The CM choke, on the other hand, requires less inductor core and installation area since it uses the

resultant phase currents, which are much less than individual currents. The inductance of the CM choke is
decided based on the attenuation requirement and on the range of switching frequency (fsw) components to
be attenuated. Keeping the resonance frequency (fres) less than fsw will attenuate all the CM current
components but it demands a high value of inductance, whereas fres >> fsw results to non-elimination CM
current [3].

beside the several mitigation approaches, to address EMI problems in pulse width modulation (PWM)
inverter based electric drives.

A number of papers have been published on the use of CM-choke and passive filters along between
inverters and the motors.

At present, adding a filter is

a common method, which can be achieved by connecting a filter between the inverter and the
motor.

the inverter output

CM- chokes and passive filters has been applied to

applied at the output of inverter.

Many different types of approaches have been suggested in the literature to address this problem.

CM EMI

the neutral voltage at the motor neutral terminal (dvN/dt).

The dv/dt characteristics of the power semiconductor devices and the inherent electrostatic parasitic
capacitances between the rotor, stator windings, and the frame of the electric machines can lead to [2],
[15], [16]:

1. Ground current,
 2. Shaft voltage and resultant bearing currents,
3. Conductive and radiated electromagnetic interference (EMI) and
4. Over-voltage and resulting non-uniform distribution of voltage across motor windings.

Despite a large reduction of CM current over a wide frequency range, the analysis reveals that there exist
two overshoots, one at lower frequency and the other at higher frequency ranges, as a result of low and
high frequency resonances respectively. Moreover, the study has indicated that the higher frequency
resonance is influenced by inductance of the CM filter or CM choke while the low frequency resonance is
influenced by the neutral filter in addition to cable and motor impedances.

Figure 4 shows the CM current simulation results for a conventional three-wire topology with CM choke,
a four-wire topology with a previously proposed (reference) topology, and a four-wire system with a
newly proposed topology.

As demonstrated in figure 4 compared to the well-filtered three-wire topology, a significant CM current

reduction has achieved over the whole frequency range with the newly proposed four-wire neutral filter
topology. Furthermore, when compared with the topology introduced in [9], with the newly proposed
topology a fairly significant CM current reduction in the frequency range of less than 50 kHz was
realized, where the effectiveness of [9] was limited.

As it can be observed from figure 5, at a frequency range less than 50 kHz, since the CM current with the
newly proposed topology is lower than the filtered conventional three-wire topology and topology
introduced in [9], the CM current difference (deviation) is positive. The CM current deviation in the latter
case (the case of figure 5(b)) is even more than the case of figure 5 (a). This is mainly due to the filter
topology proposed in [9] originally had a negative influence in attenuation of CM current in the frequency
range of less than 50 kHz.

Furthermore, another neutral filter topology for the 2.2 kW motor is designed based on the design
principle presented in [9]. is also re-designed and simulated using the methodologies outlined in the work
for comparison purposes. As a result, the reference topology's parameters were set to C nNatt1 = 10 nF,
CnNatt2 = 90 nF, and RnNatt = 200 Ω.

at a frequency range less than 50 kHz, since the CM current with the newly proposed topology is lower
than the filtered conventional three-wire topology and topology introduced in [9], the CM current
difference (deviation) is positive. The CM current deviation in the latter case (the case of figure 5(b)) is
even more than the case of figure 5 (a). This is mainly due to the filter topology proposed in [9] originally
had a negative influence in attenuation of CM current in the frequency range of less than 50 kHz.

In order to validate the effectiveness of the proposed filter topology, two self-developed filter circuits; one
designed according to [9] and the other designed based on the topology proposed in the current paper were
implemented and tested on a 2.2 kW three-phase permanent magnet servo motor drive system
manufactured by MOOG GmbH. The dc-link voltage (Vdc) is 200 V and the switching frequency (f sw) of
the inverter is 10 kHz.

motor is in standstill is again considered,

The experiment

The CM current measurement is carriedout s of the developed filters were validated experimentally on a
600-V, 2.2-kW permanent magnet servo motor drive system with a switching frequency of 5-kHz and a
cable length of 3-meter. For CM current measurement, a commercial current sensor (A6303 from
Tektronix) is used. Figure XXX shows the photograph of the experimental set up.

In this experiment the CM current of the filtered topologies were measured when the motor is stand still
and also when rotating with constant speed. Both theoretically and practical measurement revealed that the
maximum CM current is measured when the motor is at standstill. Figure yyy shows the practical
measured CM current at stand still.

Figure xxx indicates that the simulated and experimental results of

The tested prototype revealed that the simulation and experimental results fit together very well.

The experimental result revealed the theorical simulation result

From the experimental result it can be observed that, the proposed topology can effectively attenuate CM
current in a four-wire drive system.

determined by using an Agilent oscilloscope and three Agilent high-voltage probes

tdifference of )is simulated

simulation of the four-wire electric drives topology with and without neutral filter.

a 1uF Based on the in this research

The current study proposes an optimal neutral filter topology for reducing the negative effects of low
frequency resonance in the four-wire drive.

2 shows the CM and DM impedance relation between the measured and simulation a of a motor under
test.

As

Accordingly, Figure 2 shows comparison between

.
 Comparison of experimental and simulation of CM and DM impedance of the motor under test

the high-frequency model of the motor has been developed based on the motor CM and DM
characteristics impedances.

In the current research the following procedures have been followed. Firstly, the high-frequency (HF)
model of the motor has been developed based on behavior modeling methods presented in [14].

the
neutral point of the motor is connected to the middle
point of the dc-link capacitor bank
the 4wDC topology is implemented
using a conventional three-phase inverter plus adding a capac-
itive divider in the dc bus

According to [15] a four-wire electric drive architecture is preferred for fault tolerant operation and
position sensor-free control, according to literature. Furthermore, the four-wire drive system has been
shown to be suitable for the reduction of CM currents [8], [9].

As it can be observed from figure 1 the electric drive system model consists of three parts: inverter, power
cable and electric motor.

As indicated in [7] and [9], both the CM choke and CM current filter performed satisfactorily in
eliminating CM current in both three-wire and four-wire electric drives systems at higher frequencies.
However, at lower frequency, particularly below 100 kHz, the performance of both CM choke and CM
filter degrades.

In order to optimize space, in this research we have used CM choke instead of CM filter to make
comparison with three-phase three-wire drive system.

As it can be observed from the

transfer function of the proposed neutral filter to inverter can be written in Laplace domain
Figure 1 shows the high-frequency model of a four-wire electric drive system with CMMF.

In the current research the following procedures have been followed. Firstly, the high-frequency (HF)
model of the motor has been developed based on behavior modeling methods presented in [14].

In this study, a XXXKw,YYYY type of motor is employed to conduct the experiment.

the motor CM and DM characteristics impedances.

Accordingly, an Agilent 4294A Precision Impedance Analyzer is utilized to measure the CM and DM
impedance of the motor. Secondly, the designed neutral filter topology is tested via simulation on PSIM
software and practical implementation.

In the current paper a XXXKw ,YYYY type of motor is used for experiment. the high frequency model of
the cable and the motor is carryout through behavior modeling methods presented in [14].

In this study, a XXXKw,YYYY type of motor is employed to conduct the experiment. The motor high-
frequency model is implemented using the behavior modeling methods described in [8]. The ABB TCL
555 impedance analyzer is used to measure the amplitude and phase of the motor's CM and DM
impedance from 40Hz to 10MHz.

1.size values of RLC elements based on the desired cut-off frequency

2. adjust the values of parameters, specially, values of L and R to avoid over-voltage and damping
effects.

resonance frequency

From the frequency responses of the two impedances, a high-pass filter can be obtained
 Resistance is what provides the damping. Reduce the resistance will reduce the damping
and increase the Q.

Lower the effective resistance in the circuit. If it is a parallel resonant circuit, increase the
resistance value. If it is a series resistance circuit, lower the resistance value.

To

To excite the zero sequence, the neutral point of the motor

must be accessible. This is not usually the case in commercial
PMSM. However, the modification of the motor to get access to
the neutral point is minimal.
the neutral point
N
must be connected to the
dc link point
O
.
For that reason, the neutral point
is connected to the dc link through a filter, as shown in Fig. 1.

The resultant will be high pass filter in which the impedance is higher at lower frequency and low in
lower frequency range.

The cut-off frequency of the filter is exactly equal to the resonance frequency of the RLC ckt.
 Firstly, the transfer function of the neutral topology circuit is computed shown in equation (1)

As it can be seen from figure 2, the proposed topology has two parallel circuits,

Therefore, the objective of this research is to propose an optimized neutral filter topology that can
eliminate CM current particularly at lower frequency in four-wire electric drive system as a continuation
of the work presented in [9].

is used for both simulation and experiment.

In the current study, a new modified neutral filter topology along with the conventional CM choke is
proposed to improve the CM current reduction performance of the four-wire drive system at lower
frequencies, particularly less than 50 kHz.

The performance of the proposed neutral topology was tested on PSIM, power electronic simulation
software and practical implementation. A standard 2.2 kW permanent magnet servo motor available in the
laboratory is used for both simulation and experiment. The simulation and experimental results show that
a reasonable reduction of CM current was achieved at lower frequencies, especially below 50 kHz, where
the CM current is substantial in the previous studies.
 According to [7] a transfer function approach CMMF design method has been presented by
introducing three design criteria. These are resonance and reflection criterion, rise time (du/dt)
Criterion, and Efficiency Criterion. Besides the bulky size of the filter, if the design principle of [16] is
strictly followed, a Satisfactory CM current elimination can be obtained in a three-wire electric drive
topology.

In Fig. 1, the inverter is modeled with ideal switches, while the cable and the motor are modeled with a
HF equivalent circuit according to [14] and [9].

Later, in [6,] a four-wire system with an improved neutral topology was introduced to improve the CM
current reduction capabilities of the four-wire drive system in a lower frequency range. Though, the new
topology, has improved the CM current elimination in higher frequency range, still the CM current
increases slightly under 100kHz

Later, in [9], a novel four-wire system with an optimized neutral topology has been introduced to improve
the CM current elimination capability of the 4-wire drive system in lower frequency ranges. Through the
analysis of the transfer function between neutral and inverter-output voltage, the proposed neutral-filter is
proved to be effective based on design criteria such as the projected start frequency of CM current
reduction, power loss, neutral current, overshoot of resonance, and available space for the topology.
In [8], it is found that a four-wire inverter-fed electric drive system can provide a far higher CM current
reduction than a well-filtered three-wire electric drive system in a wide frequency range. However, the
proposed four-wire topology has a negative effect on CM current reduction at lower frequencies less than
100 kHz.

Again in [9], a novel four-wire system with an optimized neutral topology has been introduced for better
elimination of CM current in lower frequency ranges. Through the analysis of the transfer function
between neutral and inverter-output voltage, the proposed neutral-filter is proved to be effective based on
design criteria such as the projected start frequency of CM current reduction, power loss, neutral current,
overshoot of resonance, and available space for the topology.

due to the complimentary transistors, ACC cannot be used for a higher voltage application.

are used for limited voltage ratings.

rating limitations,

Since the rating of are limited, the active EMI filter has limitations.

rated at 600 V or higher are presently not As a result of rating limitation of Due to

However, the performance of ACC is limited by many factors, such as the accurate measurement of
system CM signals and impedance, and the reliability of the compensation circuit

A hybrid EMI filter technique can

to improve the overall attenuation and bandwidth.

Because of the low price and designed easily, passive filters are currently more popular
Practice

1
CN > (1)
4 π f start2 ω LM
1 2
DM
 1
CN > (2)
4 π f start2 ω LM
2 2
DM

Since EMI filter design is not an exact science, and it involves a cut-and-trial process t o select the exact
filter components, about four criteria were proposed. These are neutral current criterion, filtering start
frequency criterion, resonance and reflection criterion, and du/dt criterion.

A. Neutral current criterion

The objective of this criteria is to maintain the neutral current minimum. Thus, to keep the neutral current
minimum, the neutral impedance (ZN) has to be greater than 1/sCN [10]. The optimum value of CN can be
obtained through parameter sweep simulation. Thus, the minimum limits of C N_1, CN_2, and RN_2 can be
obtained by applying neutral current criteria on equations (8) and (9).

1
RN > (3)
2
sCN

CN CN
(4)
1 2
<C N
C N +C N 1 2

B. Filtering start frequency criterion

As presented in [9], [10] of the two resonances present in CM filters, the high-frequency resonance is
dominated by the inverter-output filter while the low-frequency resonance is influenced by Z N. Hence to
start the filtering from a lower frequency range, the high-frequency resonance (fres2) should be less than,
fstart, as indicated in (10).

1
f res2= > f start
CN CN
2 π LM
√ DM
( C N +C N
1
1 2

2
) (5)

Were, LM_DM – denotes inductance of the machine.

Thus, the minimum limits of CN_1, CN_2 is given in (11).

CN CN 1 2 1
> (6)
4 π f start2 ω LM
2
CN + CN
1 2 DM

C. Resonance and reflection criterion

To avoid over damping of the step response of Z N_1 (s), the reflection coefficient (ζ) of the transfer
function has to be greater than or equal to 1 [7]. Equation (4) can be represented in a typical damped
second-order system as shown in (12).

1 s
Z N ( s )=
1
CN 1
( s 2+
1
RN CN
1
s+
1
LN C N
1 1 1
) (7)

Now, apply reflection criterion proposed in [7], the maximum limit of RN_1 shall be obtained from (15).

CN
R N ≤2
1
√ LN 1
1
(8)

D. du/dt criterion
The aim of this criterion is to keep the du/dt of the neutral filter less than the maximum recommended
du/dt value. The actual du/dt value can be obtained from time constant of the filter transfer function (τ) as
indicated in (14).

0.63V p 1
τ= = =2 R N C N (9)
dv /dt ζ ω o 1 1

Were, vp denotes the peak voltage

For less CM current, the du/dt of the inverter has to be less than or equal to the recommended maximum
du/dt [7].

du/dt ≤ (du/dt) CM filter_max (10)

combining equations (14) and (15), the ratio of LN_1 to RN_1 can be obtained as expressed in (16).

0.63∗1
V
LN 3 DC
≥1
(11)
RN du
1
2( )
dt CM filter max

RN_1 should be as small as possible.

Since, the parasitic capacitance of the line is lower than the parasitic capacitance of the machine, the dv/dt
of neutral is higher than dv/dt of the line.

The actual dv/dt value of the filter transfer function can be obtained from time constant (τ) of the transfer
function as indicated in (8).
 0.63V p 1
τ= = =2 R N C N (12)
dv /dt ζ ω o 1 1

Were, vp denotes the peak voltage

For less CM current, the du/dt of the inverter has to be less than or equal to the recommended maximum
du/dt [7].

du/dt ≤ (du/dt) CM filter_max (13)

combining equations (14) and (15), the ratio of LN_1 to RN_1 can be obtained as expressed in (16).

0.63∗1
V
LN 3 DC
≥ 1
(14)
RN du
1
2 ( )
dt CM filter max

RN_1 should be as small as possible.

E. Neutral current criterion

The objective of this criteria is to maintain the neutral current minimum. Thus, to keep the neutral current
minimum, the neutral impedance (ZN) has to be greater than 1/sCN [10]. The optimum value of CN can be
obtained through parameter sweep simulation. Thus, the minimum limits of C N_1, CN_2, and RN_2 can be
obtained by applying neutral current criteria on equations (8) and (9).

1
RN > (15)
2
sCN

CN CN
(16)
1 2
<C N
C N +C N 1 2

F. Filtering start frequency criterion

As presented in [9], [10] of the two resonances present in CM filters, the high-frequency resonance is
dominated by the inverter-output filter while the low-frequency resonance is influenced by Z N. Hence to
start the filtering from a lower frequency range, the high-frequency resonance (fres2) should be less than,
fstart, as indicated in (10).

1
f res2= > f start
CN CN
√
2 π LM DM
( 1
1

C N +C N
2

2
) (17)

Were, LM_DM – denotes inductance of the machine.

 Thus, the minimum limits of CN_1, CN_2 is given in (11).

CN CN 1 2 1
> (18)
4 π f start2 ω LM
2
CN + CN
1 2 DM

G. Resonance and reflection criterion

To avoid over damping of the step response of Z N_1 (s), the reflection coefficient (ζ) of the transfer
function has to be greater than or equal to 1 [7]. Equation (4) can be represented in a typical damped
second-order system as shown in (12).

1 s
Z N ( s )=
1
CN 1
( s 2+
1
RN CN
1
s+
1
LN C N
1 1 1
) (19)

Now, apply reflection criterion proposed in [7], the maximum limit of RN_1 shall be obtained from (15).

CN
R N ≤2
1
√ LN 1
1
(20)

H. du/dt criterion
The aim of this criterion is to keep the du/dt of the neutral filter less than the maximum recommended
du/dt value. The actual du/dt value can be obtained from time constant of the filter transfer function (τ) as
indicated in (14).

0.63V p 1
τ= = =2 R N C N (21)
dv /dt ζ ω o 1 1

Were, vp denotes the peak voltage

For less CM current, the du/dt of the inverter has to be less than or equal to the recommended maximum
du/dt [7].

du/dt ≤ (du/dt) CM filter_max (22)

combining equations (14) and (15), the ratio of LN_1 to RN_1 can be obtained as expressed in (16).

0.63∗1
V
LN 3 DC
≥ 1
(23)
RN du
1
2( )
dt CM filter max

RN_1 should be as small as possible.

24
, PR is the maximum allowable power loss, and fsw is the switching frequency of the inverter.

LN
R N ≥2
√ CN
(25)

0.63∗1
V
LN 3 DC
≥ (26)
RN dv
2 ( )
dt max
2 PR
CN ≤ 2 (27)
V DC f sw (1−e−6 )

Accordingly, after once determined the optimal capacitance value (C N_opt), to limit the neutral current
below 1% of the rated current, CN can be estimated by using (7).

CN ≤ CN opt

(28)

Since EMI filter design is not an exact science, and it involves a cut-and-trial process t o select the exact
filter components, about four criteria were proposed. These are neutral current criterion, filtering start
frequency criterion, resonance and reflection criterion, and du/dt criterion.

Aiming to reduce the dv/dt value by 1/10

d v tot d vN
I CM =C L +CM + I others
dt dt

Where, CL – is the parasitic capacitance between the cables and the ground

CM – is the parasitic capacitance between the winding and the motor frame

dvtot/dt – is rate of change of voltage in all the three phases (wires)

dvN/dt – is rate of change of voltage in the neutral wire

The original sources of CM current in ASD are the rise-time of the inverter output ( dvtot/dt) and the neutral
voltage at the motor neutral terminal (dvN/dt).

As it can be seen from the above equation, the peak value of the output CM current in inverter system, is
influenced by the parasitic capacitances, and rate of voltage rises (dv/dt) of the CM voltage.
The capacitance value of the proposed neutral filter circuit can be obtained by using neutral criteria and
efficiency criteria. While LN and RN can be determined by using the resonance criteria and dv/dt criteria.

The objective of neutral criterion is to maintain the neutral current minimum.

The transfer function of the proposed neutral filter can be written in Laplace domain as the sum of equations (4)
and (5), where ZN_1 (s) is the resultant of LN_1, CN_1, and RN_1 and ZN_2 (s) is the resultant of CN_2 and RN_2.

s LN R N
Z N ( s)= 2
1 1
(29)
1
s L N C N RN + s L N + R N
1 1 1 1 1

RN
Z N ( s)= 2
(30)
2
1+s R N C N 2 2

As seen in (4) and (5), at low frequency range the magnitude of Z N_2 is a very high while Z N_1 is very
low impedance. With increasing frequency, however, the impedance of both Z N_1 and ZN_2 decreases,
resulting in the system's lowest impedance. The maximum impedance is reached at resonance
frequency of ZN_1. Hence the combination of all these responses makes the network a high-pass filter.

At lower frequency range, since both the capacitors CN_1 and CN_2 are open-circuited; the proposed circuit is
approximated to a parallel RN_1 LN_1 circuit with RN_2 coupled in series. Therefore, the approximated
impedance will be given in (6).

s L N RN
ZN (s ) ≈ 1 1
+ RN ≈ RN (31)
s LN + R N
1 1
2 2

On the other hand, at higher frequency since the inductance act like open circuit, the equivalent circuit is
approximated to two cascaded parallel RC circuit. Hence the resultant impedance is approximated to (7).

1 1
ZN (s ) ≈ + (32)
s C N RN 1 1
s CN RN 2 2

At lower frequency (3) can be approximated to (4) and with increasing frequency Z N(s) can be the
impedance of both
As a further simplification, at lower-frequency (LF), the impedance of motor (Z M_DM) can be approximated
to ωLM, and it is much larger than ZM_DM. Therefore, the resonance occurs at ωLM = ZN(s).

From figure 2 a further approximation is required

[10]. The optimum value of CN can be obtained through parameter sweep simulation.

capacitor in parallel with a series resistor and inductor

Therefore, to limit the neutral current below 1% of the rated current,

The capacitance value (CN) is obtained

In both cases, the inductance value of the CM filter and CM choke affects the transfer function between
neutral topology and inverter output of the drive system, which later influences the amplitude of CM
current. A higher CM current elimination at a higher frequency range is achieved by increasing the
inductance values of CM choke, and CM filter whereas CM current elimination at a lower frequency range
can be achieved through neutral filter topology (Z N). Therefore, in the current research, an optimized
neutral filter topology is proposed to mitigate the negative effect of low-frequency CM current in a four-
wire electric drive system. The proposed circuit constitute of a series connection of parallel RC and RLC
circuits. Figure 3 shows the proposed neutral filter topology.

Figure 1: The proposed neutral filter topology

To consider the effect of filter circuit, we have to multiply by the transfer function of the filter.

As a result, the voltage measured between the neutral point and the ground (motor frame) is indicated in
(1),
neutral voltage (VN) of the four-wire drive system can be described as (1).

Moreover, since CM voltage (V CM) is a voltage measured between neutral point to ground (motor frame),
therefore, VN_G can be treated as VCM.

3 (Z C +Z N )
K 4 w (s)= DM
(3)
ZC + ZM
DM DM
+ s LFCC +3 (Z C +Z N )
DM

Where K4W(s) transfer function of a four-wire drive topology with a CM choke.

In (3) transfer function of the four-wire drive topology with system in s-domain,

Where VN(s), VCM(s), and K4W(s) indicate the four-wire drive topology's neutral voltage, CM voltage, and
transfer function in the s-domain, respectively

Equations (1) indicates a voltage measured between neutral point and ground (motor frame), the CM
voltage (VCM) andrespectively.

Since CM voltage (VCM) is a therefore, neutral to ground voltage (VN_G) can be treated as VCM.

The above equation (1) works for a balanced 3phase system without considering the effects of filter, cable
and motor impedances.

To consider these effects the above equation has to be multiplied by the transfer function of the drive
system. Accordingly, (2) shows equation of neutral voltage

onsidering the ransfer function of figure 2 with CM choke is added.

Where VN(s), VCM(s), and K4W(s) indicate the four-wire drive topology's neutral voltage, CM voltage, and
transfer function in the s-domain, respectively.
, and it should ideally be zero. However, in practical electric motors driven by pulse width modulation
(PWM) the value of CM voltage (VCM) is always non-zero. The actual value of V N_G can be obtained by
multiplying VCM by transfer function of the drive system.

K nN ( s ) V CM (s ) K nN ( s )
I n= = (V a +V b +V c ) (2)
Zn 3 Zn

In this research work the effect of a passive neutral filter combined with a CM choke on CM current
reduction in a four-wire electric drive topology is studied.

Firstly, the transfer function of

the main duty of the filters is to reduce the magnitude of the transfer function
Therefore, filtering effect of CMMF on the 4-wire topology is investigated here [8].

Firstly, a CMMF for 3-wire inverter is designed

In this paper the influence of passive neutral filter along with CM choke on CM current reduction in four-
wire electric drive topology has been analyzed.

This paper presents the design and analysis of neutral filter in four-wire electric drive topology.

The performance of the proposed strategy is analyzed in two different 4-wire topologies. In the first
topology, T

three-phase electrical drives, based on two different 4-wire topologies, are shown

The theoretical transfer function between neutral and inverter-output voltage is well illustrated in [10]
through a Laplace domain analysis. The analysis depicted that the CM current is influenced not only by
the filter parameters but also by the cable and motor impedances. Since the CM impedance of power
cables and motors is much larger than the DM impedance [14], [15] the effect of ZC_CM and ZM_CM can be
neglected. Equations (1) and (2) show the transfer function between inverter output and neutral topology
with CMMF and CM choke, respectively.

3 ( Z C +Z N )
K 4 w ( s)=
DM
(33)
CMC
ZC + Z M +s L¿F +3 ( Z C + Z N )
DM DM DM

Z C + Z M +3 ( Z C + Z N )
K 4 w ( s)= K 3 WF ( s )
DM DM DM
(34)
CMF
Z C +Z M + s L¿ F K 3 WF ( s ) +3 ( Z C + Z N )
DM DM DM

Were, K3WF(s) denotes transfer function of the CM filter. It is given by (3).

C¿ + 1
K 3 w (s)=S R¿ 2
¿¿ (35)
s L¿ C ¿ + s R¿ C ¿ +1 ¿ ¿¿ ¿