It is very difficult to identify one or two key components used in radiated immunity testing as being the most important. The broadband E-field sensor most certainly is a key component because we rely on it for an accurate indication of the generated E-field. Also, the anechoic room makes it possible to have an enforceable immunity standard, and the software enables efficient testing while ensuring thoroughness.
While equally important, antennas and amplifiers too often are taken for granted. But when selected incorrectly, singly or in combination, they can be the source of serious errors. Worse yet, you may not know when a problem exists.
Because of the sound engineering effort that went into its production, ENV61000-4-3, formerly IEC 1000-4-3, is an excellent immunity test reference. Consider the positive changes that now are in place and which should serve as a solid reference for other standards:
The transmit distance has increased from 1 m to 3 m. This ensures more uniform exposure and reduces proximity effects between the antenna and the equipment under test (EUT).
The performance specification for the anechoic room now ensures reasonable site-to-site correlation. Before the anechoic-room requirement, the test results in unlined (bare metal) rooms were chaotic at best. Room resonance and reflections made it impossible to ensure repeatable test results, not only from site to site but within the same room. Also, the maximum field-level variation at each test frequency must not exceed 6 dB.
The standard requires that the field strength throughout the defined test area does not fall below the required test level. If the required test level is 10 V/m, the field variations in the defined test area will range from 10 V/m to 20 V/m if the field variations are 6 dB.
Amplitude modulation is required to better simulate ambient conditions. While questionable in the way it is specified, the reason for requiring amplitude modulation is sound. Questionable because the amplitude modulation is added to the unmodulated carrier frequency. This increases the amplifier power requirement by 5.1 dB. There also is no requirement in the standard to ensure that the power amplifier accommodates the modulated carrier without suffering severe amplitude compression.
In light of these changes for the good, we now can better assess the effects of the antennas and power amplifier on a radiated immunity test.
The Power Amplifier
Some look upon the power amplifier as a brute-force component that only needs to deliver raw power without shutting down or oscillating. More experienced users, however, recognize that the amplifier must be linear to preserve the reasonable fidelity of the amplitude-modulated waveforms that IEC testing requires.
Those who must test to the Bellcore specification know that reasonably low harmonics also are important. The Bellcore specification calls for harmonics of <20 dB.
At the other extreme, those relatively new to EMC testing may attempt to impose impractical and unnecessary requirements on the power amplifier—requirements such as low noise figure and low intermodulation distortion. While such requirements may be important for preamplifiers, they are unnecessary for high-power amplifiers operating well above the amplifier’s noise level. As for intermodulation distortion, this is important for multichannel applications such as cell-phone base stations and cable TV amplifiers.
EMC power amplifiers only need to process single-frequency waveforms that are either modulated or unmodulated. Also, the modulations most often required for EMC testing are relatively simple and require an insignificant increase in bandwidth.
Amplifier Specifications
Most power amplifiers are specified for linear and saturated power. Because amplitude modulation is specified for IEC 1000-4-3 testing, the saturated power rating of the amplifier is of no value. To ensure reasonably good fidelity of the intended test waveform, the amplifier must operate within its linear range.
Linear operation is the point where the amplifier is compressed by no more than 1 dB. Driving the amplifier beyond this point results in rapidly increasing compression and high harmonic power. In combination with the antenna performance, high harmonic power can result in test fields that have significant or even dominant harmonic E-field levels.
The Antenna
While apparently the simplest of all the required test equipment for radiated immunity testing, the antenna deserves serious consideration and caution. With the transmit distance being moved from 1 m to 3 m, antenna gain takes on greater importance. Neglecting the effects of moving from the near field to the far field, the theoretical increase in power will be nine times (power varies as the square of the transmit distance). Equally important are the exposure angle on the EUT and the need to ensure that the defined test area, a 1.5-m2 vertical plane, is illuminated adequately.
Uniform EUT Illumination
With the use of biconical, log periodic, and biconical/log periodic-type antennas, the EUT illumination is reasonably uniform across the 1.5-m plane at a transmit distance of 3 m. Most of the field variations encountered in the typical anechoic room with these antennas are the result of small reflections from the imperfect absorber-lined room surfaces (Figure 1).
Of a more serious consideration for antennas is the gain vs frequency. We can examine three common antenna types to illustrate the importance: the log periodic, the biconical, or biconical/log periodic types, and horn antennas (standard and ridged types).
Log Periodic Antennas
The log periodic antenna could be considered the perfect antenna for modern radiated immunity testing because it exhibits virtually constant gain and beam width vs frequency. The only drawback with this antenna type is the radiating phase center, which moves with frequency along the length of the antenna.
The highest frequencies radiate from the smallest elements at the front of the antenna and the lowest frequencies from the larger back elements. Depending on the frequency range and lowest frequency of the log periodic antenna, the effective transmit distance may be as much as 4 m or more.
The effective beam width and gain of the typical EMC log periodic antenna are determined by three of its dipole elements at any one transmit frequency. However, we often stretch the definition of log periodic performance beyond the normal design limits. In such cases, the lowest specified or used frequency may only employ the last two dipole elements of the antenna. This, in addition to the increased transmit distance (due to the location of radiating phase center), will further degrade the effective E-field producing efficiency.
Biconical or Biconical/Log Periodic-Type Antennas
Since the log periodic antenna already has been covered, that portion of the biconical/log periodic antenna will not be addressed here. Of greater importance is an understanding of the basic biconical antenna and the bow-tie portion of the biconical/log periodic antenna (collectively referred to as a biconical antenna).
Both of these antenna types are specified for operation from approximately 25 MHz to 300 MHz; however, most users limit its upper frequency use to 200 MHz. With many log periodic antennas starting at 200 MHz, it makes good sense to take advantage of the higher antenna gain.
The operation of the biconical antenna from approximately 80 MHz to 200 MHz is reasonably flat. Typical gain variations are -1 dBi to +1 dBi. Gain in dBi is referenced to a theoretical isotropic, point-source radiator. On the other hand, the operation of the biconical antenna down to 25 MHz is not flat, and this is where serious problems arise. While testing below 80 MHz is not required by the ENV 61000-4-3, it is commonly performed to comply with the medical directive and Bellcore test requirements.
One thing is immediately apparent when we inspect the manufacturer specifications for the biconical-type antenna. We find best-case gain figures of approximately -14 dBi at 25 MHz. When we consider the significance of this low gain figure, we see that at least 25 times more power is required at 26 MHz than at 80 MHz.
Hypothetically, if a 100-W amplifier were adequate at 80 MHz for a 10 V/m test, then 2,500 W would be needed at 26 MHz for the same test specification. In practice, this is even worse because the effective antenna gain in a typical compact room at 26 MHz is even lower (Figure 2).
The gain for most EMC antennas is derived from antenna-factor measurements made on an outdoor antenna range. This is acceptable for emission testing on an open-area test site. With the limited size of the compact anechoic room and the fact that they are below cutoff at 26 MHz, the effective antenna gain in the room is significantly lower than the published gain figure (the dashed curve in Figure 2).
Since the power required for testing at 26 MHz is unrealistically high, we must conclude that radiated immunity testing at 26 MHz is impractical at best. How is it that we seem to carry out such testing at 26 MHz with reasonable power levels?
Combination of Power Amplifier and Antenna
We now must consider what we know about the typical power amplifiers and the typical EMC antennas. First, the power amplifiers have the potential to generate significant harmonic levels when they are overdriven. Secondly, certain antennas, such as the biconical types from 26 to 80 MHz, have significant gain slopes (increasing gain with increasing frequency).
Combining the amplifier and antenna, we have the potential to attenuate the desired frequency, 26 MHz, and enhance the harmonic power by virtue of the antenna gain characteristics and the anechoic room performance. In fact, most of the radiated testing claimed to have been performed at 26 MHz may have been, in all likelihood, performed at the harmonic frequencies (Figure 2).
While we addressed one of the worst and very common testing errors due to the combination of power amplifier and antenna, we must remember that similar problems can occur elsewhere in the test frequency range. The harmonic levels may not dominate, as in this example, but they could become a significant part of the indicated field strength.
Horn Antennas
Horn antennas are not recommended for IEC testing below 1 GHz. While horn antennas have higher gain and appear to require less power, they also have narrower beam widths. Narrow antenna beam widths will contribute to the field variations that are already present due to the imperfect anechoic room performance. The combination of the narrower beam horn antenna and the room-generated field gradients across the test area may make the room effectively noncompliant.
You could argue that a narrow antenna beam will reduce the illumination of the walls, floor, and ceiling of the room. In this hypothesized case, there is little contribution from the imperfect anechoic room. However, in this scenario, we also are assuming that the peak of the radiation pattern will be the center of the defined test area. If this is the case, every decibel in roll-off due to the narrow antenna pattern will have to be made up with increased transmit power.
For example, if the antenna pattern is 6 dB down at the edges of the test area, 6 dB or four times more power is needed to comply with the standard. The field in the defined test then will be 10 V/m to 40 V/m. Remember that 3 dB more power is required when only room reflections contribute to the field variations.
Testing Above 1 GHz
The soon-to-be-requirement for testing above 1 GHz is the frequency range where horn antennas are most commonly used. This raises some interesting questions: What are the realistic options above 1 GHz to ensure compliant testing? What about the use of traveling wave-tube amplifiers above 1 GHz? What impact will they have on a reliable and efficient test?
About the Author
Timothy D’Arcangelis is the EMC manager at Antenna Research Associates. His career in RF and microwave systems includes 26 years in developing high-power broadband amplifiers, E-field sensors, antennas, TEM cells, multichannel controllers, and software. He has held several EMC-related positions, from product engineer to president. Antenna Research Associates, 11317 Frederick Ave., Beltsville, MD 20705, (301) 937-8888.
Copyright 1998 Nelson Publishing Inc.
June 1998