Evaluation Method and Modeling Analysis of the Common Mode Noise Suppression Capability of Full-Bridge Transformers

Abstract

The effective capacitance of the common mode port serves as a critical metric for assessing the common mode noise suppression capability of transformers in power converters. Conventionally, the evaluation of transformers in single-ended topologies, such as flyback converters, using a network analyzer necessitates a reference static point and a dynamic point at the transformer port. However, a full-bridge transformer without a center tap lacks a reference static point in both the primary and secondary stages. Consequently, this paper proposes an innovative measurement technique to evaluate the common mode noise suppression capability of full-bridge transformers. This method accounts for the intrinsic parameters of the transformer and refines the high-frequency equivalent circuit model for accurate measurement. Ultimately, the validity of the proposed model is confirmed through experiments conducted on a CLLC converter prototype, offering the industry a straightforward and efficient approach to assessing and testing the common mode noise suppression performance of transformers without a center tap.

1. Introduction

In recent years, CLLC converters, LLC converters, DAB converters, and other isolated DC/DC converters have found widespread application in the field of renewable energy. As the switching frequency of isolated converters continues to improve, the power density of magnetic components has also increased. Electromagnetic interference (EMI) remains one of the primary concerns in the design of isolated converters [1,2,3]. The transformer, a critical magnetic component in isolated DC/DC converters, plays a significant role in the conduction of common mode noise in these power converters.

For the current DC/DC converter topology, the differential mode conduction emission mechanism is clear and easy to handle, while the common mode conduction emission pathway is more complex, which is the focus [4,5,6] of analysis. According to the analysis of the noise source, coupling path, and sensitive equipment, the switch tube of the bridge arm will be regarded as the main noise source for an isolated converter. The common mode conduction path is mainly divided into two parts: the first path is through the isolated converter bridge arm switch tube (or semiconductor diode) equipotential moving point conduction to its radiator and then through the ground capacitance into the earth; the second path is through the isolation-type converter transformer in the primary winding coupling to the secondary winding and then through the ground capacitance into the earth. For the first noise flow path, the existing engineering technology connects the static point of the converter’s primary side and the secondary side’s static point respectively through the heat sink of the switch tube (or semiconductor diode) for bypass.

For the second common mode conduction path, the potential distribution on the winding is closely related to the magnitude of the common mode noise. Therefore, it is also difficult to control the common mode noise of isolated converters at present. Based on the measurement method to evaluate the transformer common mode noise suppression ability of the scheme, it is also very rare.

The literature [7] proposes using a network analyzer to measure the S21 parameter and evaluate the common mode noise suppression ability of the transformer. The transformer is regarded as a filter to measure the insertion loss. The smaller the S21 parameter, the better the CM noise suppression capability of the transformer. In the design of filters, S21 is typically used to represent insertion loss (IL). IL reflects the degree of loss and attenuation of signal power before and after the use of the filter. The greater the insertion loss, the more attenuation there is, and the better the performance of the filter. For the full-bridge two-terminal topology, because each port of the transformer is directly connected to the semiconductor device, each port has a drastic potential change; that is, the four ports of the transformer are potential moving points. Because the port has no static point of potential stability, the above evaluation method cannot find a suitable circuit connection mode to ensure that the transformer measurement potential distribution and the actual working situation are consistent; existing literature research shows that for any double-winding transformer, one can use the independent six capacitance model to describe the transformer [8,9,10,11,12,13] with a center tap. However, the relevant capacitance parameters are generally measured using finite element simulation or two-port equipment. For a full-bridge type centerless tap transformer, the absence of a center tap increases the difficulty of obtaining parameters. Secondly, the accuracy of the electromagnetic parameters of the finite element simulation of the planar PCB transformer is better, but for Litz wire winding or other winding structures, winding, and material multiple factor mixing, it is difficult to extract parameters through finite element software. Therefore, this kind of modeling method also has certain limitations. In addition to the use of the simulation method to propose parameters, the use of the frequency domain method and time domain method to analyze the common mode noise is also one [14,15,16,17,18] of the common means. The modeling of this kind of method is complicated, and the test results should be improved by trial and error.

If the transformer loss characteristics and EMI characteristics can be comprehensively considered in the transformer design stage, the cost and time involved in later transformer optimization can be saved to a great extent, and through the measurement of transformer loss and the definition of the common mode noise mechanism at high transformer frequency, the development of high-power density of the switching power supply has very important research significance.

In this paper, a new measurement method of transformer common mode noise is proposed, which considers the situation in which the transformer has no center tap in the full-bridge double-ended topology and expands the application range of the test object. It has certain engineering application value.

2. Traditional Transformer Common Mode Noise Suppression Capability Evaluation Method

2.1. Transformer Common Mode Noise Transmission Mechanism

In the actual work, there is a gradient potential distribution in the primary and secondary winding of the transformer, so there are induced charges between the turns of the primary winding, the turns of the secondary winding, and the layers between the primary and secondary winding. The main transmission mechanism of the common mode noise current in the transformer is as follows: because the port of the primary side winding and the port of the secondary side have dynamic potential distribution, the electric field is coupled between the primary and secondary side windings so that the charge induced by the secondary side can form a displacement current through the distribution capacitance to the ground.

It can be seen that the capacitance used to measure the common mode current path of the transformer is a two-port transfer impedance parameter, which represents the transformer’s ability to suppress common mode noise in practical work.

2.2. The Definition of the Effective Capacitance of the Common Mode Noise Port

For static structural capacitors, since there is no potential distribution, physical structural parameters such as area and equivalent spacing between winding layers are determined. The calculation of the structural capacitance of the high-frequency transformer can generally use the approximate parallel plate capacitance calculation formula. Under the premise of the transformer winding tightly wound, the structural capacitance can be calculated using Formula (1):

$$C_0={\epsilon }_0{\epsilon }_r{\displaystyle \frac{wl}{d}}$$

(1)

where ${\epsilon }_0$ is the dielectric constant of air; ${\epsilon }_{\mathrm{r}}$ is the relative dielectric constant of the material; w is the width of the winding; l is the perimeter of the transformer windings; and d is the equivalent spacing of the windings.

It should be noted here that in the design process of some transformers, the center column may be round, so this type a cylindrical capacitor is used to equivalent the winding layer of the primary secondary side.

As shown in Figure 1, the capacitance between the primary and secondary sides of the transformer (without a shielding layer) can be measured using a measuring device such as an LCR meter and impedance analyzer. However, the two measurement connection methods shown in Figure 2 make it easy to determine that the potential above the winding is the same, whether it is the primary or secondary side winding. Therefore, only with the static structural capacitance does this measurement obtains Cps. The static structural capacitance depends on physical structural parameters such as the opposite area between the windings, insulation spacing, etc. However, when the actual circuit works, the potential distribution above the windings is not static and unchanged, so only the static structural capacitance can not effectively characterize the transformer’s common mode noise conduction capacity.

As shown in Figure 3, the displacement current of the net charge induced by the electric-field coupling induced by the sub-side winding of the transformer through the distribution capacitance to the ground is actually the common mode noise current flowing through the transformer.

In order to facilitate the analysis of the effective capacitance of the common mode noise port, similar to the fundamental wave analysis of the circuit, we need to assume certain conditions:

• The leakage inductance of the primary side and the secondary side of the transformer is very small.
• Assume that the port voltage of the transformer winding is uniformly and linearly distributed along each turn winding.
• The primary side winding and secondary side of the transformer are single-layer winding, and the turn ratio is n:1.

Based on the above prerequisites, assuming that the width of the winding is w, the original side potential distribution is U_S, and the corresponding secondary side potential is U_p, then the potential difference between the primary and secondary sides can be expressed as follows:

$$\Delta U=(U_P-U_S){\displaystyle \frac{x}{w}}$$

(2)

Therefore, the amount of charge induced by the secondary side windings of the transformer is as follows:

$$Q={\displaystyle {\int }_0^w{C}_0\cdot \Delta U}\cdot {\displaystyle \frac{dx}{w}}$$

(3)

In Formula (3), C₀ is the transformer structure capacitance, which can be measured using the impedance analyzer. The calculated amount of charge is reduced to the port voltage U_P on the original side, and the effective capacitance of the common mode port expressed by the lumped parameter is:

$$C_Q={\displaystyle \frac{Q}{U_P}}={\displaystyle \frac{\frac{1}{2}(U_P-U_S)C_0}{U_P}}$$

(4)

When the number of turns of the primary side winding is much higher than that of the secondary side, it can be known that:

$$C_Q={\displaystyle \frac{C_0}{2}}$$

(5)

From the above analysis, it can be seen that there is a certain linear proportional relationship between the static structural capacitance and the effective common mode port capacitance represented by lumped parameters, but it is not exactly equal. The actual transformer should analyze and calculate the relationship between the two according to the specific structural theory, but because the actual structure of the transformer, in addition to the winding structure as an influencing factor, also includes different winding, different insulation materials, and other influencing factors, it is difficult to directly measure the effective capacitance of the common mode port through simulation modeling analysis or efficient means.

For the transformer in the power converter, because it is in the transmission path of common mode noise, in addition to its own voltage conversion and isolation characteristics, we can actually regard it as a common mode filter, which is used to suppress the conduction of common mode noise.

According to Figure 4, the transformer is regarded as a common mode filter, and the transformer insertion loss measurement curve is measured directly with the EMI receiver. The insertion loss is defined as shown in Formula (6):

$$I_L=S_{21}=20\lg \left|{\displaystyle \frac{U_2}{U_1}}\right|(\mathrm{dB})$$

(6)

where U₁ represents the magnitude of the voltage over R₂ without C_Q, and U₂ represents the voltage over R₂ with C. R₁ and R₂ are the source impedance of the receiver 50 Ω and the RF input impedance of the receiver 50 Ω, respectively.

The calculation Formula (9) of C_Q can be obtained through the simultaneous solution of Formulas (7) and (8). Thus, the effective capacitance of the common mode port can be calculated and evaluated. In this way, in practical engineering applications, we can measure the insertion loss of the transformer, calculate the effective capacitance of the transformer common mode port, and reflect the design of the transformer common mode noise suppression ability through its size.

$$U_1={\displaystyle \frac{R_2}{R_1+R_2}}{v}_{ac}={\displaystyle \frac{1}{2}}{v}_{ac}$$

(7)

$$U_2={\displaystyle \frac{R_2}{\frac{1}{jw{C}_Q}+R_1+R_2}}{v}_{ac}={\displaystyle \frac{50}{\frac{1}{jw{C}_Q}+100}}{v}_{ac}$$

(8)

$$C_Q={\displaystyle \frac{1}{2\pi f\sqrt{{10}^{\frac{40-I_L}{10}}-{10}^4}}}$$

(9)

According to the above calculation process, we only need to measure the corresponding transformer insertion loss through the network analyzer—we can obtain the C_Q change trend under the corresponding frequency—so as to facilitate further evaluation of the common mode noise suppression ability of the transformer. However, through analysis, it is not difficult to find that this method has certain limitations:

(1)

The transformer measured as shown in Figure 4 needs to have clear moving points and static points. The analysis of potential static points in

Table 1. Common isolated topology potential critical point analysis.

Topology Types	Whether the Static Potential Point Exists or Not
	Primary Side Winding	Secondary Side Winding
(a) Flyback	√	√
(b) Forward Conversion Circuit	√	√
(c) Push-pull	√	√
(d) Half bridge	√	√
(e) Front stage full bridge + full wave rectification	×	√
(f) Front stage full bridge + bridge type rectification	×	×
(g) Front full bridge + double current rectifier	×	×
(h) Front stage full bridge + voltage doubling rectifier	×	×
Method	Applicable topology	Precision
Literature [19]	(a), (e)	(a), (e)-General
Literature [7]	(e)	(e)-General
Literature [20]	(a)	(a)-Higher

shows that not all topologies have potential static points. Therefore, the application of this measurement method is limited.

(2)

In the above derivation process, the equivalent circuit model is too idealized. It is believed that the magnetic parameters of the transformer have no effect.

3. New Method for Evaluating the Common Mode Noise Suppression Capability of the Transformer

3.1. Construction of Static Potential Point

For the measurement method shown in Figure 4, it is generally used in single-ended isolated topology. There are clear potential moving points and potential resting points.

As shown in Figure 5, taking the CLLC converter as an example, the common mode noise transmission path inside the transformer contains four potential jump points, A, B, C, and D. Therefore, the measurement method in Figure 4 is no longer applicable. Therefore, this paper presents a method of constructing a potential static point by using an RF transformer as an auxiliary tool to measure the effective capacitance of a common mode noise port. The measurement method is shown in Figure 6.

The measurement method is shown in Figure 5. Ideally, when the RF transformer is connected in parallel to the primary winding of the transformer since no secondary winding is used, the RF transformer is equivalent to an autotransformer. According to the working principle of the autotransformer, when the voltage is applied to the two ends of the 1 pin and the 2 pin, the two ends of the 3 pin will also produce the same size and the same direction of the induced voltage. Therefore, the RF transformer is used to simulate the port potential change characteristics of the full-bridge transformer. And because the 1:1 autotransformer is in parallel with the primary and secondary side windings of the transformer respectively, the potential of the center tap is the same as that of the midpoint of the winding impedance. Therefore, point E can be regarded as the potential static point of the primary side winding; in the same way, the secondary side winding of the transformer can also be connected to the RF transformer to create a potential static point. In this way, after the construction of the potential static point, we can measure it according to the method of measuring the single-ended topological transformer.

For the RF transformer as an auxiliary tool, we also need to evaluate the rationality of its use. The RF transformer adopts the RF transformer manufactured by Coilcraft company, model PWB3010L. According to the user manual data introduction, its bandwidth size is 0.0035 MHz~125 MHz. The insertion loss of the transformer is a maximum of −3dB. Through the above parameters, it can be seen that the performance of the RF transformer is better. However, in the measurement of the new method, we only used the primary side winding of the RF transformer, so we needed to further measure and analyze the performance of the single-side winding. As shown in Figure 6a, in order to verify the performance of single-side winding as an autotransformer, the insertion loss test of the autotransformer was measured by using network division.

As can be seen in Figure 6 above, the measurement range of the network analyzer is set within the range of 0.1–30 MHz. When the interval is 100–500 kHz, the maximum attenuation of the autotransformer is −1.29 dB, and when it is within the range of 30 MHz, the maximum attenuation of the autotransformer is −4.85 dB. It can be seen that in the entire measurement range of conduction, the unilateral winding of the RF transformer as an autotransformer can maintain good performance.

3.2. New Equivalent Circuit Model

As shown in Figure 7, based on the measurement method and its equivalent circuit model illustrated in Figure 4, two magnetic components with magnetic field inductances of Lm1 = 400 μH and Lm2 = 4 μH were measured. If the calculation method as shown in Formula (9) is followed, the obtained effective common mode port capacitance should exhibit a 20 dB/10-fold frequency change trend. However, it is not difficult to see from the change trend in the figure that the change trends of the S21 parameter do not all show a 20 dB/10-fold frequency change trend. Therefore, the traditional single-capacitance equivalent circuit model (the equivalent circuit model shown in Figure 4) has certain limitations. Therefore, for the new measurement method, we also need to propose a new equivalent circuit model to further improve the calculation accuracy of the effective common mode port capacitance of the transformer.

In order to facilitate the description of the flow process of the displacement current, in Figure 8, we use the common transformer model to characterize the center-tapped transformer and the center-tapped transformer. In Figure 8, R is the equivalent resistance of the transformer windings; Lm is the transformer excitation inductance; and L_k is leakage inductance.

Whether it is a full-bridge transformer or a transformer in a single-ended topology, because the transformer is in the circulation path of noise, the displacement current formed by the charge induced by the secondary side will resonate with the magnetic parameters of the magnetic component during the conduction process. In the traditional measurement scheme shown in Figure 4, the assumption given is that the excitation inductance is large enough, thus ignoring the influence of the excitation inductance. However, in the measurement results shown in Figure 7, the variation trend of insertion loss under different excitation inductances cannot be described by the equivalent circuit of a single capacitor.

For the full-bridge transformer, the transformer with a center tap and the transformer without a center tap are shown in Figure 8. The mechanism of both measurement methods is the same. Among them, the full-bridge transformer with a center tap can be measured according to the traditional measurement method as shown in Figure 4.

From the measurement method, we can see that because the potential static point of the primary side and the secondary side of the transformer with the center tap is clear, there will be no displacement current flowing through the secondary side winding. Therefore, only the influence of primary field inductance and equivalent leakage inductance on insertion loss is needed. Similarly, for the new measurement method, we also need to consider the influence of excitation inductance and equivalent leakage induction in combination with the connection mode of the transformer model and actual measurement.

For the connection mode without a center tap, as shown in Figure 9, we added the excitation inductance and the equivalent leakage inductance into the equivalent circuit. Because the permeability of the magnetic core is not constant, and its own permeability will also change with the change in frequency, Z_LM is used to characterize the impedance characteristics of the excitation inductance in the equivalent circuit, and Z_LK is used to characterize the impedance characteristics of the equivalent leakage inductance. Generally speaking, the impedance value of the excitation inductance in the low-frequency band is larger, the influence of leakage induction can be ignored, and the proportion of leakage induction after the high-frequency band is larger, so the total impedance of a test is the sum of the equivalent leakage impedance and excitation impedance, and the effective capacitance of the common mode port is carried out in different frequency ranges for the RF transformer so that L₁ = L₂ and mutual induction M = 1.

The improved high-frequency circuit model mainly considers the influence of the RF transformer, excitation impedance, and equivalent leakage impedance on measurement.

In order to facilitate the calculation, it can be made:

$$Z_{n1}={\displaystyle \frac{1}{j\cdot \omega \cdot C_q}}+R_2+{\mathrm{Z}}_{\mathrm{RE}}$$

(10)

$$Z_{n2}={\displaystyle \frac{1}{\frac{1}{2\cdot \mathrm{j}\cdot \omega \cdot L_2+{\mathrm{Z}}_{\mathrm{LM}}+{\mathrm{Z}}_{\mathrm{LK}}}+\frac{1}{2\cdot \mathrm{j}\cdot \omega \cdot L_1}}}$$

(11)

According to Formula (6), we can obtain the improved insertion loss Expression (12) by substituting Formulas (10) and (11).

$$I{L}_{\mathrm{Lm}2}=20\lg \left(\left|{\displaystyle \frac{100}{50+\frac{1}{\frac{1}{{\mathrm{Z}}_{\mathrm{n}1}}+\frac{1}{{\mathrm{Z}}_{\mathrm{n}2}-\mathrm{j}\mathsf{\omega }\mathrm{M}}}}}\cdot \right.\right.\left.\left.{\displaystyle \frac{1}{\frac{1}{Z_{n1}}+\frac{1}{Z_{n2}-\mathrm{j}\omega M}}}\cdot {\displaystyle \frac{1}{Z_{n1}}}\right|\right)$$

(12)

4. Experimental Verification

In order to further verify the accuracy of the calculation model and the high-frequency circuit model, a measurement platform was set up in this study, and the measured S₂₁ parameters of the transformer were compared with the calculation model. In this paper, two design schemes of magnetic parts are shown. Scheme A uses a 12 mm wide copper foil up-and-down parallel winding system. Scheme B adopts a 20 mm wide copper foil, and the winding method adopts an ordinary structure arrangement (P-S). The transformer winding is made of copper foil winding, the number of turns of copper foil on the primary side is 24, the number of turns of copper foil on the secondary side is 15, and the magnetic core model is DMR95. The test connection mode of magnetic parts is shown in Figure 5. The real copper foil transformer is shown in Figure 10.

The improved measurement method needs to pay attention to the impedance of the RF transformer as it is much larger than the part to be measured. At present, in the mainstream charging pile module on the market, the transformer inductance in the DC/DC module is generally about 100 μH~200 μH. The single winding inductance of the wire art RF transformer used in this study is 780 μH. The excitation inductance of the magnetic component is about 140 μH, enough to meet the measurement needs. If the impedance of the RF transformer is too small, part of the noise current will not flow through R₂, resulting in measurement errors.

In the calculation, it is necessary to first measure Z_LM, and Z_Lk, in the calculation model, where the impedance parameters measured by the open circuit of the secondary side include the excitation inductance and the equivalent leakage inductance of the transformer. The impedance parameters measured by the short circuit of the secondary side are the equivalent leakage inductance to the original side. The corresponding impedance curves are solved respectively through the data of the impedance analyzer.

According to the comparison of measurement results in Figure 11, in the low-frequency band of 0.1–1 MHz, the three curve trends of scheme A and scheme B are basically consistent. In the low-frequency band, the insertion loss difference of scheme A is 4 dB at most. In the range of 0.7 MHz to 4 MHz, the trend is basically the same. In this frequency band, the overall upward trend of 20 dB/ 10 frequency is maintained. This shows that at this stage, the main common mode noise port effective capacitance plays a leading role, that is, C_Q plays a major role. Therefore, its size can be effectively evaluated through this upward trend. For the band of 10–30 MHz, no matter scheme A or scheme B, there is a certain error after the resonance point of the three curves.

As can be seen in Figure 6 above, the measurement range of the network analyzer is set between 0.1 MHz and 30 MHz. When the interval is between 100 kHz and 500 kHz, the maximum attenuation of the autotransformer is −1.29 dB, and within the 30 MHz range, the maximum attenuation of the autotransformer is −4.85 dB. At this time, it is difficult to ensure that the “0” position is the potential center of the transformer. This is also a limitation of the test. It is necessary to ensure that the impedance of the autotransformer is large enough. Therefore, when dealing with different types of transformers, different RF transformers need to be selected to meet the measurement requirements. This is also the direction for further expansion of the thesis in the future.

For the two transformer winding methods used, it is not difficult to find that the PS winding method of scheme B makes the positive area of the copper foil winding larger. The upper and lower parallel winding method of scheme A makes the positive area of copper foil winding relatively small. As shown in Figure 12, the insertion loss of scheme B is about 20 dB larger than that of scheme A in the linear segment region.

According to CISPR-16-1 [21], within the frequency range of 150 kHz to 30 MHz for conducted electromagnetic interference noise testing (Figure 13), the values of the internal components of the LISN are specified as C₁ = C₂ = 1 μF, C₃ = C₄ = 0.1 μF, L₁ = L₂ = 50 μH, and R₁ = R₂ = 1 kΩ. Structurally, the LISN can be regarded as a low-pass filter. For the input power supply, inductors L₁ and L₂ present low impedance, allowing the input power supply to supply power to the device under test (DUT) normally. For the noise in the conducted interference frequency band, inductors L₁ and L₂ present high impedance, while capacitors C₁ to C₄ present low impedance. The interference noise current on the input side of the power supply is bypassed by capacitors C₁ and C₂, and the interference noise current iL and iN generated by the EUT flows through C₃, C₄, R₁, R₂, and the EMI receiver.

In order to further compare the rationality of the test data, a prototype platform with 1 kW power as shown in Figure 14 was built in this study for noise testing, with an operating frequency of 200 kHz. The prototype topology adopts CLLC topology. The parameters of the transformer and the resonant cavity of the CLLC are shown in

Table 2. Prototype parameters.

Parameter Type	Parameter Size
Input voltage	Vin 400 V
Input current	Iin = 205 A
Output voltage	Vout = 328.8 V
Output current	Iout = 2.97 A
Efficiency	97.71%
Resonant inductance	Lrp = 26.8 μH Lrs = 18.24 μH
Leakage inductance	Lkp = 3.7 μH Lks = 2.6 μH
Excitation Inductance	Lm = 142.3 μH
SiC Mosfet (S1–S8)	C3M0030090K

.

In order to avoid the interference of common mode noise through other paths, in the measurement process of the prototype, for the potential moving point such as the bridge arm switch tube (or semiconductor diode) of the isolated converter to its heat sink and then flowing into the ground through the ground capacitance, the primary heat sink and secondary heat sink are connected to the potential static point on the primary side and the potential static point on the secondary side, respectively, to suppress the noise.

As shown in Figure 15., we compared the results according to the noise. When the magnetic parts in scheme B are running, the overall common mode noise of the prototype is higher than that in scheme A, which is basically consistent with the trend of insertion loss. The trend of individual points in the low-frequency band is somewhat inconsistent with the model results, but the overall noise level is positively correlated with the insertion loss measurement results. The accuracy of the model is illustrated on the side of the noise measurement results page. This measurement method does not have too complicated a calculation formula, and the measurement method is relatively simple. It is helpful to judge the common mode noise suppression ability of transformers quickly and simply in engineering applications.

5. Conclusions

This paper analyzes the mechanism of transformer suppression of common mode noise and draws the following conclusions through the above research analysis and comparison:

1. The traditional insertion loss model used to evaluate the common mode noise capacity of transformers cannot be applied to transformers with no definite static points.

2. In view of the problem that the traditional measurement scheme cannot measure the full-bridge transformer, the new measurement method and the corresponding improved equivalent circuit can better characterize the insertion loss of the full-bridge transformer.

3. The new measurement model can more accurately characterize the transformer insertion loss at high frequency. At the same time, the improved circuit model makes clear the physical meaning of the transformer insertion loss curve, that is, in the low-frequency band (linear segment), it is determined by the inductive parameters of the port winding and the dynamic capacitance parameters of the transformer. In the high-frequency band, the effective capacitance of the common mode port and the leakage inductance of the winding will resonate.

4. The new measurement method was compared with actual measurements and calculations. In a certain frequency band range, the calculation results of the model and the measured results have high consistency, and the changing trend of both is the same in the entire frequency band range. Thus, the reliability and accuracy of the new measurement method have been verified, which provides a reference for EMI optimization design and the evaluation of full-bridge non-center tap transformers.