Automotive technology innovations—such as digital FM radio broadcasting, remote keyless entry, tire pressure monitoring systems, GPS, satellite digital audio radio service, Bluetooth, and Wi-Fi—have brought new complexities to automotive antenna design. At the same time, the continual increase in clock speeds and the higher density and more complex structure of today's integrated circuits (ICs), printed circuit board (PCBs), and connectors mean that many components can act as an antenna, transmitting signals to other components that happen to be in the area. The potential for electromagnetic interference (EMI) is increasing due to many factors: an increased number of embedded control units (ECUs), higher onboard diagnostic (OBD-II) data rates, increased number of controller area network (CAN) lines, etc.
The electromagnetic behavior of the entire vehicle can be experimentally tested only when the first complete prototype is available. Problem resolution at this phase of the process is expensive and time-consuming. Furthermore, the complexity of today's electronics and the huge number of possible product configurations offered by many companies make it impossible to fully test most automobile models.
A new generation of simulation tools makes it possible to predict and correct in advance major things that can go wrong with the electronics. For example, simulation can identify EMI that is emitted by high-speed electronics components and determine the effect of that radiation on vehicle subsystems. Unlike physical testing, simulation makes it possible to simultaneously consider the effects of potentially conflicting design requirements, such as keeping electronics cool while avoiding unintended emissions.
There are several automotive standards designed to reduce the probability of EMI through laboratory testing. One of the most important standards is ISO 11451-2, intended to determine the immunity of vehicles to electrical disturbances from off-vehicle radiation sources. An antenna radiates a vehicle in an anechoic chamber while electronic subsystems are operated to ensure their performance is not affected. The prototype and equipment required to perform this test is very expensive, and the test takes long periods of time, which limits the number of times that it can be performed during a development cycle.
The greatest challenge in simulating this test is the large computational domain required to model the air region. The hybrid finite element and boundary integral method (FEBI) helps to overcome this challenge by using the boundary integral—a method of moments (MoM) solution for Sommerfeld's radiation condition—as the interface boundary for the finite element solution. This eliminates the need to model the air region while providing an exact mathematical calculation of the far-field radiation condition. The technique can be used to perform full-vehicle simulation including complex geometries and dielectric materials without having to simulate the air regions, resulting in accurate simulation with fewer computing resources.
In a study, the electric field distribution and antenna radiation far-field pattern at 1 GHz for a complete vehicle simulation according to ISO 11451-2 was conducted using the conventional finite element method (FEM) approach. The air region was modeled for the entire room, including the absorber elements on the side walls. For this particular simulation, 89% of the total number of elements were used to model air. The model was solved using the domain decomposition method (DDM) on a high-performance computing platform with 12 nodes in 310 min with 75 GB of RAM.
The same test was simulated using the FEBI method. The big air box that comprises most of the model when using FEM was replaced with two much smaller conformal air boxes whose outer surfaces are very close to the antenna and vehicle. The absorber elements were replaced with the integral equation (IE) boundary, which yields the same results. The antenna far-field patterns for the two different methods are nearly identical, indicating that FEBI is essentially equal to FEM in accuracy. Yet, FEBI required only 28 min of solution time and 6.8 GB of RAM on the same 12-node HPC platform. So both solution time and computational effort were reduced by a factor of 10 by using FEBI.
The FEBI approach can also be used to test the immunity of ECU modules. A PCB connected to an engine wiring harness is introduced into the simulation. The transmitted signal travels from a sensor, located at the bottom of the engine, to the PCB via a wiring harness.
To understand the role of the wiring harness, two simulations were performed with all of this geometry plus the car and source antenna. In the first simulation, three wiring harness cables were connected to the PCB. In a second simulation, the wiring harness was removed, and a random CAN J1939 signal was applied directly into the connector on the PCB. The results show a resonance on the PCB when it is connected to the wiring harness. The frequency of this resonance is a function of the cable length that is attached to the PCB. The results also show that the coupling between the source antenna and the PCB is increased by over 30 dB between 152 MHz and 191 MHz when the cable harness is attached to the PCB.
The impact of EMC on automotive components must also be considered. The operation of the airbag and infotainment systems depends on microcontroller unit (MCU) speed. Operating speed of the MCU, in turn, depends upon the quality of the power that it receives. Poor PCB design can cause a 100+ mV drop, consequently reducing MCU performance by 40 to 60 MHz, so the PCB must be designed to ensure the performance of the MCU. Power integrity is a global issue involving chip, package, and PCB design, so it cannot be properly addressed in individual domains. Existing package design or selection methods rely on very rudimentary chip power estimations, such as total power, so PCB designers often have no information on transient characteristics of chip power consumption.
Compact models such as the Chip Power Model (CPM) can be used in systems-level simulations for low-power design of ICs. A CPM is a compact SPICE equivalent circuit model that captures the full-chip switching current signatures as well as the parasitic network of the chip power network. The chip power model can enable IC power-aware package selection design; during the post-layout stage, it can be used for IC package coverification as well as PCB power delivery network design and optimization.
Juliano Fujioka Mologni, Senior Application Engineer, ESSS, and Markus Kopp, Lead application Specialist, ANSYS wrote this article for AEI.