Some EMC tests are about as exciting as watching grass grow. However, I must admit that today's computer control beats the previous procedures that included highly manual testing followed by lengthy and often very tedious rechecking, recording, and analyzing results.
Throughout the years, the electric field (EF) radiated susceptibility/immunity test remains the most fun EMC test to run. It's so easy to produce the required EF at any desired frequency that we can really concentrate on testing the EUT.
Performing the RS Test
Radiated susceptibility (RS) testing in its various forms is mandated by a number of EMC specifications including MIL-STD-461, MIL-STD-464, and the EU. Those of you who have read the article titled “EMC Failures Happen” in the December 2007 issue of EE-Evaluation Engineering know that passing an EMC test won't guarantee that the unit will be immune to EMC problems in its operational environment. But it helps.
RS testing, illustrated in Figure 1, requires the EUT to be illuminated by a low-, medium-, or high-level EF. If it will fit, the EUT is placed in a shielded enclosure with a layout that represents its normal operational configuration as closely as possible. The setup includes placing it on a ground plane made from material representative of the actual platform.
Figure 1. MIL-STD-461D/E/F RS103 Measurements* = EF Display and Sensor Required for RS103
A transmit antenna is placed in front of the EUT's most susceptible RF pickup area at the separation distance prescribed by the test specification, typically 1 meter for military and 3 meters for the EU. The EF is established at the specified frequencies using a signal generator and an RF power amplifier to drive the transmit antenna. It may require a suite of signal generators, power amplifiers, and antennas to cover the entire frequency range.
The test signal is modulated using a frequency and waveform that correspond to worst case. In the event these are unknown, use the modulation called out in the specification.
The frequency range for the test is slowly swept at the prescribed EF level or higher. This can be done by a computer or manually.
Table 1 indicates the MIL-STD-461F susceptibility scan speeds. At frequencies where the EUT is susceptible, the scanning is stopped, the EF is reduced to the susceptibility threshold level, the first level is recorded, and then the EF is reduced by at least 6 dB. Now the EF level is increased until the susceptibility condition reappears. This second level is compared with the first, and the lowest level is the susceptibility. Performing the test in this way avoids the problems of hysteresis in the measurement.
Table 1. Susceptibility Scan Speedsfo = tuned frequency
Most test labs execute susceptibility tests manually because of the randomness of susceptibility occurrences. It's difficult enough to establish the susceptibility criteria. But then these criteria must be monitored and provided as feedback to a computer to tell it to stop when susceptibility occurs, automatically adjust RF levels to 6 dB below the susceptibility point, and perform a retest.
Humans do a much better job of correlating what may appear as random, nonrelated incidents. It's not uncommon for an absolutely unexpected, unplanned susceptibility condition to occur.
RF Test Equipment
The RS test requires equipment to create the EF and monitor it just to ensure that the EF is there. Following is a list of the equipment that would be used to do a MIL-STD-461E/F RS test:
Transmitters to Create the EF
• ??LISNs: Used to standardize the power input RF impedance.
• RF Signal Generators: Any standard generator with modulation that covers the frequency range.
• ??Modulation Generator: Used with the standard generator to provide the required modulation.
• ??Power Amplifier (PA): An RF signal booster used with the standard generator because no generator has enough output to produce the required EF.
• ??Transmit Antennas: Produces the EF. The EF required in conjunction with the gain of the transmit antenna determines the size of the PA.
• ??Directional Coupler: Because of antenna impedance mismatch variations, the directional coupler is required to determine forward power when precalibrating the EF.
• ??Power Meter: Used with the directional coupler.
Receivers to Ensure the EUT Is Exposed to the Correct EF
• ??EF Sensors: 10 kHz to 1 GHz. Used at the EUT to determine incident EF strength. Sensors that cover the 1-GHz to 18-GHz range also are available.
• ??Receive Antennas: Used at the EUT to determine incident EF strength in place of EF sensors; 1-GHz to 10-GHz double ridge horns and 10-GHz to 40-GHz antennas as approved by the procuring activity.
• ??Attenuator: Used to protect the measurement receiver and reduce errors from antenna VSWR.
• ??Measurement Receiver: Used with the receive antennas. One or more may be required to cover a test frequency range; could also be a spectrum analyzer.
• ??Data Recording Device: Connects directly to receiver output or a computer used to control the receiver.
From an EMC perspective, there's nothing unusual about a test setup that uses signal generators, RF power amplifiers, and antennas until it's time to perform such a test. Then try to find that equipment. All the equipment is readily available except for large broadband RF power amplifiers and antennas.
Since the antenna performance determines the amplifier power requirement, it's necessary to know the worst-case antenna gain and how far we need to squirt the RF to size the amplifier. Figure 2 shows how to calculate the power amplifier requirements based on antenna characteristics.
The same set of amplifiers can be used with a wide assortment of transmit antennas. And there is a wide assortment, each with very different characteristics. It would be great if one antenna could be used to generate the RF field across the entire frequency range, but dimensional restrictions and antenna Q limit the maximum antenna bandwidth to about a decade. Table 2 shows the most popular antenna types used in the different frequency ranges.
Table 2. Antenna Types vs. Frequency
Generating high-amplitude EF strengths in the 10-kHz to 30-MHz frequency range is difficult because antenna dimensions are very small with respect to a half wavelength, making the antenna efficiency very poor. As a result, there are some antenna alternatives such as the GTEM cell and parallel plate/triline used when testing smaller EUTs. For larger EUTs, the size of the line limits the usable upper frequency.
There always has been a struggle regarding the sizes of the EUT, the shielded enclosure, and the antennas. To minimize distortion and antenna loading, when an antenna is used in a shielded enclosure, the ends should be kept away from the wall by at least 0.5 meter. For small EUTs, the size of the antenna determines the enclosure size.
In the early days of RFI/EMI/EMC testing, RS tests were performed by feeding the 100,000-??V modulated output of a standard signal generator into a 41-inch monopole, tuned dipole, or horn antenna. E-fields weren't monitored.
MIL-STD-826 (1964), the first attempt at a tri-service standard and the basis of a number of procedures in MIL-STD-461, changed all that. Then, RS field strengths were monitored by antennas placed to the side or behind the transmit antennas.
Now for MIL-STD-461 measurements, we've shrunk the antennas, grouped three together aligned along the X-Y-Z axes, added amplification to make up for their inefficiency, called them EF probes, and placed them on or in close proximity to the EUT. To minimize EF probe susceptibility and field distortion, most utilize fiber-optic interfaces.
The EU EMC tests use an alternative approach in which the EF is precalibrated. Figure 3 shows the EU 16-point EF uniformity requirements.
Figure 3. EU EF Uniformity +6/-0 dB
Defining Susceptibility
We want to determine if the EUT will operate properly in an adverse RF environment. This can be defined by duplication of previously measured RF environmental levels or compliance with an EMC specification. Susceptibility to radiated EM energy primarily is due to RF pickup on wires and cables and generally results in malfunctions or degradation of performance. The latter often can be tolerated, but malfunctions cannot.
The problem of establishing pass/fail criteria for susceptibility is determining how much degradation is tolerable before we conclude that the EUT is not working properly. Beware of any specification that states that the characteristics of the EUT during the susceptibility tests cannot change from those measured in the laboratory sans RF.
Four characteristics greatly influence the susceptibility of the EUT: frequency, amplitude, spatial relationships, and timing (FAST). They often are used as a culling approach to analyze EMC problems.
Frequency
For an RF device, in-band susceptibility generally stems from the culprit frequency or its harmonics coupling at the victim's tuned frequency, harmonic, or IF. Out-of-band or non-RF device susceptibility generally results from the culprit frequency coupling into a circuit through an RF response window created by wire, cable, or parasitic resonances.
Cable resonance is one of the most often occurring problems so failures frequently occur in the 30-MHz to 300-MHz range. Because the response frequencies are unknown, the entire RF spectrum must be scanned during an RS test. Signals must be modulated to determine if the system is susceptible to audio rectification.
Amplitude
The interfering signal adds to the EUT internal noise. If the amplitude of the interfering signal being coupled into the EUT is at the same level as the intended signal, most likely the system will malfunction. Consequently, the amplitude of the interfering source energy level determines the susceptibility.
Spatial Relationships
If the EUT is susceptible along a particular axis, then the field-generating antenna should be aligned with this axis during the test. It's not a big problem when the test is performed in an ordinary metal box enclosure, but it is in an anechoic enclosure.
The EU requirements handle this problem by rotating the EUT. Military tests generally probe the unit for the worse emission locations and assume that will correspond with the worst susceptibility locations.
Timing
Susceptibility occurs only when both the culprit and victim are ON. The ease of determining susceptibility then depends on whether the simultaneous operation of the culprit and the victim make their timing relationship appear random, periodic, or continuous.
It's possible to use FAST analysis to understand what circuits, subsystems, or equipment are likely to respond to the susceptibility signal and under what circumstances this may occur. As an example, the operational transfer function of an analog circuit is continuous. A slight change in the input results in a large change in the output. Accordingly, analog circuits respond to much lower RF susceptibility signals than digital equipment, typically on the order of 40 dB. Once the analog signal is contaminated, there is no way to clean it up.
Digital circuits, on the other hand, require significantly higher susceptibility signal levels before malfunctions occur. These malfunctions generally are the result of a change of state of the logic devices and take place suddenly with a corresponding loss of data. Prior to this point, digital devices will operate properly.
During susceptibility testing, there is a requirement to place the EUT in its most susceptible operating mode. The difference in the behavior of analog and digital circuits really complicates determining what mode is the most susceptible.
The number of modes multiplies the test time accordingly. If the EUT has one mode, the test is performed once. If it has 10 modes, the test is performed 10 times. Fortunately, MIL-STD-461 only requires a sufficient number of modes to ensure that all circuitry is evaluated.
Since real-time monitoring is necessary, tests usually are performed using built-in test equipment (BITE). This procedure is augmented by special test software, custom interface circuits, and the creative use of fiber-optic interconnects, isolation transformers, closed-circuit TV, acoustic couplers, telescopes, shotgun microphones, canary circuits, and any other thing necessary to monitor the operation of the EUT.
Care must be exercised to ensure that these equipment modifications, special circuits, and software do not change the susceptibility of the EUT. Honesty is important. It's generally up to the test director to be sure that the test is being performed properly.
Test Conditions
Typically, there are three distinct areas used for EMC testing: the measurement equipment area, the EUT test environment area, and the exercise-simulation equipment area. The EUT test environment area for RS generally is a shielded enclosure.
The test configuration should isolate these three areas. In the simplest arrangement, the shielded enclosure is placed between the other two areas. In a more extensive setup, three or four shielded enclosures are configured so each area is contained within its own shield (Figure 4).
A number of groups, such as the FAA, the FCC, the IT department, and our neighbors, would be upset if we just started generating RF energy without regard to the environment. As a result, most RS tests are performed in a shielded enclosure.
There are three primary types of shielded enclosures: standard metal boxes, anechoic enclosures, and reverberation chambers. The standard metal box enclosure more closely represents the actual operational environment, especially for military equipment. Most military aircraft, ships, tanks, Humvees, communications shelters, Quonset huts, and field desks are made of metal. The military even uses shielded tents made from nickel-plated nylon.
However, there are problems associated with testing in a standard metal box: enclosure resonance, reflections from the walls, and antenna loading. The biggest problem is the metal box itself, which is a resonant cavity. At resonance, the EF levels can be increased as much as 35 dB to 40 dB. The field distribution is not uniform, and since the Q is very high, it's possible to mistake a field-intensity increase at room resonance as EUT susceptibility. The same is true regarding narrowband emissions from the EUT.
We can fight the problem, or we can work with it. If you want to fight the problem, resonance can be checked and the effects significantly reduced by detuning the cavity. This can be done by opening the door or using portable anechoic panels.
Panels work best. These typically are 8.5 ft high and made from a 4 ft x 8 ft sheet of ?? inch plywood with absorber material on one side and aluminum foil on the other side. These roll-around panels can be used in absorber-enhanced enclosures.
Unlike the EU EMC requirements, MIL-STD-461 does not require a semi- or fully anechoic enclosure, only that RF absorber material be placed above, behind, and on both sides of the EUT from the ground plane to the ceiling with additional absorber located behind the test antenna from the floor to the ceiling. The absorber material must provide at least 6-dB attenuation from 80 MHz to 250 MHz and 10 dB or more above 250 MHz.
The requirement to place the EUT no closer than 30 cm to the absorber material, combined with absorber material that could be 24 to 30 inches thick, means that the material significantly reduces the working volume inside the shielded enclosure. The enclosure working volume is determined by the size of the EUT or the size of the antennas, whichever is larger.
A reverberation chamber works with the enclosure resonances and reflections. It has a useable volume approximately 50% of the total and can be smaller than an absorber-lined enclosure. A rotating metal reflector sweeps the maximum field strength produced by resonance and reflections throughout the enclosure to assure that the EUT has been adequately illuminated by the EF.
A shielded enclosure is a test equipment item. It has or should have a model number and a serial number. It is not a room. It is a large metal enclosure with a power line and other types of filters that contains or excludes RF energy. It should be placed on a calibration schedule just like all other test equipment.
In the case of an enclosure, make sure that the internal ambient is 6 dB or more below the specification limit. Data from the calibration is kept in a file. Lastly, unless the test sample is huge or it must do something strange during its operation, open area test sites (OATS) normally are not used for performing RS tests.
RF Safety/RADHAZ
MIL-STD-461D/E/F calls out 200 V/m. With 20-dB to 40-dB resonant gain inside an enclosure, the E-field strength in a shielded enclosure where a 200-V/m field is being generated could be 2,000 V/m to 20,000 V/m at some locations within the enclosure.
Beware that high-level RF fields are hazardous to your health. There is plenty of information that discusses how exposure of this type can cause the formation of cataracts or thermal tissue damage.
The FCC OET Bulletin 65 covers RF hazards and provides the equations for calculating how long you can be exposed before a hazard exists. These equations are based on thermal heating, but they do consider antenna gain, modulation type, and antenna separation.
Not a lot of information cites actual health hazards for lower-level RF fields because of a lack of concrete proof that such dangers exist. The premise is that observable biological effects do not necessarily mean that there is a biological hazard. But of course, it doesn't rule out that possibility either.
This disagreement about what constitutes an RF hazard level is reflected in the differences between the Russian RADHAZ and U.S. safety levels. The Russian levels are based on the field strength at which there are observable effects; the U.S. levels are based on the field strength where thermal damage occurs. Consequently, the Russian maximum field-strength levels are much, much lower than in the United States.
If observable biological effects are occurring, then there is likely to be an RF hazard, even if we haven't yet determined what the hazard is. Accordingly, precautions should be taken. Call it prudent avoidance, but why take the chance?
Measurement Error and Uncertainty
During an RS test, an attempt is being made to control the RF environment. But testing problems are compounded because of variations in the characteristics of the EUT, measurement equipment, and test setup/facility. These variations can result in large differences in EMC measured data, often as much as 40 dB.
It is not possible to make measurements without the measurement process/equipment disturbing the data. Some of the errors will be random, others systematic. If the errors are reasonably independent, a calculation of the total uncertainty can be made by combining their standard deviations.
This uncertainty value will tell how much potential error there is in the measurement. However, it does not tell us anything at all about the probability of susceptibility capture. That is, what is the probability that a complex system will be in an internal state where it is susceptible concurrent with the presence of a signal with frequency and modulation characteristics that would cause susceptibility?
If you are interested in learning more about the statistical nature of EMC measurement uncertainty, check out IEC CISPR 16-4. It deals only with the random and systematic errors associated with making measurements on a steady-state device to an EMC standard. It does not address how well the standard's test environment mimics the actual operational environment.
About the Author
Ron Brewer currently is a senior EMC/RF engineering analyst with Analex at the NASA Kennedy Space Center. The NARTE-certified EMC/ESD engineer has worked full-time in the EMC field for more than 30 years. Mr. Brewer was named Distinguished Lecturer by the IEEE EMC Society and has taught more than 385 EMC technical short-courses in 29 countries and published numerous papers on EMC/ESD and shielding design. He completed undergraduate and graduate work in engineering science and physics at the University of Michigan. e-mail: [email protected]
FOR MORE INFORMATION
on FCC OET Bulletin 65
Click here
on RF/ELF handbook for health professionals
Click here