EMC: the driving factors
06 April 2018
As demand for electric vehicles (EVs) increases, achieving electromagnetic compatibility (EMC) becomes ever more challenging. This piece explains the problems the industry faces – and stresses the importance of EMC best practices.
In the densely packed automotive environment, digital control units containing sensitive electronics (with multiple voltage rails: 3.3V, 2.5V, 1.8V, and even as low 1V), sensors and actuators – as well as modern communications systems such as USB, Bluetooth and Ethernet – must co-exist with the high-voltage and high-current electronics of the electric drivetrain and battery systems.
This creates a situation where unintended electromagnetic (EM) field effects and noise interference (to, for instance, wireless communication links, navigation or entertainment functions – or even safety-critical functions) are of growing concern.
All modern vehicles contain electronic control units (ECUs), typically a master and several slaves, connected via bus protocols such as controller area network (CAN bus) and Fast Ethernet. The extent of the ‘EMI problem’ grows proportionally to both the number of ECUs deployed and their complexity. One example of this is the use of integrated circuits (ICs) with shrinking geometries and increasing transistor counts, and signals clocking at higher speeds on- and off-chip.
An ECU’s capabilities, such as speed and efficiency, are of course governed by the soundness of the hardware and software design, plus the layout of its PCBs. However, an ECU’s ruggedness and reliability are heavily dependent on adequate electromagnetic compatibility (EMC), which is becoming much harder to realise for all forms of EVs currently on drawing boards.
The main components of an EV’s drive system are one or more electric motors, a power converter, power cables, and power sources. These power sources may include batteries, fuel cells, mains electricity and power generated through regenerative braking. These elements make up the vehicle’s power distribution system (PDS), which has considerably higher voltages and power levels than a combustion engine-only driven vehicle. Power switching circuits are a notorious source of EMI, as power transistors switch in and out high currents – at high frequencies – into inductive loads.
The proximity of high field strength EM radiation (from high power cables and electric motors) to sensitive high-density electronics requires a holistic approach to designing with EMC in mind.
Ramping up the power
The electric power train is introducing voltage and power into vehicles at levels far higher than have previously been managed. For instance, depending on the vehicle concept, voltage levels are on the increase: 48V has been a contender for several years and next-generation EV battery charging voltages could be around 500V.
While considerable, the PDS is not the only noise source of EMI within the vehicle. Aside from further systems in the EV, other such sources include the vehicle’s cables, its harnesses, and of course any devices (such as smartphones or tablets) brought in by the occupants. There are also sources of EMI outside the vehicle, including phone masts, power lines, and radio and radar transmitters; and in this respect it is worth noting that the increasing use of composite materials and plastics in vehicles means fewer EM shielding opportunities.
It is clear that the future EMC behaviour of sensitive electronics, such as communication units (on-board Ethernet or vehicle-to-vehicle communication, for instance) in EVs will require a further significant reduction of the EMI generated by the electric power train components. In some cases, unaddressed EMC issues are merely an inconvenience (for example, one EV OEM is unable to supply AM radios with one of its models because of interference issues). However, when EMI has the potential to disrupt control systems, it becomes a far more serious problem.
Some ECUs feature ARM core-based processors and high speed memory devices. Even FPGAs, once considered too expensive for mass market automotive, are now being used to provide the ECU with firmware upgrade option. They are also used for the high speed and parallel processing of data – as required by some embedded vision-based Advanced Driver Assistance Systems (ADAS).
ADAS points us to another trend in the industry: we are seeing increased reliance on vehicular electronic systems, which are becoming safety critical as we head towards full autonomy. In other safety critical industries, such as aerospace or the nuclear power industry, redundant systems and majority voting help to combat instances of possible data corruption.
But for most vehicle systems, redundancy is not an option, due to cost, weight and time- and volume-to-market reasons, so resilience to EMI (and its generation) rests firmly with design-for-EMC practices.
In the automotive sector, a good design is one that strikes the perfect balance between system performance, size, cost, and time-to-volume production. However, virtually every design decision effects the overall EMC characteristics of a PCB. Minor changes, such as varying a component’s placement and the routing to or from it, can have a have a big impact. Thankfully, all ECAD systems include rule checkers. These can, for example, assess the design data geometry and look for stances where signal crosstalk may occur, or where decoupling may be required – all in a quick and easy manner without exhaustive simulation setup.
These are, however, post-layout checks. It is always best to design with EMC in mind – rather than embark on a trial-and-error exercise. Common-sense PCB design rules include:
- Keeping clock frequencies as low as possible and rising edges as slow as possible (within the scope of the requirements spec);
- Placing the clock circuit at the centre of the board, unless the clock must also leave the board (in which case place it close to the relevant connector);
- Mounting clock crystals flush with the board and grounding them;
- Keeping clock loop areas as small as possible;
- Locating I/O drivers near to the point at which the signals enter/leave the board; and
- Filtering all signals entering the board.
Simulations are useful, as long as you know what to expect: it is just as important to have confidence in pass results, as it is to understand why a simulation fails. Also, the simulations are unlikely to provide any steer on what the radiation levels are. For a more advanced analysis, full EM simulations are required.
The combination of emerging and planned higher voltage/power drive trains with ever-sensitive electronics – which will increasingly be for safety critical applications as we head towards autonomy – means it is more important than ever to design with EMC in mind. In this respect, your ECAD tool, a fundamental understanding of EM ‘sources and victims’, and adherence to best practice and ‘design with EMC in mind’ rules will all serve you well.
Contact Details and Archive...