From the technical perspective, creating a universal standard for EMC tests performed worldwide would not be difficult, and it would eliminate so many problems. For example, if the same model computer system is purchased by the U.S. Navy, the Cleveland Clinic, a college student in Michigan, a Japanese businessman, and a manufacturer in Paris, the EMC requirements are different at each location. That same model computer must be tested to and meet multiple EMC requirements, each with its own test procedures. This forces manufacturers to perform similar but different EMC tests based on the belief that the same product will perform differently depending on who owns it and where it is being used.
All the EMC requirements alluded to—for example, military, medical, FCC, VCCI, EU, CISPR, and IEC—have sound testing procedures. In some cases, the procedures used for the different EMC requirements are based on the same parent standard.
The goal of the various test standards is the same. The procedures aren’t all that different and could be standardized. However, different requirements continually are being developed for specific problems. As a result, the cost of the test program grows unnecessarily because meeting all the laboratory EMC tests that can be conceived won’t guarantee that the equipment will work in its intended environment.
The goal should be all for one standard—one standard for all. The emphasis should be improved productivity; lower manufacturing, test-facility, and test-equipment costs; and a reasonable assurance that the product will not have an EMC problem.
Limits Depend Upon the Use
It is true that farming equipment used on a typical day in a rural environment might only need to operate in a 1-V/m to 3-V/m average RF field strength while a product used on a modern military aircraft carrier might have to operate in a 27,000-V/m peak RF field strength. These differences are significant. It may not be possible to have a single limit that applies to every user environment, but there certainly is no reason why we have to make those measurements so many different ways. Differing standards result in differing and often conflicting approaches to measurements.
What This Article Does
This two-part article takes a brief look at the tests called out by MIL-STD-461E, FCC Part 15, and the EU requirements and examines some of the major problem areas that result from the differences in test procedures. These problem areas can be loosely divided into test sample problems and measurement system problems.
The first part of this article addresses test environment problems, and the second will address the measurement problems. A study of these areas also will make it easier to understand why it is so difficult to compare data from one test standard with another and helps to support the need for a universal EMC test standard.
Globalization of the electronics marketplace and the military emphasis on using commercial off-the-shelf (COTS) equipment strongly substantiate the need for standardized tests. Multiple limits simply could be superimposed on the data to determine if the test sample met various requirements. Translation from one requirement to another would be easy.
First let’s look at a universal test sample. This product has analog and digital circuits to perform its intended function, one or more power supplies or battery chargers, signal and control cables, possibly an antenna or optical link, and input/output devices with associated displays.
This article will not discuss antenna radiated or conducted emissions as a separate issue. But if the antenna is removable, and the RF power output can be handled by standard laboratory test equipment, antenna emission tests always should be done as conducted measurements to minimize the errors created by the interaction of the radiated energy with the environment.
Figure 1 shows a universal test sample setup for EMC testing. The product can have four types of EMC coupling problems: radiated and conducted emissions and radiated and conducted susceptibility. Test-sample size can make a big difference in its radiation characteristics.
Table 1 indicates the military and commercial tests that would be required for the test sample. Regardless of the EMC test standard, the product must be set up in a controlled environment. This includes providing the test sample with standardized power to compare results from one lab to another.
The artificial measurement environment is a good example of the many differences in the EMC measurement standards. The test environment is a trade-off between real-world evaluation and laboratory repeatability. Each approach has its proponents. The choice is determined by the level of testing that is necessary to evaluate the test sample within the range of acceptable risk.
In Situ Testing
In situ testing is the best. The test sample is installed into its actual operating environment and EMC tests are preformed while the system is running. The test data then applies exactly to the operation of the product in a multiple-emitter environment.
This is a big advantage over testing approaches that use a single-frequency susceptibility environment. Unfortunately, each test sample must be individually tested because each operational environment is unique. Also on the downside, if the test sample fails, it’s too late to make inexpensive changes.
Earlier military EMC standards used simulated installation. Since tanks, ships, airplanes, personnel carriers, and transportable equipment shelters are metal, the tests performed in an ordinary metal-box shielded enclosure would approximate testing the unit in its operational environment, plus the highly reflective enclosure walls would minimize the need for rotating the test sample. This approach is used in optional MIL-STD-461E mode stirred tests at the higher frequencies.
Unfortunately, metal-box shielded enclosures introduce errors into the measurements, resulting from multipath reflections (some signals are in phase, some are out of phase), enclosure resonance, and antenna loading. This plays havoc with repeatability, and if the same product is tested in different-sized enclosures, the data will be different.
Open Area Test Sites
Semi-anechoic and anechoic chambers and open field and free space (open field + absorber on the ground in the specular region) are used to eliminate or minimize errors introduced principally by reflections and will minimize variations in measurements between facilities. The semi-anechoic and anechoic chambers generally are shielded enclosures lined with absorber materials and suffer some of the problems of metal-box enclosures. There still will be resonance, some reflections off the sides of the absorber material, and antenna loading.
The open-field locations are more homogeneous and nearly reflection free, but they are not protected from the environment. Adding shelter increases reflection problems in some frequency ranges which are determined by the shelter material.
Around the world, power-line impedance varies from a few tenths of an ohm to 600 W or more depending on the impedance of the power source and whether the power is supplied to the test sample through metal conduit, buried underground, swinging from power poles, or delivered by batteries. If the power-line source impedance is greater at some frequencies than the test sample impedance, RF noise will flow to the test sample. If the impedance is lower, RF noise flows from the test sample to the power source.
To standardize the radiated and conducted emissions behavior of the power line so measurements from different labs can be compared, its impedance is controlled by a Power Line Impedance Stabilization Network (PLISN). This is a combination voltage probe/50-W filter that isolates the power source from the test sample, standardizes the source impedance at 50 W, and permits the measurement receiver to be capacitively coupled to the test-sample power line.
For radiated emissions, the PLISN is terminated in 50 W. Each ungrounded power line requires a PLISN. Unfortunately, the PLISN impedance is not 50 W across the entire measurement frequency range and must be calibrated at the test sample’s equivalent load current. Even then, significant errors can be encountered due to impedance mismatch between the PLISN and the measurement receiver. At this time, most power-line emissions tests use the same type PLISN.
Measurement distances vary a lot. The radiated emissions measurement distance is a good example. The military uses 1 meter, the FCC either 3 or 30 meters, and the EU 10 meters. Even then, all of these distances are better than trying to use the VDE 100-meter test distance.
The arguments for different distances are based on the assumption that these distances represent typical user distances. That may have been how the model was set up and justified, but there are so many exceptions to these distances in real life that just about any standardized test distance could be used.
Distance does affect the measurements. As the distance from an RF radiator increases, the field characteristics change from near field to transition region to far field, with significant variation in the wave impedance. The EMC community tends to neglect the transition region and makes the assumption that the near field ends where the far field begins.
For commercial tests, there also has been some attempt to make measurements in the far field. However, the near-field and far-field characteristics also depend on the size of the measurement antennas as well as the size of the test-sample radiators, and that is not considered during testing.
For test samples whose radiator dimensions are very small compared with the measured wavelength (D << l), the near-field/far-field interface occurs at R = l/2p. As the frequency increases, the test sample radiator dimensions are no longer small compared to the measured wavelength, and for D = l/2, the near-field/far-field interface occurs at R = D2/2l.
Since these effects are not considered by the current specifications, there is nothing unique about the current measurement distances. That being the case, we could pick one that lowers overall testing costs. The cost of a fully anechoic 10-meter measurement site can run into the millions of dollars.
In our quest to understand what is necessary to create a universal EMC test standard, the second part of this article will address measurement errors, antenna characteristics, receiver detector function, and receiver impulse bandwidth.
About the Author
Ronald W. Brewer is an internationally recognized EMC authority named a Distinguished Lecturer by the IEEE EMC Society. The NARTE-certified EMC/ESD engineer has worked full time in the field for more than 30 years, specializing in EMC systems design, integration, and shielding. Mr. Brewer completed undergraduate and graduate work in engineering at the University of Michigan, serves on the IEEE EMC Society Board of Directors, and recently was featured in the CBS History Channel Special “Three Air Crashes: Common Links.” P.O. Box 1221, Leesburg, VA 20177, 703-727-4150, e-mail: [email protected]
Return to EE Home Page
Published by EE-Evaluation Engineering
All contents © 2002 Nelson Publishing Inc.
No reprint, distribution, or reuse in any medium is permitted
without the express written consent of the publisher.