Professional Documents
Culture Documents
Flyback Converter
SaberRD Design Example
This Design example demonstrates the EMI Analysis of a Flyback Converter. The design is tested for
various techniques to reduce common-mode noise. The common-mode noise from the converter is
compared with the CISPR and FCC standard limits to see if the design complies with the standards.
The transformer is modeled to include the capacitive coupling between windings and leakage
inductances. The use of FFT calculation, available in Saber, is demonstrated to check if the design
complies with the industry standards for EMI specifications. The use of AIM scripting language is
also introduced.
Synopsys Inc. |700 East Middlefield Road, Mountain View, CA 94043 | www.synopsys.com
EMI Analysis of Flyback Converter |SaberRD Design Example |
Contents
INTRODUCTION 2
FLYBACK TRANSFORMER 3
LINE IMPEDANCE STABILIZATION NETWORK (LISN) 3
LINE FILTER 3
CM AND DM NOISE MEASUREMENT 3
FFT IN SABERRD 5
RUNNING THE ANALYSIS 7
HOW TO RUN THE SIMULATION 7
RESULTS AND OBSERVATIONS 8
EMI ANALYSIS BY NOISE MEASUREMENTS USING FFT CALCULATION 8
CONCLUSION 11
REFERENCES 12
Introduction
Switch Mode Power Supplies are being used extensively because of the reduced size, weight, and
cost of the equipment. A Switch Mode Power Supply is a converter in which power switches
operated at high frequencies deliver the required electrical power in the required form. The recent
trends are increasing switching frequencies to reduce the size of the energy storage components
and improve the performance/efficiency of these power converters. The consequence of this trend
is the interference. The high voltages and currents in the converter topology are switched at high
frequencies resulting in Electromagnetic Interference (EMI) with other equipment in the
neighborhood.
Various government bodies have instituted standards, which set specific limits on the quantities of
radiated and conducted noise emissions in order for a product to be sold within that country. In the
United States, these regulatory bodies are the Federal Communications Commission (FCC) and the
Department of Defense (DOD). In Europe, all standards are set by the European Economic
Consortium (EEC). There is also an international body called the International Special Committee
on Radio Interference (CISPR), a committee of the International Electrontechnical Commission
(IEC), which has no regulatory authority but sets standards that can then be adopted by individual
nations in order to facilitate international trade. This example aims at testing the design for meeting
the FCC and CISPR standards for Class-A equipment.
The noise can be of Differential Mode (DM), where the circuit forms a closed path with out-of-phase
currents and applying Kirchoffs current Law in the loop gives zero. The other noise is the Common-
mode (CM) noise, where the circuit loop has in-phase noise currents.
This design example focuses on Conducted Emissions emanating from the Flyback Converter and
the filtering techniques being implemented to bring the CM noise within acceptable limits.
The converter is operated in open loop mode with the switch operated at 100kHz switching
frequency as the area of interest is the measurement of noise and not the closed loop operation of
the converter. The input is also a dc source though the actual source is ac. This is to reduce the
complexity in taking the circuit to steady state for noise measurements.
Flyback Transformer
This is modeled as a linear transformer with the turns ratio of 10:10:1 (primary : compensating :
secondary). The leakage inductances, bypass capacitances, cross coupling capacitance, and other
parasitic elements are modeled to allow for noise currents to pass between all three windings of the
transformer. The compensating winding is placed to aid for one of the noise filtering techniques
that will be explained in the following sections.
Line Filter
A common method to filter Common-mode noise is the CM choke in combination with the Y Caps.
The CM Choke has two windings, which are coupled to each other and cancels the CM noise. The Y
caps form the path to ground for the cancelled CM currents.
Figure 2: Conduction Emission Limits for FCC Part 15, Subpart J and CISPR, Publication 22.
The techniques implemented in the design example for reducing the common-mode noise are:
A common-mode choke coil is installed in the input section of the power supply line to suppress
common-mode noise. Since magnetic flux cancels out inside the ferrite core, impedance is not
produced for differential mode current. The magnetic saturation problem is small. Common-mode
choke coils are suited for common-mode noise suppression on lines with large current flow, such as
AC/DC power supply lines. Since they do not affect signal waveform, they are also suited for
common-mode noise suppression on lines where signal waveform distortion causes a problem,
such as video signal lines.
The Y-capacitor returns noise to the noise source in the following order: Y-capacitor metallic
casing stray capacitance noise source.
The values of L and C are selected in such a way that the magnitude of common-mode voltage at
lower frequencies is less than the limit set by the FCC and CISPR standards. As the frequency
increases, gain rolls off and the magnitude at higher frequencies will automatically be attenuated.
Passive Cancellation
A flyback converter can be modified to incorporate passive CM noise cancellation technique easily.
In this case, a compensation winding (NC) is added to the main power transformer/flyback
transformer. The NC uses small gauge wire as the currents that are intended into this winding are
just the noise currents and it would be a fairly cheap addition to the transformer. This winding can
be interwound with the primary winding for a 1:1 turns ratio.
The other added component is the compensation capacitor CCOMP. This is used to generate the anti-
noise current that will cancel out the parasitic noise current generated by C PARA (parasitic
capacitance). The value of CCOMP is determined by the size of CPARA and the turns ratio NP:NC. If this
ratio is 1:1, then CCOMP should be set equal to CPARA; otherwise it should be sized such that
FFT in SaberRD
It is required that the electrical signals be transformed to frequency spectrum to check with the
standards frequency spectrum limits. Snapshot in Figure 3 shows the location of FFT calculation in
Waveform Calculator.
To access the FFT calculation, select Analyze Tab Waveform Calculator Wave button FFT.
The output of this calculation is the frequency spectrum of the signal.
Upon giving the required waveform and selecting FFT calculation from the waveform calculator as
shown in Figure 3, a window shows up as seen in the Figure 4.
The tool allows selecting the number of points to be displayed/calculated. The Time start and Time
stop options are helpful to select the portion of the waveform, which is in steady state and has the
required details for running FFT calculation. In addition to selecting the Windowing function, the
tool can calculate the THD/SNR/SINAD values. Detailed information on FFT Calculation is provided
in SaberRD Documentation.
Provide the necessary values, and select the required options for performing FFT on the signal.
Click OK on this window, and click the waveform button on the waveform calculator. This shows up
the FFT waveform, which is the magnitude and phase plots.
In case of common-mode noise measurement, the waveform from the LISN is taken and the
common-mode voltage is measured. The frequency spectrum for this voltage in steady state is
obtained by the FFT calculation and the resultant waveform is compared with the standards shown
in Figure 2.
The process of taking the common-mode voltage signal to waveform calculator to getting a final
output showing a comparison with standard limits can be automated using the scripting language
called AIM. The AIM scripting language allows the users to create customized workflows/interfaces
so that the advanced features of SaberRD are utilized to perform common functionalities. In this
example, a script is available which when executed, gives graphs that clearly shows the frequency
spectrum of common-mode voltage compared with the FCC and CISPR standards limits. Details of
how the AIM scripting language can be used are given in SaberRD Documentation.
To automate the simulation process, an experiment is created to run all the four conditions. The
experiment is transient_analysis_for_fft.ai_expt. Once the analysis is complete, the
plotfiles are generated and saved on the working directory. To automate the process of analyzing
the waveforms for FFT and comparison with FCC and CISPR standards, an AIM script
(fft_proc.aim) is available. Run the AIM script and four graphs are created, which has plots of
common-mode noise compared with the limits of FCC and CISPR standards. A graph is plotted to
show the amplitudes of common-mode voltage waveforms for each filtering technique
implemented in the design.
source fft_proc.aim
7) The AIM script runs and five graphs shown in Figure 7, Figure 8, Figure 9, Figure 10 and
Figure 11 opens up.
It can be observed that the common-mode noise voltage amplitude is reducing with each filtering
technique, explained in the previous sections, inserted. Now, the next step is to analyze the
common-mode voltage noise level with respect to frequency.
In Figure 7, the measured Common-Mode voltage waveforms are plotted. It can be seen that the
peak-to-peak value is reducing when filters are introduced. The values are in line with the
Experiment Report shown in Figure 6.
Synopsys Inc. | EMI Analysis in SaberRD 8
EMI Analysis of Flyback Converter |SaberRD Design Example |
It can be seen in Figure 8 that the noise levels are beyond the limits defined by FCC and CISPR
standards. It is required that necessary filtering be provided to sell this product in the respective
markets.
From Figure 9, it can be seen that the noise levels reduced and the noise levels are marginally above
the limits of FCC and CISPR standards.
It is established in Figure 10 that the common-mode choke and Y capacitor filter which is designed
to reject the noise signals above the limit at lower frequencies is working fine. The gain at higher
frequencies rolls off at 40dB/dec after the corner frequency and hence the higher frequency noise
signals are attenuated to levels, which are far below the limits. The noise levels are marginally
below the FCC and CISPR limits before the cut-off frequency. It is always better to have some
margin so that any other parasitic elements may cause the noise levels to rise.
Finally, Figure 11 shows that the common-mode noise is having greater margin with respect to the
limits of FCC and CISPR standards. This is the best choice where the filters are optimally designed
and are making the design/equipment pass by a comfortable margin.
Conclusion
This design example details the procedure for performing EMI Analysis on power converters using
Flyback Converter topology. It is understood that the design without filtering does not pass the FCC
and CISPR standards under Class A equipment category. The two individual filtering techniques
show a reduction in the noise levels but when both the filtering techniques are employed, the noise
test is passed with comfortable margin.
SaberRD provides a very useful FFT calculation for this purpose and gives accurate results. The
waveforms from different plotfiles resulted from each analysis are automatically run for FFT and
the results are plotted in an easily readable format by using AIM script. Hence, SaberRD proves to
be a promising solution for running EMI Analysis.
References
1) Daniel Cochrane, Passive Cancellation of Common-Mode Electromagnetic Interference in
Switching Power Converters, M.Sc (EE) Thesis, Virginia Polytechnic Institute and State
University.
2) Module 11: Conducted Emissions -
http://www.egr.msu.edu/em/research/goali/notes/module11_conducted.pdf
3) Mark Steffka, Conducted Emissions and Power Supply Filters, IEEE EMC Society Central
Texas Chapter, January 19, 2011.