The 2003 IEEE International Symposium on Electromagnetic Compatibility (EMC)

11-16th May 2003, Istanbul, Turkey

  

Perspective of an Automotive Electronics Design Engineer

 

There are few, if any, EMC conferences specifically tailored to Automotive EMC.  Most test engineers and designers of products requiring compliance with Automotive EMC specifications have to attend more generic EMC conferences, where these are within their companies' training budget.  Many design engineers cannot justify attendance at these generic conferences, as their automotive content is too low.  Consequently the number of automotive related EMC topics at general EMC conferences remains low and likewise it appears that interest in this area is similarly low.  This assumption has recently been proved incorrect; the launch of the Automotive EMC Professional network (www.AutoEMC.net) last month has highlighted that there is significant interest in automotive specific EMC, the network having gained over 1000 subscribers within a month.

The Istanbul Hilton Conference centre, venue for IEEE Symposium.

   

This review is of the seven papers (of almost 400 in the proceedings) that are specifically automotive related and were presented at the IEEE Symposium.  A summary of some of the other papers that might be of interest to electronics engineers working in the automotive field is also provided.

 

Automotive Specific Papers

 

The Automotive EMC Test Chambers at General Motors, Milford, Michigan USA

Wiles

 

This paper outlines the recent improvements and modifications to the General Motors (GM) test facility in Milford.  The facilities are very impressive, offering 4 large chambers capable of immunity and emissions testing of the whole range of GMs’ vehicles (including busses and large trucks).  The emissions capability is relatively straightforward for such large chambers and although not to be taken lightly is not as impressive as the immunity capability.  The chambers are capable of continuous field strengths of 200V/m up to 30MHz and 100V/m up to 1GHz with pulsed fields exceeding 100V/m at 9.5GHz so far proven, they are also capable of tolerating 200V/m at the upper GM frequency limit of 18GHz.  The chambers also feature vehicle-sized turntables with in-built dynamometers.

 

The facilities are without a doubt “state-of-the-art” for Automotive EMC testing and undoubtedly are an impressive sight.  This paper was denied a better impact by having to be presented in the “Open Forum” (poster) session, possibly due to its relatively low technical content (no new test methods or specification impact conclusions).  The paper does, however, give a very good overview of the facilities required to provide whole vehicle testing and implies the enormous cost this level of testing must incur (costs are not explicitly mentioned).

 

Double Bulk Current Injection: A New Harness Setup to Correlate Immunity Test Methods

Cuvelier, Rioult, Klingler & Fischeux.

 

The paper provides an analysis of why the three common methods of automotive immunity testing (free-field, stripline and Bulk Current Injection; BCI) produce such differing and incomparable results.  The primary reasons are related to the impact of the field on the harness in each test method; free field is synchronous field across the harness length, stripline is a moving field along the harness length and BCI is a point impact method.  The authors then went on to use double BCI probes (at either end of a given harness) to produce similar field impact and enable the DBCI method to reproduce induced currents in a harness that replicate free field or stripline test methods.  The DBCI method is then taken a stage further and a “U” shaped harness configuration is suggested that enables all three test methods to produce similar induced current effects in a harness from either of the three test methods. 

During conference breaks the poster sessions were relatively busy.

 

Although the final “U” harness method is an interesting method for producing test correlation, it seemed slightly impractical and unrepresentative of in-vehicle harness configurations.  The use of BCI on-vehicle was not investigated and this would be a good avenue for further study.  Having come across instances where problems on-vehicle were not found in the test chamber, it would be useful to try the DBCI method on-vehicle to simulate the test method (free-field or stripline) that a product had “passed” in the test laboratory.

 

Analysis of the CISPR 25 Component Test Setup and its Validation

Kris & Mullner

 

The paper analyses the origin of repeatability problems in the CISPR-25 radiated emissions test configuration, particularly above 108MHz where the LISN impedance is undefined and over 200MHz where the harness exceeds an integer of the radiating wavelength and radiating pattern effects are observed.  The authors suggested use of site attenuation (SA) characteristics to correlate between OATS and anechoic chambers.  This is not especially new as normalised site attenuation (NSA) methods can and have been applied to anechoic chambers in the past for validating results.  What was new (to this delegate at least) was the proposed use of five point source emitters for the SA, although due to symmetry this can be reduced to three emitters.  The new method is being used successfully for Automotive EMC Laboratory Recognition Program (AEMCLRP) accreditation by the authors.

 

Another interesting point that the authors’ highlighted was the effect of the ground plane bonding on the emissions, particularly in the vertical polarisation.  The effect of ground plane (table) bonding has been examined before and the results given here show some of the lower frequency (near 50MHz) resonance effects other papers have highlighted (some of which are included in the papers references).  The paper includes simulations of the various ground bonding schemes.

 

Although the paper was specifically aimed at CISPR-25, due to the similarity of the test methods, the results apply equally to 95/54/EC radiated emissions testing.

 

Automotive EMC Test Harnesses: Standard Lengths And Their Effect On Conducted Emissions

O’Hara & Colebrooke

 

This paper originates from the frustration of having to supply multiple cable harnesses for different EMC tests and the desire to investigate why such a plethora of harness lengths have developed.  The paper examined 5 standard lengths (0.2m, 0.5m, 1m, 1.5m and 2.0m) under a single test method (conducted emissions) to examine what difference using the “wrong” cable length has on this test and if results from a single cable length could be used for all test types.  Comparison of wired and chassis ground return is also included.

Technical sessions had good attendance despite being split over 6 different auditoria.

 

The conclusion is that generally the overall levels are not significantly affected in this test method, but the onset of resonance for longer harness lengths is the primary reason for using a short harness for this test method.  The harness requirements in CISPR-25 in particular are obviously well thought out, even if the reason for the differences is not always obvious to the test engineer or designer.

 

Automotive Equipments EMC Modelling for Electrical Network Disturbances Prediction

Poinsignon, Matossian, Mazari & Duval

 

The paper proposes a relatively simple modelling method for transient and conducted emissions based on the Thevenin approach.  An example of a wiper motor on a Renault Laguna vehicle is provided and shows a very high degree of correlation between simulation in Matlab and measured response.  One of the main appealing features of this simulation approach is the relatively simple extraction of the model parameters and model construction in Matlab.  The method could be modelled in many other mathematical modelling software and even some circuit simulation programs that allow control of voltage and current sources via mathematical equations. 

 

Unlike some of the more complex simulation schemes proposed elsewhere, this paper gives a method that can be easily used by most electrical engineers and would allow provision of models to integrators that are simple enough to produce reasonable whole-system simulations within relatively short simulation timescales.

 

RF-Properties of Automotive Traction Batteries

Hoene, Guttowski, Saikly, John & Reichl

 

A paper on modeling of standard automotive lead-acid batteries has been provided by BMW engineers at a past IEEE conference (reference not available at time of writing) and this paper takes a similar approach to Li-ion and NiMH batteries intended for traction use (i.e. in hybrid or electric vehicles).  The paper goes further than previous reports in deriving a model that can be directly related to some of the chemical as well as physical properties of the battery, while at the same time maintaining a relatively simple equivalent circuit model.

 

The parameters for the model are fully explained and could be relatively easily adapted for other chemistries and/or physical plate arrangement (the target batteries were cylindrical and prismatic construction).  Another well thought out inclusion in the model is the skin effect, making the internal impedance frequency dependant, a factor easily overlooked.  The correlation between model and measured response are extremely close from DC to 40MHz (the limit of the test equipment used) and simulation results up to 100MHz look sensible.  

 

EMI Coupling from Automotive Traction Systems

Weber, Guttowski, Hoene, John & Reichl

 

The paper examines the mechanisms for coupling in electric vehicle (and hybrid vehicle) installations where there is a high power electric drive content in the cable network.  The paper proposes new test methods for assessing the coupling from these new high power drive sources and offers models for these installations to assess the impact of high power/high voltage drives on lower voltage/lower power networks.  The simulation results are in close correlation to measured phenomena.

The paper makes a very useful comparison between the effects of shielded cable and filtering on the vehicle networks, offering methods of assessing the cost-weight-effectiveness trade-off that is usually necessary in vehicle design (usually with cost winning out).  The results also allow the authors to make suggestions for limit levels for these new power networks that correlate with existing CISPR-25 approaches while being determined specifically for the higher power requirements of traction systems.

Accompanying exhibition was well supported by the EMC industry.

 

Editors Note: AutoEMC.net have approached all these authors to request permission to publish these papers in the technical papers section of the website.  Please check this section if any of the above papers are of interest.

 

Other Interesting Papers

 

There are two areas that may be of interest to automotive engineers, the areas of aerospace that shares many of the environmental EMC constraints and those papers concerned with EMC at the PCB and IC level.  There were several papers on cable EMC that may also be of interest to systems integrators in the automotive industries, but these were primarily targeted at industrial installations and computer networking.

 

The most interesting aerospace paper concerns the use of personal electronic devices (PEDs) in commercial aircraft.  Although more severe in commercial aircraft due to the possibility of superposition of multiple PEDs, the implied effect has potential application in drive-by-wire systems and in a vehicle any PED will be physically closer to the control surfaces.  Most other aerospace papers concentrate on the testing issues and in particular there are several papers on reverberation chamber testing, some on smaller products as well as the aerospace applications.

 

There was a discussion session on grounding in chips and systems that included a paper on grounding for ungrounded systems that really only highlighted the areas of difficulty rather than offering solutions.    Other relevant papers in the emissions and immunity session included predicting maximum emissions from ground bounce using only resonance characteristics of the power plane, predicting the field at 1m from PCBs and IC's based on near field measurements; relevant as this is the measurement distance for ESAs in most automotive emissions testing.  The paper on immunity of ADCs highlights that the integrator and multi-vibrator are the dominant susceptible circuits within an integral ADCs.  Signal integrity and thermal properties were considered simultaneously in a paper that found multiple ground stitching helped with both situations and another paper provided a very simple mathematical method of estimating the shielding effectiveness of multiple ground stitching.  A technical session on EMC in microelectronic systems was a particular highlight for IC level EMC papers with work demonstrating significant progress in the measurement and modelling of IC level EMC and some very impressive results on visualisation of chip level EMC for design improvement and on-chip/off-chip decoupling. The modelling of IC EMC characteristics and provision of relevant EMC information from IC manufacturers is still considered to be a few years away and will probably be supplied on an “as-requested” basis, but some IC vendors are clearly working in this area and making significant progress in their understanding of IC level EMC issues and the impact on PCB and system wide EMC problems.

 

Other interesting sessions included two on ESD, one aimed at high technology manufacturing industry and another in the chemical processing and hazardous environment arena, both still had some relevance to general ESD in any industry or application.

 

Overview

 

There was definitely inadequate coverage to justify this conference for an automotive electrical designer, even when the IC and PCB level papers are included.  The majority of the papers are aimed at testing and test houses and also a significant number of papers on simulation and numerical methods, hence the conference was primarily an “EMC for EMC Engineers and Academics” event rather than targeting design solutions for EMC and practical engineering methods.  The papers that are relevant to the automotive designer are in general very good and deserve a wider automotive audience than they achieved at this conference.

 

Martin O’Hara

24th May 2003  

 


© www.AutoEMC.net 2003                                               TOP OF PAGE                                                  HOME

 

Breakdown of papers presented by application.

 

 


© www.AutoEMC.net 2003                                               TOP OF PAGE                                                  HOME