National Semiconductor Corporation Simulations from Scott Piper Bandwidth and Risetime, Tim J. Sobering, SDE Consulting, Technote 2, May

99 99