Academia.eduAcademia.edu

Figure 52 – uploaded by dongsheng zhao

See full PDFdownloadDownload figure
Figure 5-19 shows a model of a motor in high frequency. We use switch to represent commutation. In simulation, we mark the housing of the motor as reference, and every coil inside the motor is floating to reference. In time domain, the highest peak of voltage in the coil terminal appears when motor leaving commutation. The spectrum of commutation noise is shown at Figure 5-20. commutation. The spectrum of commutation noise is shown at Figure 5-20.

Figure 5 -19 shows a model of a motor in high frequency. We use switch to represent commutation. In simulation, we mark the housing of the motor as reference, and every coil inside the motor is floating to reference. In time domain, the highest peak of voltage in the coil terminal appears when motor leaving commutation. The spectrum of commutation noise is shown at Figure 5-20. commutation. The spectrum of commutation noise is shown at Figure 5-20.

Related Figures (106)

Table 1-1: International and regional standards used in automotive industry Some more requirements are imposed by product manufacturers themselves for customer satisfaction. It is also called automotive Original Equipment Maker (OEM) standards. These standards are not easily accessible because they are classified. They are imposed for the purpose of insuring a reliable, quantity product. These stricter regulations are used to obtain a good reputation from the standpoint of quality control of its product. Almost each automotive manufacturer has his own additional regulation. In NEDAP, a test handbook summarized from variant standards is used for validation test. Configurations, specifications, and procedures are abstracted so that the test items and criteria are fixed during the test. The reason is that some regulations give several options for the same measurement items, and some organizations give regulations on
Table 1-1: International and regional standards used in automotive industry Some more requirements are imposed by product manufacturers themselves for customer satisfaction. It is also called automotive Original Equipment Maker (OEM) standards. These standards are not easily accessible because they are classified. They are imposed for the purpose of insuring a reliable, quantity product. These stricter regulations are used to obtain a good reputation from the standpoint of quality control of its product. Almost each automotive manufacturer has his own additional regulation. In NEDAP, a test handbook summarized from variant standards is used for validation test. Configurations, specifications, and procedures are abstracted so that the test items and criteria are fixed during the test. The reason is that some regulations give several options for the same measurement items, and some organizations give regulations on
Figure 1-2: Designable Parameters determine EMI level These designable parameters determine how severely a transient contributes to noise
Figure 1-2: Designable Parameters determine EMI level These designable parameters determine how severely a transient contributes to noise
The vehicle sunroof system normally consists of three main parts. That is, the sunroof 2.1 Structure of sunroof system of the glass panel. The sunroof structure is fixed in the car body for safety reason. Vehicle sunroof system
The vehicle sunroof system normally consists of three main parts. That is, the sunroof 2.1 Structure of sunroof system of the glass panel. The sunroof structure is fixed in the car body for safety reason. Vehicle sunroof system
Figure 2-2: Electrical connection of a typical sunroof system
Figure 2-2: Electrical connection of a typical sunroof system
Figure 3-1 shows how one line segment with the length Az is approximated with an electric circuit. Figure 3-1: The equivalent circuit of one line segment of two-conductor line can be replaced with a lumped-circuit model which is related to per-unit-length
Figure 3-1 shows how one line segment with the length Az is approximated with an electric circuit. Figure 3-1: The equivalent circuit of one line segment of two-conductor line can be replaced with a lumped-circuit model which is related to per-unit-length
Here, C; and C, are added to represent the capacitances between one conductor and the reference, and C; and C, are the capacitances between another conductor and reference. The inductance of the reference is ignored here. The mutual inductance between two conductors is also not considered, because the main issue is how DM current between two conductors introduces CM current in reference. If the wavelength of the highest frequency component from signal source is much longer than the largest dimension of transmission line, we say that the line is
Here, C; and C, are added to represent the capacitances between one conductor and the reference, and C; and C, are the capacitances between another conductor and reference. The inductance of the reference is ignored here. The mutual inductance between two conductors is also not considered, because the main issue is how DM current between two conductors introduces CM current in reference. If the wavelength of the highest frequency component from signal source is much longer than the largest dimension of transmission line, we say that the line is
Figure 3-3: A sample motor cable resistance versus frequency Figure 3-3 shows the wire resistance versus frequency of a sample motor cable. 3.2.2 Inductance The wire inductance plays a crucial role in determining the specification of circuit in
Figure 3-3: A sample motor cable resistance versus frequency Figure 3-3 shows the wire resistance versus frequency of a sample motor cable. 3.2.2 Inductance The wire inductance plays a crucial role in determining the specification of circuit in
Figure 3-4: Inductance between two parallel wires
Figure 3-4: Inductance between two parallel wires
Figure 3-5: Inductance between wire and metal plane
Figure 3-5: Inductance between wire and metal plane
Figure 3-6: Wire inductance of a sample cable versus frequency
Figure 3-6: Wire inductance of a sample cable versus frequency
Figure 3-7: The Ratio of wire inductance with a current probe to without a probe These two experiments explain why there exists a deviation between prediction and
Figure 3-7: The Ratio of wire inductance with a current probe to without a probe These two experiments explain why there exists a deviation between prediction and
Figure 3-8: One element of a periodic trapezoidal waveform The PWM signal is a periodic trapezoidal wave as Figure 3-8 shows.
Figure 3-8: One element of a periodic trapezoidal waveform The PWM signal is a periodic trapezoidal wave as Figure 3-8 shows.
Figure 3-9: Asymptotes for Spectrum of trapezoidal waveform Figure 3-9 shows two asymptotes of signal spectrum for a trapezoidal wave. By changing some designable parameters, we know the effects of these parameters.
Figure 3-9: Asymptotes for Spectrum of trapezoidal waveform Figure 3-9 shows two asymptotes of signal spectrum for a trapezoidal wave. By changing some designable parameters, we know the effects of these parameters.
Figure 3-10: One element of a periodic triangular waveform Another special waveform we discussed here is the triangular waveform, which is very
Figure 3-10: One element of a periodic triangular waveform Another special waveform we discussed here is the triangular waveform, which is very
For Fourier series, we have the expansion coefficient as front edge is extremely short, normally 100ns. The fall edge follows an exponentia
For Fourier series, we have the expansion coefficient as front edge is extremely short, normally 100ns. The fall edge follows an exponentia
Table 3-1 Requirement for radiated emission Radiated emissions are the electric and magnetic fields radiated by one device that may be received by other devices. Although radiated emissions include both electric and magnetic fields, regulatory agencies only require that electric fields be measured for certification. According to the standards [22], [23] and [24] of automotive industry, the concerned frequency range starts from 150kHz to 1GHz. A radiated emission limit is shown in Table 3-1.
Table 3-1 Requirement for radiated emission Radiated emissions are the electric and magnetic fields radiated by one device that may be received by other devices. Although radiated emissions include both electric and magnetic fields, regulatory agencies only require that electric fields be measured for certification. According to the standards [22], [23] and [24] of automotive industry, the concerned frequency range starts from 150kHz to 1GHz. A radiated emission limit is shown in Table 3-1.
Figure 3-7: The predominant contribution of near-field and far-field component
Figure 3-7: The predominant contribution of near-field and far-field component
For an elemental magnetic dipole, the field intensity vector is [27], 3.4.3 Radiated emissions model
For an elemental magnetic dipole, the field intensity vector is [27], 3.4.3 Radiated emissions model
In a real circuit, stray capacitance between two terminators of load needs to be considered. Figure 4-2 shows the improved model. 4.2 Real switch modeling Here, V; is the voltage between two terminals of the switch, and 7 is the current flowing along the loop. When switch U, is closing, the current will exponentially rise up to V,/R; with a time constant of L;/R;. In another transient, when the switch is opened with the assumption that transient time is zero, the loop current i abruptly changes to 0. At that moment, voltage over inductance L/(di/d?) is infinitely large, and voltage V; over the switch is also infinitely large to make Equation 4.1 balance. No component in real life can sustain infinite voltage, so it must be a different case for real circuit.
In a real circuit, stray capacitance between two terminators of load needs to be considered. Figure 4-2 shows the improved model. 4.2 Real switch modeling Here, V; is the voltage between two terminals of the switch, and 7 is the current flowing along the loop. When switch U, is closing, the current will exponentially rise up to V,/R; with a time constant of L;/R;. In another transient, when the switch is opened with the assumption that transient time is zero, the loop current i abruptly changes to 0. At that moment, voltage over inductance L/(di/d?) is infinitely large, and voltage V; over the switch is also infinitely large to make Equation 4.1 balance. No component in real life can sustain infinite voltage, so it must be a different case for real circuit.
Figure 4-5: The waveform of oscillation when P;=10 Figure 4-5 shows the waveform of oscillation when P;=10.
Figure 4-5: The waveform of oscillation when P;=10 Figure 4-5 shows the waveform of oscillation when P;=10.
Figure 4-4 shows the waveform of oscillation when P;=1. Figure 4-4: The waveform of oscillation when P;=1
Figure 4-4 shows the waveform of oscillation when P;=1. Figure 4-4: The waveform of oscillation when P;=1
Figure 4-6: The waveform of oscillation when P;=25
Figure 4-6: The waveform of oscillation when P;=25
Figure 4-7: Comparison of the waveform of oscillation when P,=20 and 2 Similar criterion can be achieved by simulation. We define P2 as
Figure 4-7: Comparison of the waveform of oscillation when P,=20 and 2 Similar criterion can be achieved by simulation. We define P2 as
Figure 4-8: Comparison of the waveform of oscillation with different current variation
Figure 4-8: Comparison of the waveform of oscillation with different current variation
The setup of the experiment is shown as Figure 4-9 below. Figure 4-9: Experiment setup to verify switch model We set up an experiment for verification. It is done with a real Printed Circuit Board (PCB) board. A LISN is inserted to make impedance stable, and a relay driver is controlled by switch. We only consider the voltage ripple introduced by a switch. The mechanical characteristic makes the rotor inside the motor start rotating 7-8 ms after the relay is closed. Therefore, the influence of the motor can be ignored at the switch transient time. The input resistance of relay driver is replaced with a resistor. The cetiin af the eyvneriment ic chown ac Fionre 4A_O helaw,
The setup of the experiment is shown as Figure 4-9 below. Figure 4-9: Experiment setup to verify switch model We set up an experiment for verification. It is done with a real Printed Circuit Board (PCB) board. A LISN is inserted to make impedance stable, and a relay driver is controlled by switch. We only consider the voltage ripple introduced by a switch. The mechanical characteristic makes the rotor inside the motor start rotating 7-8 ms after the relay is closed. Therefore, the influence of the motor can be ignored at the switch transient time. The input resistance of relay driver is replaced with a resistor. The cetiin af the eyvneriment ic chown ac Fionre 4A_O helaw,
Figure 4-10: Simulation diagram for a switch model
Figure 4-10: Simulation diagram for a switch model
Figure 4-11: Switch model verification simulation result (100ps/div) Figure 4-12: Switch model verification measurement result (100us/div)
Figure 4-11: Switch model verification simulation result (100ps/div) Figure 4-12: Switch model verification measurement result (100us/div)
Figure 4-14: Switch model verification measurement result (10ps/div)
Figure 4-14: Switch model verification measurement result (10ps/div)
similar between simulation and measurement. Let us expand the time scale further. Figure 4-15: Switch model verification simulation result (1ps/div)
similar between simulation and measurement. Let us expand the time scale further. Figure 4-15: Switch model verification simulation result (1ps/div)
Figure 4-16: Switch model verification measurement result (1ps/div) From above figures, we found the high frequency component fy appears as predicted. Doing FFT transform upon the signal, we got the spectrum of the noise source which is shown below. Two resonance frequencies can be found easily. They are 200 kHz and 4.5 MHz.
Figure 4-16: Switch model verification measurement result (1ps/div) From above figures, we found the high frequency component fy appears as predicted. Doing FFT transform upon the signal, we got the spectrum of the noise source which is shown below. Two resonance frequencies can be found easily. They are 200 kHz and 4.5 MHz.
Figure 4-17: Switch model verification simulation result in frequency domain 4.4 Conclusion
Figure 4-17: Switch model verification simulation result in frequency domain 4.4 Conclusion
Figure 5-1 to 5-3 shows pictures for each part of a motor. commutator
Figure 5-1 to 5-3 shows pictures for each part of a motor. commutator
Figure 5-4: The suppression circuit and its location grounding suppression capacitors. A typical suppression circuit is illustrated in Figur 5.2 Coil model housing. Here we call the shunt one as shunt suppression capacitor, the other two as Of course, primary modeling should be based on coils, but not on turns. Some references [15] deduce a model by assuming coil made of turns connected in series, and each turn is represented by an equivalent circuit including mutual effects be However, when a motor wit it is a really hard approach o do it respectively. Analysis starting with coi ween turns. h a different structure and coil geometry is adopted at SCU. s is not too insignificant and will include enough phenomena we need to concern. Another advantage of dividing a mo or into coils is that the running time for simu shorter with a high turn number motor. — ii? 7 +14 “a « 4 41 4 ce 4 7 eqyoie ation is far
Figure 5-4: The suppression circuit and its location grounding suppression capacitors. A typical suppression circuit is illustrated in Figur 5.2 Coil model housing. Here we call the shunt one as shunt suppression capacitor, the other two as Of course, primary modeling should be based on coils, but not on turns. Some references [15] deduce a model by assuming coil made of turns connected in series, and each turn is represented by an equivalent circuit including mutual effects be However, when a motor wit it is a really hard approach o do it respectively. Analysis starting with coi ween turns. h a different structure and coil geometry is adopted at SCU. s is not too insignificant and will include enough phenomena we need to concern. Another advantage of dividing a mo or into coils is that the running time for simu shorter with a high turn number motor. — ii? 7 +14 “a « 4 41 4 ce 4 7 eqyoie ation is far
Figure 5-5: Equivalent inductance versus frequency of one coil in DC motor
Figure 5-5: Equivalent inductance versus frequency of one coil in DC motor
Figure 5-6: High frequency equivalent circuit of a coil (1)
Figure 5-6: High frequency equivalent circuit of a coil (1)
The curves of impedance we measured are shown at Figure 5-7. SS OT eS Choosing the value of these components for model obviously is important and difficul MATLAB is used to compare the calculated impedance value to measurement resu By a curve fitting, we update the coil model with adjusted values. This procedure called “back-annotation” method. It can also be done by numerical techniques. W increase the density of frequency points around resonance frequency to get a more dete Because winding of the coils and distribution of parasitic capacitance are not totally tl same, resonance frequency is not exactly the same between impedance test results. TI deviation is within 10%. Here we use a typical test result to model the coil. The curves af imnedance we meaaired are chawn at Fioure S_7 deviation is within 10%. Here we use a typical test result to model the coil.
The curves of impedance we measured are shown at Figure 5-7. SS OT eS Choosing the value of these components for model obviously is important and difficul MATLAB is used to compare the calculated impedance value to measurement resu By a curve fitting, we update the coil model with adjusted values. This procedure called “back-annotation” method. It can also be done by numerical techniques. W increase the density of frequency points around resonance frequency to get a more dete Because winding of the coils and distribution of parasitic capacitance are not totally tl same, resonance frequency is not exactly the same between impedance test results. TI deviation is within 10%. Here we use a typical test result to model the coil. The curves af imnedance we meaaired are chawn at Fioure S_7 deviation is within 10%. Here we use a typical test result to model the coil.
Figure 5-9: High frequency model of one coil in DC motor
Figure 5-9: High frequency model of one coil in DC motor
Figure 5-8: High frequency equivalent circuit of a coil (2)
Figure 5-8: High frequency equivalent circuit of a coil (2)
Figure 5-12: How common mode current is generated from motor cable.
Figure 5-12: How common mode current is generated from motor cable.
Figure 5-14: Simulation result of coil current Here we get the prediction of the current flowing into LZ; and L). Now, we give a derivation using circuit theory. We also split the waveform into 4 phases: in phase (a), current of L; is increasing; in phase (b), current starts to decrease; in phase (c), current keeps on decreasing; in phase (d), current is suddenly changed. We expand it in time axis to make it better visible.
Figure 5-14: Simulation result of coil current Here we get the prediction of the current flowing into LZ; and L). Now, we give a derivation using circuit theory. We also split the waveform into 4 phases: in phase (a), current of L; is increasing; in phase (b), current starts to decrease; in phase (c), current keeps on decreasing; in phase (d), current is suddenly changed. We expand it in time axis to make it better visible.
Figure 5-15: Phases of Current transient Two asymptotes illustrated by a dashed line represent two ultimate current values if the
Figure 5-15: Phases of Current transient Two asymptotes illustrated by a dashed line represent two ultimate current values if the
The simulation results of a current in the motor cable are shown at Figure 5-17. Figure 5-17: Simulation result of Differential Mode current
The simulation results of a current in the motor cable are shown at Figure 5-17. Figure 5-17: Simulation result of Differential Mode current
Figure 5-18: Measured DM Current with 3m cable (0.2A/div 0.5ms/div)
Figure 5-18: Measured DM Current with 3m cable (0.2A/div 0.5ms/div)
-igure 5-20: Commutation noise spectrum without suppression (simulation)
-igure 5-20: Commutation noise spectrum without suppression (simulation)
Figure 5-21: Commutation noise converts to CM current (measured)
Figure 5-21: Commutation noise converts to CM current (measured)
Figure 5-22: Commutation noise converts to DM current (measured)
Figure 5-22: Commutation noise converts to DM current (measured)
Figure 5-23: Commutation noise spectrum with different ground suppression capacitor (simulation) Figure 5-23 shows three curves, the highest one is the noise level with 47pF grounding loop area which makes it radiates inefficiently. Noise caused by DM current is ignored,
Figure 5-23: Commutation noise spectrum with different ground suppression capacitor (simulation) Figure 5-23 shows three curves, the highest one is the noise level with 47pF grounding loop area which makes it radiates inefficiently. Noise caused by DM current is ignored,
Figure 5-24: Commutation noise spectrum with different shunt suppression capacitor (simulation) two curves, the higher one is the noise level with 470nF, and the lower one is simulation Effect of increasing shunt suppression capacitor is also evaluated. Figure 5-24 shows
Figure 5-24: Commutation noise spectrum with different shunt suppression capacitor (simulation) two curves, the higher one is the noise level with 470nF, and the lower one is simulation Effect of increasing shunt suppression capacitor is also evaluated. Figure 5-24 shows
in time domain. resonant mode in measurement. This is a potential noise source. The motor case,
in time domain. resonant mode in measurement. This is a potential noise source. The motor case,
Figure 5-26: Mechanism of CM current in grounding network
Figure 5-26: Mechanism of CM current in grounding network
Figure 5-27: Conducted Emission Transient when motor is switch off suddenly The transient noise when motor is switch off has a much more frequency content than
Figure 5-27: Conducted Emission Transient when motor is switch off suddenly The transient noise when motor is switch off has a much more frequency content than
Figure 6-1: Setup to validate behaviour of the relay PSPICE component library, while the model of the motor is the result of previous
Figure 6-1: Setup to validate behaviour of the relay PSPICE component library, while the model of the motor is the result of previous
Figure 6-2: Three resonant modes established with relay closing
Figure 6-2: Three resonant modes established with relay closing
Figure 6-3: Motor circuit with resonant loop when relay opening
Figure 6-3: Motor circuit with resonant loop when relay opening
Figure 6-5: Simulated CM emission spectrum with motor housing grounded
Figure 6-5: Simulated CM emission spectrum with motor housing grounded
Figure 6-4: Simulated time domain transient of CM current
Figure 6-4: Simulated time domain transient of CM current
Figure 6-8: Simulated CM emission spectrum with increased grounding wire inductance The resonance /4, f; disappear and /; decrease to 6.3MHz. As shown in this figure, th
Figure 6-8: Simulated CM emission spectrum with increased grounding wire inductance The resonance /4, f; disappear and /; decrease to 6.3MHz. As shown in this figure, th
10 Q, the simulated CM emission spectrum is shown in Figure 6-9. Figure 6-9: Simulated CM emission spectrum with grounding wire impedance increased Sa, fs are decaying rapidly, so no oscillation is constructed, f5 is also suppressed. In fact, wire resistance is in series connected in the loop3 and loop6, the amplitude of /4, fs is decreased. CM current along the ground wire converted from potential difference in loop5 is also decreased by increased wire impedance. 4. The grounding suppression capacitor. A a ee ce Be DR ee ee I ee ec ee ee ce Be ote ASR TS ate = a hee 1 ew 2 OVER A cg hee 2 fa, fs are decaying rapidly, so no oscillation is constructed, f¢ is also suppressed. In fact
10 Q, the simulated CM emission spectrum is shown in Figure 6-9. Figure 6-9: Simulated CM emission spectrum with grounding wire impedance increased Sa, fs are decaying rapidly, so no oscillation is constructed, f5 is also suppressed. In fact, wire resistance is in series connected in the loop3 and loop6, the amplitude of /4, fs is decreased. CM current along the ground wire converted from potential difference in loop5 is also decreased by increased wire impedance. 4. The grounding suppression capacitor. A a ee ce Be DR ee ee I ee ec ee ee ce Be ote ASR TS ate = a hee 1 ew 2 OVER A cg hee 2 fa, fs are decaying rapidly, so no oscillation is constructed, f¢ is also suppressed. In fact
Figure 6-11: Simulated CM emission spectrum with switching time increased
Figure 6-11: Simulated CM emission spectrum with switching time increased
Figure 6-12 Simulated CM emission spectrum with shunt suppression capacitor increased
Figure 6-12 Simulated CM emission spectrum with shunt suppression capacitor increased
simulated CM emission spectrum changes a lot in high frequency as Figure 6-13 shows. Figure 6-13 Simulated CM emission spectrum with wire inductance of motor cable decreased
simulated CM emission spectrum changes a lot in high frequency as Figure 6-13 shows. Figure 6-13 Simulated CM emission spectrum with wire inductance of motor cable decreased
Figure 6-14 Simulated CM emission spectrum with another varistor
Figure 6-14 Simulated CM emission spectrum with another varistor
Figure 6-15 Simulated CM emission spectrum with battery input capacitor increased
Figure 6-15 Simulated CM emission spectrum with battery input capacitor increased
set via parameter in the switch model which is measured in experiment. The switching device, a MOSFET, was modeled in this study as a switch parallel with two capacitors, as shown in Figure 7-1. The junction capacitor C; involved in the reverse recovery current and the capacitance C;, as the parasitic capacitance between drain and ground plane because drain has biggest voltage sudden change and soldering area. Both C; and C; can be obtained by measuring. The turn on/off time of the switch is set via parameter in the switch model which is measured in experiment. Figure 7-1: Model of MOSFET for high frequency
set via parameter in the switch model which is measured in experiment. The switching device, a MOSFET, was modeled in this study as a switch parallel with two capacitors, as shown in Figure 7-1. The junction capacitor C; involved in the reverse recovery current and the capacitance C;, as the parasitic capacitance between drain and ground plane because drain has biggest voltage sudden change and soldering area. Both C; and C; can be obtained by measuring. The turn on/off time of the switch is set via parameter in the switch model which is measured in experiment. Figure 7-1: Model of MOSFET for high frequency
We use Figure 7-2 to illustrate the transient of PWM switch.
We use Figure 7-2 to illustrate the transient of PWM switch.
Figure 7-3: Main DM current loops when SCU is working at PWM mode The measurement results in time domain of these three DM currents are shown in Figure 7-4. In phase 1, MOSFET is switched on firstly. The current from the battery flows to the motor then through MOSFET and returns to the battery. In phase 2, the MOSFET is urned off, the current in the inductors cannot change instantaneously, the voltage polarity across the inductors will be reversed due to electromagnetic force when di/dt <C his voltage will forward bias diode, then the diode take the current loop cover. The block in Figure 7-2 is embrace with high di/d¢ current, which emits radiation efficiently ike a loop antenna. We call the current in this loop as diode current. Besides diode current, current with ripple is produced between power supplier and SCL between SCU and motor when motor is running at PWM mode. Their locations are
Figure 7-3: Main DM current loops when SCU is working at PWM mode The measurement results in time domain of these three DM currents are shown in Figure 7-4. In phase 1, MOSFET is switched on firstly. The current from the battery flows to the motor then through MOSFET and returns to the battery. In phase 2, the MOSFET is urned off, the current in the inductors cannot change instantaneously, the voltage polarity across the inductors will be reversed due to electromagnetic force when di/dt <C his voltage will forward bias diode, then the diode take the current loop cover. The block in Figure 7-2 is embrace with high di/d¢ current, which emits radiation efficiently ike a loop antenna. We call the current in this loop as diode current. Besides diode current, current with ripple is produced between power supplier and SCL between SCU and motor when motor is running at PWM mode. Their locations are
Figure 7-4: The DM current waveform when motor running in PWM mode (measurement) We found a small ripple in the falling edge of the waveform of diode current. It is caused by a new resonant mode which does not exist in real product. This resonant mode is produced by an extend wire inserted into diode current loop to contain the clamp of current probe. We found a small ripple in the falling edge of the waveform of diode current. It i
Figure 7-4: The DM current waveform when motor running in PWM mode (measurement) We found a small ripple in the falling edge of the waveform of diode current. It is caused by a new resonant mode which does not exist in real product. This resonant mode is produced by an extend wire inserted into diode current loop to contain the clamp of current probe. We found a small ripple in the falling edge of the waveform of diode current. It i
The diode current has a highest level in these DM signals. It starts from 100dB LE A at The DM current at motor cable is 70dBu A at 400kHz and follows a faster decay slope
The diode current has a highest level in these DM signals. It starts from 100dB LE A at The DM current at motor cable is 70dBu A at 400kHz and follows a faster decay slope
These DM current can be predicted by simulation as Figure 7-6 shows.
These DM current can be predicted by simulation as Figure 7-6 shows.
The simulation and measurement result of diode current are compared at Figure 7-7. Figure 7-7: Diode current comparison between measurement and simulation The amplitudes of two waveforms are 1.36A and 1.2A. The shape is slightly different because current probe has less transfer impedance in low frequency. The decay time constants are 3us and 4.2us respectively which are still in the same order. Both waveforms have rise time about 100ns. The predicted diode current agrees well with the measurement result in the frequency range up to 3MHz. Therefore, the diode current is predictable. Via the formula of Equation 3.17 and Equation 3.18, we can predict radiated emission
The simulation and measurement result of diode current are compared at Figure 7-7. Figure 7-7: Diode current comparison between measurement and simulation The amplitudes of two waveforms are 1.36A and 1.2A. The shape is slightly different because current probe has less transfer impedance in low frequency. The decay time constants are 3us and 4.2us respectively which are still in the same order. Both waveforms have rise time about 100ns. The predicted diode current agrees well with the measurement result in the frequency range up to 3MHz. Therefore, the diode current is predictable. Via the formula of Equation 3.17 and Equation 3.18, we can predict radiated emission
Figure 7-9: Measured radiated emission with setup1
Figure 7-9: Measured radiated emission with setup1
Figure 7-10: Mechanism of resonance mode in motor cable
Figure 7-10: Mechanism of resonance mode in motor cable
mechanism is shown at Figure 7-11 and 7-12. Figure 7-11: Two Unbalance Wires carrying DM currents drive CM current
mechanism is shown at Figure 7-11 and 7-12. Figure 7-11: Two Unbalance Wires carrying DM currents drive CM current
Figure 7-12: A proposed noise reduction method by twisting tracks
Figure 7-12: A proposed noise reduction method by twisting tracks
We use current probe to measure the DM current. The voltage output of the current orobe is connected to a spectrum analyzer (ADVANTEST R3131) or to an oscilloscope -YOKOGAWA DL1540). To monitor weak signals, a pre-amplifier (KALMUS 737LC) is used to amplify weak voltage signal. This amplifier owns a constant gain of 44dB up 0 1GHz. Racaiice the calibration of current nrohe is normally done with a 50O terminator. if the We arrange all the | EUT ( Equipment Under Test) on a table with a height of 80-90cm
We use current probe to measure the DM current. The voltage output of the current orobe is connected to a spectrum analyzer (ADVANTEST R3131) or to an oscilloscope -YOKOGAWA DL1540). To monitor weak signals, a pre-amplifier (KALMUS 737LC) is used to amplify weak voltage signal. This amplifier owns a constant gain of 44dB up 0 1GHz. Racaiice the calibration of current nrohe is normally done with a 50O terminator. if the We arrange all the | EUT ( Equipment Under Test) on a table with a height of 80-90cm
Figure 8-5: Diagram of CISPR-25 radiated emission module testing setup
Figure 8-5: Diagram of CISPR-25 radiated emission module testing setup
Figure 8-6: Motor and SCU in different wiring configurations We also show these configurations via Figure 8-6. Here, the status of S;, S; and S3 corresponds to three connections in Table 8-1.
Figure 8-6: Motor and SCU in different wiring configurations We also show these configurations via Figure 8-6. Here, the status of S;, S; and S3 corresponds to three connections in Table 8-1.
Table 8-2: Measurement configurations
Table 8-2: Measurement configurations
The first comparison is done with different polarity. In the setup of the measurement, the antenna is placed perpendicular to the cable of the motor, which is the worst situation. Two measurements are done with vertical polarity and horizontal polarity. We draw the results into Figure 8-7. Figure 8-7: Comparison between horizontal and vertical polarization measurement results when motor running in PWM and wiring configuration is setup5
The first comparison is done with different polarity. In the setup of the measurement, the antenna is placed perpendicular to the cable of the motor, which is the worst situation. Two measurements are done with vertical polarity and horizontal polarity. We draw the results into Figure 8-7. Figure 8-7: Comparison between horizontal and vertical polarization measurement results when motor running in PWM and wiring configuration is setup5
Figure 8-8: Comparison of measurement results when motor is running in always-on mode and in PWM mode, while wiring configuration is setup4
Figure 8-8: Comparison of measurement results when motor is running in always-on mode and in PWM mode, while wiring configuration is setup4
8.4.3 Comparison between different wiring configurations Figure 8-9: Comparison of radiated emission with different wiring configurations when motor is running in PWM mode In this comparison we can see that the difference between fields below 2MHz are negligible. One lump is found in setup] around 9MHz, and setup2 comes out two lumps around 5MHz and 9MHz, while setup4 is accompanied by two lumps in 350kHz and Q9MHz.
8.4.3 Comparison between different wiring configurations Figure 8-9: Comparison of radiated emission with different wiring configurations when motor is running in PWM mode In this comparison we can see that the difference between fields below 2MHz are negligible. One lump is found in setup] around 9MHz, and setup2 comes out two lumps around 5MHz and 9MHz, while setup4 is accompanied by two lumps in 350kHz and Q9MHz.
800kHz is excited by PWM signal and it lifts the level about 10dB higher.
800kHz is excited by PWM signal and it lifts the level about 10dB higher.
Figure 8-11: Comparison between setup1 and setup5 To see the improvement of the third wire, the following comparison, were made.
Figure 8-11: Comparison between setup1 and setup5 To see the improvement of the third wire, the following comparison, were made.
Figure 8-13: Comparison between setup4 and setup8
Figure 8-13: Comparison between setup4 and setup8
From Figure 8-15, we find distinguishing peaks are left with a low level if the wiring configuration is setup2. The reason is part DM current is converted into CM current flowing through metal plane. Therefore, the radiation produced by converted CM current is not shielded. This experiment confirms the mechanism CM current which is converted from DM current produce EMI. 8.5 Synthesis
From Figure 8-15, we find distinguishing peaks are left with a low level if the wiring configuration is setup2. The reason is part DM current is converted into CM current flowing through metal plane. Therefore, the radiation produced by converted CM current is not shielded. This experiment confirms the mechanism CM current which is converted from DM current produce EMI. 8.5 Synthesis
identify the catalogue of noise as Figure 8-16 shows. Figure 8-16: Routine to identify characteristic of noise
identify the catalogue of noise as Figure 8-16 shows. Figure 8-16: Routine to identify characteristic of noise
Related topics:
Engineering
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved