When helping customers in the electric vehicle (EV) industry with module designs, I often find that engineers tend to follow outdated "do's" and "don't do" lists in the form of EMC design rules without understanding the basics. These design rules are usually borrowed from other industries, and they are not synchronized with the latest technologies involved in the fast-paced electric vehicle industry.
In this article, taking the powertrain module as an example, I will first introduce the high-voltage EMC regulations that the powertrain module must comply with. Then, I will focus on the risks and challenges of designing such modules. The main part of this article will share the design techniques that engineers can apply to each part of the powertrain module, including grounding, front-end filter, inverter design, etc. Some examples are given to demonstrate some key design techniques.
It should be noted that the design techniques introduced in this article have the same principles as any other EMC project. Therefore, engineers in other industries can also benefit from these technologies.
The past decade has witnessed the rapid adoption of electrification in the automotive industry, with more and more hybrid and all-electric vehicles entering the market. A recent study showed that “halfway” (as done by plug-in hybrid vehicles) may not be enough to make carbon emissions comply with the new regulations, which will require more fully electric vehicles on the road [1].
The powertrain module is one of the key differentiating factors in the electric vehicle industry. Automobile manufacturers and Tier 1 suppliers have been spending a lot of resources researching and developing the most advanced electric vehicle technology. The current trend is to achieve a more compact module design with higher power density and system efficiency. For example, Nissan LEAF realizes a very compact electric powertrain module design by integrating an on-board charger (OBC), a DC-DC converter, and a junction box with an electric drive unit (EDU) [2].
This article introduces the EMC design of the powertrain module, which is composed of an electric motor, an inverter, and a mechanical gearbox. The electrical system diagram of the EV powertrain module is shown in Figure 1.
Figure 1: System diagram of EV electric powertrain module
My previous work "Uncovering the EMC Mystery in Electric Vehicle Drive Units" [3] introduced the importance of design engineers' system-level overview of EV powertrain modules. It is important to take EMC design considerations into consideration in the early stages to achieve the overall system design goals. High voltage (HV) EMC regulations and requirements are a daunting task not only for new entrants but also for established companies in the automotive industry. Therefore, we will first review the high-voltage standards and regulations applicable to electric powertrain modules in this article. Then, we will focus on the EMC challenges in the design of powertrain modules and demonstrate design techniques to solve potential EMC problems. Then, engineers will better understand how to design modules that can pass EMC requirements in an EMC test room.
CISPR 25:2016 [4] As a general EMC guideline for car developers, although car manufacturers usually have their own proprietary EMC specifications [5]. Appendix I of CISPR 25 defines the test methods for high-voltage shielded power supply systems in electric and hybrid vehicles. CISPR 36:2020 [6] was recently released, and it defines a vehicle-level electromagnetic field emission test method. It is expected that automakers will soon provide component-level electromagnetic field emission tests that reflect this standard.
(Please note that on-board chargers (OBC) require a different set of test methods related to charging, which are not covered in this article. Discussions on low voltage (LV) related EMC tests, electrical tests, or electrostatic discharge ( ESD) test.)
Annex I of CISPR 25 defines the conducted and radiated emission limits for shielded HV systems. Unshielded systems should comply with the same restrictions as shielded systems. Appendix I of CISPR 25 also introduces the HV/LV coupling attenuation test. This test is performed when the device under test (EUT) is not powered on. Essentially, the test result is the -S21 diagram of the impedance analyzer to the EUT. In order to obtain a good system decoupling behavior, it is necessary to achieve the A1 or A2 category in the minimum coupling attenuation requirements given in [4].
For RF immunity, the ISO 11452 series is related to vehicle component testing. The latest revision of ISO 11452 subparts (eg ISO 11452-4:2020) includes high-voltage component test setup and high-voltage artificial network (HV-AN). It is expected that other sub-components will adopt these high-voltage component requirements accordingly.
ISO/TS 7637-4:2020 deals with transient emissions and transient immunity on high-voltage lines.
The American and Chinese standards SAE J551-5 [7] and GB/T 18387-2017 [8] respectively define the limits and test methods for electric vehicle magnetic field and electric field emission. CISPR 36:2020 [6] Deals with electromagnetic field emissions. A recent comparative study of GB/T 18387-2017 and CISPR 36:2020 on magnetic field radiation disturbance test requirements found that GB/T 18387-2017 is more stringent [9].
Due to the high currents associated with electric vehicles, magnetic field exposure will become the most critical aspect of electric vehicles. In order to also cover this aspect at the component level, a test method must be defined.
Human exposure to magnetic fields was tested in accordance with guidelines issued by the International Commission for Non-Ionizing Radiation Protection (ICNIRP). IEC TS 62764-1 defines procedures for measuring the level of magnetic fields related to human exposure generated by electronic and electrical equipment in the automotive environment.
A summary of the HV EMC test is listed in Table 1.
The list only includes HV EMC testing, excluding LV, electrical and ESD testing
More compact, higher rated power and higher efficiency powertrain modules usually mean more EMC challenges.
Currently, the most common high-voltage rating adopted by modern automakers is 400V (Audi e-tron, Tesla, Nissan Leaf, etc.). The Porsche Taycan is the first electric vehicle on the market to use an 800V power system [10]. For the same rated power, increasing the voltage will reduce the current in the system, thereby reducing the copper loss (I2R loss). Therefore, higher voltage is associated with higher system efficiency. However, the high-voltage rating of the powertrain module is limited by factors such as the rated voltage of commercial power electronic equipment, insulation breakdown of high-voltage cables, and poor EMC performance.
The roadmap for next-generation power electronic equipment for powertrain modules has shown that breakdown voltage levels exceed 1000V. Soon, most car design companies will switch to 800V (and higher) systems to achieve higher efficiency. This predictable trend also poses a huge challenge to EMC design. As the voltage level doubles (for example, from 400V to 800V), and assuming the design has the same parasitic characteristics, due to the higher dV/dt characteristics, the noise level associated with the electric field will increase.
Another huge challenge associated with high-voltage systems is safety. EMC and safety cannot be discussed separately in high-voltage systems. The Global Electric Vehicle Safety Technical Regulations (EVS) [11] defines the maximum capacitor energy that can be stored in a Y capacitor as 0.2J. This hard limit has a profound impact on the front-end input filter design of all HV modules, because the Y capacitor is a very effective broadband noise attenuation filter, especially in the lower frequency range (starting from 300kHz).
This means that when the voltage level doubles, the available Y capacitance value decreases by 75% according to Equation 1:
Where E is the total energy stored in the Y capacitor, CY is the available Y-class capacitor, and V is the upper limit of the nominal voltage of the high-voltage system.
The relationship between the noise level (represented by the common mode current) and the available Y capacitance is shown in Figure 2.
Figure 2: As the voltage level increases, the common mode noise increases proportionally, while the available Y capacitor decreases in an inverse square trend
Since the powertrain module is the performance unit of the vehicle, its rated power directly determines the acceleration, horsepower and torque of the vehicle. Given the rated voltage of the powertrain module, higher power means higher current. Since the current is directly related to the magnetic field, higher power also means an increase in the vehicle's magnetic field.
Since the motor and inductor in the inverter are both inductive, higher current also means higher transient behavior caused by sudden state changes. The back electromotive force (EMF) or recoil voltage caused by L·di/dt may cause stress or damage to the components, if not controlled, and send huge voltage spikes propagated on the HV bus.
Insulated gate bipolar transistors (IGBT) were adopted in the early days of powertrain modules (such as the one in the early version of Tesla Model S). The switching frequency of IGBT-based power systems is theoretically limited to 20kHz. The thermal problem of IGBT usually limits its switching frequency to be lower than its theoretical value. Due to its fast switching speed (thus low switching loss) and better thermal characteristics, wide band gap devices (such as SiC MOSFET) have recently begun to replace IGBTs as the preferred device. Using SiC MOSFET-based powertrain modules can easily achieve switching frequencies of 20 kHz and above.
The disadvantage of using wide band gap devices is that their faster switching events can lead to increased electromagnetic interference (EMI) related problems. The rise time of a silicon carbide MOSFET can be as small as a few nanoseconds, resulting in a slew rate of 50-200V/ns [12]. In order to achieve such fast speed characteristics, the gate driver is equipped with a short and high peak pulse current function, which may also cause EMI problems.
In order to power microprocessors, gate drivers, and analog and digital integrated circuits (ICs), multiple power supplies are usually integrated in the design of the control unit board, the most common of which is the switch mode power supply (SMPS) unit.
The switching frequency range of SMPS units such as buck and boost converters in automotive applications is usually between 150kHz and 500kHz. The rise and fall times of the switch can be as short as a few nanoseconds. Compared with power switching devices, the noise spectrum shows less energy, but the frequency range covered is much wider. Figure 3 shows the switching noise curve between the LV buck converter and the HV inverter.
Figure 3: Comparison of noise curves of switching equipment
Bearing currents are mainly caused by electrostatic discharge, magnetic asymmetry (caused by unbalanced three-phase windings), common mode voltage, and high switching rate [13].
The bearings in the motor have moving metal balls or rollers in a fixed metal shell. A very thin lubricant layer is located between the two components, so it has a high capacitance and can carry high displacement currents. Because the lubricant is too thin and the bearings are not perfect, occasionally electrical failures occur, and even two metal parts are in direct contact. Therefore, the bearing current is partly capacitive, which generates current pulses during each switching period, and partly is random high current spikes [3]. This random breakdown will result in a very high random peak current, resulting in high quasi-peak noise during EMI scanning.
Bearing current will cause the bearing of the motor to age, thereby shortening the service life of the powertrain. The bearing current circulating in the power system can also cause conduction and radiated emissions, as shown in Figure 4.
Figure 4: Bearing current caused by fast switching circulates through parasitic capacitance
It is not uncommon to see many ground symbols in a module design, although many EMC experts [14] [15] have emphasized clarification of the use of the term "ground". A project I reviewed recently had more than 5 ground symbols in a schematic. It is very confusing to see all these symbols at first, let alone how the ground is connected.
In Figure 1, the term "reference" is used instead of the term "ground." It should be noted that the circuit ground is not necessarily the same as the EMC ground. For simplicity and clarity, there can only be one EMC ground, that is, the metal chassis of the device, or what we call an RF reference. The metal casing of the module has a contact point with the vehicle chassis (through direct bonding or mechanical fixing); therefore, we treat the metal casing as a chassis reference.
The high-voltage line is a high-voltage design reference, and it should be isolated from the vehicle chassis reference (via dielectric or Y capacitor). LV design should use 0V as the reference point, which should be the only design reference point. The idea of separating analog and digital ground points is based on a misunderstanding and is not a good design method [16].
The connection between the 0V reference voltage (on the PCB or on the connector pins) and the chassis reference voltage introduces inductance [16]. RF current will inevitably flow in these inductances, creating a noise voltage that helps drive radiation. This is shown in Figure 5. Therefore, the inductance of the connection should be minimized in the module design. In the scheme of reducing the inductance between two reference points, the most effective method is to increase the direct contact area between 0V Ref and Chassis Ref. This is demonstrated in Figure 6. In most cases, multiple contact points around the edge of the PCB and the corners will form a quasi-360 degree contact.
Figure 5: The connection between the 0V reference and the chassis reference creates impedance for the RF noise current
Figure 6: Demonstrates good contact between PCB 0V reference and chassis reference to reduce inductance
The high-voltage cables of the powertrain module are usually shielded, and the shielding terminal and the module housing have closely spaced contacts (360 degrees) [17]. The reason for having low impedance bonding is the same as we explained in the previous section. However, the mechanical quality of the contactor design needs to be considered very carefully, because high temperature, aging, vibration, and chemical intrusion can damage the connection configuration over time. Figure 7 shows an example of a shield connector failure.
Figure 7: Example of weak contact of shielded connector
Front-end filter design is essential for electric powertrain modules because it helps block noise from the inverter power switch. It can also suppress noise transmitted from the outside of the module housing through the HV DC wiring.
There are many types of front-end filters, including the two-stage filter shown in Figure 1. In Figure 1, the L1 and C2 configurations form the first-stage low-pass filter. The second stage filter consists of CMC and Y capacitor. Note that the first stage filter, together with the DC link capacitor C1, effectively acts as an ap (CLC) filter.
Due to its high voltage and high current characteristics, the saturation of the magnetic core needs to be considered when designing high-voltage inductance components. Nanocrystalline materials have very high saturation magnetization. Since windings carrying such high currents inevitably increase the size and weight of the module, toroidal or elliptical nanocrystalline cores are often used for common mode suppression chokes in automotive applications. They are effective in the frequency range between 150kHz and 120MHz.
For power system applications, the magnetic core can be used as an inductor on a single power line (HV) or as a CMC on two power lines. It should be noted here that designers may find that CMC is not required due to other good EMC practices, such as nanocrystalline cores. However, it is better to allow Murphy's law and make proper design from the beginning so that kernels can be added later if necessary [18].
Compared with the use of CMC, the benefits of using Y capacitors include great high-frequency conducted emission attenuation (usually effective from 5 MHz), smaller size, lighter weight, and no saturation problems. The connection of the Y capacitor to the chassis reference should also be designed to achieve very low impedance. The internal equivalent series resistance (ESR) and equivalent series inductance (ESL) are the main factors that affect the effectiveness of the Y capacitor. The impedance imbalance of the two Y capacitors will also affect the common-mode filtering performance. These shortcomings can be remedied by layout or the use of alternative components (such as X2Y capacitors).
X2Y capacitors (see Figure 8) [19] provide excellent EMI filtering and can be used to replace CMC in applications where size, weight, and cost are limited by design (Figure 8(a)). Two capacitors are balanced in parallel to eliminate mutual inductance. X2Y capacitors also provide a shielding effect. Alternatively, as shown in Figure 8 (b), the X2Y capacitor can be configured as a decoupling capacitor with ultra-low inductance. Currently, 500V X2Y components do not meet automotive standards, but this component is worthy of attention because manufacturers may upgrade high-voltage components, so they meet the AEC-Q200 standard.
Figure 8: X2Y capacitor connection in the electrical system, (a) filter configuration; (b) decoupling configuration [19
The DC link includes a high-voltage DC bus and a DC link capacitor. The DC bus should be designed as short as possible and close to each other to reduce the loop area between them. The design of the DC bus capacitor should be able to cope with high voltage, high frequency switching ripple and high temperature. For full power operation of the powertrain module, the capacitance value should be large enough. Low ESR and ESL film capacitors and electrolytic capacitors are often used in powertrain modules.
Film capacitors are widely used in powertrain module design due to their high performance and reliability. The "brick" size of film capacitors limits design freedom; therefore, it is important for design engineers to cooperate with film capacitor suppliers early in the design phase. Parameters such as ESL and ESR are critical to EMC because the self-inductance of a capacitor is caused by the geometry of the components (such as capacitor wrap, upper conductive rail, and lower conductive rail).
In some cases, manufacturers use high-performance electrolytic capacitor banks as the DC link. The current flowing through the capacitor's self-inductance generates a magnetic field. The smaller the self-inductance, the smaller the magnetic field of a given current, and vice versa. Multiple capacitors are connected in parallel with their self-inductance in parallel. However, if they are too close, their magnetic fields will interact, because when the magnetic fields are all in the same direction, the total inductance will not decrease. Therefore, the total inductance will not be reduced to 1/N (as we would expect from circuit theory or SPICE simulation).
However, if we arrange N parallel capacitors closely together, make them alternately reverse or make the self-inductance of the capacitors perpendicular to each other, their magnetic fields will tend to be opposite to each other, canceling them to some extent (because the mutual inductance is now maintained At a minimum). Since a weaker magnetic field means a lower inductance, we may be able to reduce the overall self-inductance by more than 1/N [20].
Figure 9 shows the magnetic field coupling due to mutual inductance between capacitors. It illustrates how the layout of the capacitor array can achieve lower overall inductance through its magnetic field and cancel each other to some extent.
Figure 9: Layout rules for multiple electrolytic capacitors
In [3], EMC design considerations for using wide-bandgap devices such as SiC MOSFETs are introduced. The topic itself can easily inspire some specialized articles. Therefore, we only summarize some best design practices here.
Because the common mode current ICM can be calculated by Equation 2:
ICM = Cstray · dV/dt equation. 2
Where dV/dt is the slew rate of the switching device, and Cstray is the stray (parasitic) capacitance, which can be calculated by Equation 3:
Cstray = CFET CD CL 3
CFET is the parasitic capacitance of the SiC MOSFET (mainly the drain-source capacitance CDS), CD is the freewheeling diode capacitance, and CL is the parasitic capacitance caused by the layout (for example, the capacitance between the device and the heat sink).
Reducing the slew rate helps reduce spikes or ringing, but at the expense of increased switching losses. Cstray can be reduced by choosing optimized packaging and applying good layout practices. The ringing of the switch is caused by the resonance of the LC circuit, and good layout practices that achieve low stray inductance (for example, the device is connected to the bus) can help reduce the ringing.
In order to share the large current, N SiC MOSFETs are placed in parallel. This configuration results in 1/N RDS(ON), allowing very low conduction losses. For the same reasons we explained when discussing multiple capacitors in parallel, the total ESL of the device may not be as low as 1/N. However, methods of shortening the connection, such as the connection between the device and the bus and the connection between the device and the motor winding, can minimize inductance.
Figure 10 shows the layout of the SiC MOSFET in the Tesla Model 3 powertrain module. Four MOSFETs are placed in parallel to form a switch block. All in all, there are 24 switching devices in a very compact package space with short connections to minimize parasitic inductance. Sintering the SiC MOSFET directly to the bottom of the heat sink helps to dissipate heat effectively.
Figure 10: SiC MOSFET layout in Tesla Model 3 power system inverter
Some techniques are introduced in [3], such as the use of SiC Schottky diodes in parallel with SiC MOSFETs to eliminate reverse recovery charge effects. But more and more manufacturers are integrating very fast and robust intrinsic body diodes into the device package; therefore, separate anti-parallel diodes are not required. Generally, placing the decoupling capacitor array close to the switching device is also important to reduce the ringing effect of switching events.
Compared with the power stage, low-voltage control units often have various sources of high-frequency noise. The noise spectrum of the control unit covers a wider range, as shown in Table 2.
The EMC design follows similar guidelines to the guidelines we discussed earlier, namely applying good layout practices, designing front-end and output filters on power and signal lines, and applying sufficient global and local decoupling capacitors.
CMC is often seen in the design of low-voltage power supply systems. The X2Y balancing capacitor was introduced earlier in this article. Although HV (above 400V) X2Y components have not passed automotive certification, there are a large number of AEC-Q200-compliant components [19] that can be used in the control unit of the powertrain module. [21] X2Y capacitors were introduced in the SMPS design.
At the design stage, there are two methods that can be used to reduce bearing currents, as described below.
Due to its mechanical properties, ceramic bearings or hybrid bearings (combination of steel ring and ceramic) are ideal for electric motors. Because ceramic is an electrical insulator, it can reduce bearing current and reduce arcing [22]. Alternatively, the rotor of the motor can be directly grounded through a small thrust bearing, which can be easily replaced [13] [23]. This grounding also helps prevent the motor shaft from radiating due to the stray radio frequency current induced into it.
Another way to reduce common mode noise (contributing to shaft voltage and bearing current) is to add common mode filters along the motor windings. Shielded cables can also help [13]. But for compact module design, these methods are generally not considered.
In addition to using the hardware method, a certain switching scheme can also be used to reduce the common mode voltage of the motor, thereby reducing the bearing current. The method proposed in [24] is proven to achieve high performance and low common-mode voltage and current, but unfortunately, compared with the normal PWM scheme, it does not include a spectrum analyzer evaluation of the results. The implementation of the proposed method also requires an in-depth understanding of the motor drive system.
In this article, we reviewed the HV-related EMC regulations for powertrain modules in EV applications. Then, we discussed the design challenges they presented and demonstrated EMC design techniques that can be implemented in the design phase.
Most of the technologies introduced in this article follow EMC design principles, such as reduction of parasitic parameters, 360-degree shielding, and bonding. You can also consider new passive components (such as nanocrystalline cores and X2Y capacitors), as well as software solutions to reduce common mode noise. By adopting these design techniques, engineers can be more confident that the powertrain modules they design will pass EMC testing, even for the first time!
Dr. Min Zhang is the founder and chief EMC consultant of Mach One Design Ltd, a UK-based engineering company specializing in EMC consulting, troubleshooting and training. His in-depth knowledge in power electronics, digital electronics, motors and product design has benefited companies around the world.
Your email address will not be published.
Save my name, email, and website in this browser for the next time you comment.
In Compliance is the main source of news, information, education and inspiration for electrical and electronic engineering professionals.
Aerospace Automobile Communication Consumer Electronics Education Energy and Power Industry Information Technology Medical Military and National Defense