There has been much to do about the sudden unexpected acceleration EMI connection. There is no shortage of opinions on this subject. We’ve heard from manufacturers and their EMI consultants as well as from many recognized industrial and academic EMI experts. Most manufacturers have been very supportive of each other but have been very careful not to utter direct EMI related statements as they would rather not be seen in the spotlight of any investigations aimed in their direction. I guess they think that the EMI gods have decided to target just one automobile manufacturer.
I am writing this letter because I believe that, all too often, we can get too close to a problem and overlook some obvious possibilities. Anyone that has done system and sub-system EMI trouble shooting knows, there are interactions/resonances within a single box as well as between subsystem and system boxes that can add to or subtract from the overall EMI signature of the equipment under test (EUT). Of course that works both ways making it both an emissions and susceptibility issue but the later is of most concern here. That’s pretty basic. The way the EMI compliance test system works today, when we take a product to an EMI Test House for compliance confirmation, we expect and get testing performed to a single threat over the entire relevant standards driven test range, to the “minimum” test levels that the standards allow. In fact the basic goal of today’s anechoic test chamber designs is to eliminate all signals except those that are part of the test process. This might be acceptable for stationary EUT such as refrigerators, stereos, PCs, TV, recorders etc. That is not to say that automobiles are not tested to higher standards, they are. More on this later. My point is, all electronic products are tested to survive one single EMI threat at a time.
In the mechanical engineering arena (e.g. The European Union Machinery Directive) a single fault condition is not allowed unless the safety function is maintained. In other words, the box must fail safe. The point is, because machinery is more robust than electronics, failures are fewer and more predictable, even though they share environments. In the EMI environment the same criteria must be considered along with the anticipation of multiple threats. This process will be complicated by the fact that EMI complications are more numbered and complex. In any case the product must always fail safe.
In the real world, when you have a “EUT on wheels” chances are, sometime in its travels, it is going to encounter “multiple EMI threats” (1, 2, 3 or 4 maybe more) simultaneously and just like the individual box environment mentioned above, sometimes the threats can combine, subtract and negate each other. Other times these culprit signals might add or combine in such a manner that allows them to break through the single threat tested EUT, as this condition is not tested for. It is much like a chemical reaction. One chemical added to a solution may make no noticeable difference but together with another chemical or chemicals it can change the original solution to something totally different. The number of threats can be changed with location and EMI environmental conditions.
Try this: Just as an exercise, the next time you are sitting at a stop light at a busy intersection surrounded by department stores, gas stations and other industrial sights, see if you can identify the number of potential EMI threats. e.g. is a big rig behind you with the driver keying his CB Radio? Is the wireless burglar alarm in the hardware store across the street turned on? Is the red light in front of you smart enough to sense traffic patterns? Are there high tension electric power cables or microwave towers in the area? You get the idea?
More possibilities: Now suppose a production line is being held up because some particular EMI filter components are out of tolerance or maybe not available and a substitution is made. How about if some shielded wires or some non-metallic connectors were shipped by mistake to the manufacturer? These are real issues that manufacturing people must deal with every day. I only mention a few to give you a flavor for the unpredictable. Chances are, not one of these mishaps will have any effect on the performance of the end product. Also there is always a tremendous amount of pressure to get the product out the door. The proper way to prove EMI compliance is to re-test these products. In the present day manufacturing culture this is highly unlikely. Chances are in this whole process an EMI Engineer was never even consulted. These types of decisions are made every day by well intentioned individuals who are trying to do the right thing. These issues will get resolved and not always favorably. The product will almost always get shipped. Now we have a number of automobiles on the streets and highways, with their EMI defenses compromised. Some may drive for years with no problems and some may be susceptible to multiple EMI events from day one. Some could cause the subject condition. Over time some of these issues may be addressed, after the fact, with recalls or Technical Notes to the dealers. Some will not.
Remedies: One way to find the cause of such problems is to recreate the exact EMI conditions that triggered the event in the first place. This of course, is difficult (but not impossible) because some of the threats may be moving in different directions at the same time. This also explains why, once an event like sudden unexpected acceleration occurs and the automobiles’ ignition is turned off (reset), and then re-started; the problem disappears. This, in my opinion, renders the “floor mat over the accelerator pedal defense” moot. If it was really the cause, it would still be there after the automobile was restarted (it is a mechanical condition). Another possibility is to select or create a suite of tests (at the correct levels for the application) which would assault the EUT with multiple threats at the same time. From this approach a more real world set of tests could be developed to address these real world problems.
Unfortunately there is not much that today’s standards authorities can do when the foxes have the keys to the hen house. In 1947 the FCC came to the conclusion that there were not enough public complaints on “ignition noise” to take any action. Over the years the FCC allowed the automakers (who were, of course, the closest to the problem) to call the shots. Committees were formed and disbanded but the one constant was that the automobile manufacturers held on tight to the control of test levels. The SAE (automotive standards) Specifically SAE J551, now -2, became the default automotive EMI authority in the US. The SAE as well as their global counterparts have done a fine job over the years trying to create and maintain the relevant standards. That, in itself, is not the problem. I believe the problem lies in the immunity levels being too low and the single threat method of testing being insufficient but most importantly, I question those that control the outcome.
The auto manufacturers have worked together on test levels for EMI over the years. Further, they perform most of their own EMI testing citing the high cost of test facilities and proprietary design preservation. This from the people who over the last 20 years have been dragging their feet on improving overall fleet gas mileage and suddenly, in the last year, have broken the 30 miles per gallon barrier on many of their models.
The one thing we can count on is that the automobile manufacturers will continue to declare their commitment to the safety of their automobile customers. They will continue to sell this notion on the airwaves in the form of our favorite TV commercials, but their true focus is the bottom line. Until an accredited independent organization with enough authority is allowed to step in and tell the automobile manufacturers what the realistic levels of immunity for ground vehicle applications must be, anything else is just an excursive in “due diligence” which in this context is just another legal defense strategy. Even if NASA gets involved, unless they are given complete policy making control, the results are already in.
Solution: Guess what boys; it is time to man-up and use the required bonding and shielding techniques used successfully by the military for decades.
I have to close by quoting a wise old engineer who said, before you can begin to solve any engineering problem you first “gotta wanna.” 
William D. Webb, EMC, Safety and Quality, Butler America, Shelton CT., present assignment United Technologies Corp., Sikorsky Aircraft, Stratford CT.
EMC Engineer for Mission System Integration. Contact Information: compliancehelp@yahoo.com
