Comparing Fully Anechoic Chambers to Semianechoic Limits

Greg Kiemel

A series of tests was conducted to attempt to validate the use of a fully anechoic chamber for certification to existing semianechoic limits.

Worldwide regulatory requirements specify the use of open-area test sites (OATS) or semianechoic chambers (SACs) for the measurement of radiated emissions from unintentional radiators (see sidebar for definitions of key terms).

The intent of any emissions specification is to prevent unwanted interference to radio reception. This includes preventing interference to such services as broadcast radio and TV, as well as amateur radio and emergency services. The emissions specification limits were derived from comprehensive studies that determined the interference threshold to TV and radio receivers versus the proximity of potential noise sources. These studies, which covered a large geographical area, amassed a very large body of data. Based on the results, statistically significant specification limits were derived.^1–3

In every study, the measured values were made in an environment where the reflected wave from the earth had a major effect on the measured value. It was concluded that a uniform ground plane would provide a controlled environment that not only simulated real-world conditions but also greatly enhanced the repeatability of the measurements. As a result, international standards-making bodies developed measurement procedures based on the use of test sites with a reflecting ground plane. Regulatory agencies then adopted these procedures and set specification limits for emissions measurements made at facilities using a reflecting ground plane.⁴

Since then, thousands of measurement facilities that use a reflecting ground plane to make radiated emissions measurements have been installed successfully worldwide. When devices are found compliant to the applicable limits at these facilities, the potential for interference is very small. Virtually all interference complaints are a result of a noncompliant product—not deficient measurement standards or facilities.

So why is the effort being made to improve a methodology that has a proven track record? In recent years, the number of wireless devices and their associated services has greatly increased. The resultant ambient radio-frequency interference (RFI) has also increased at most OATS facilities. Measurements in high-ambient environments are more difficult and can affect the accuracy of the results. Currently, the only alternatives to OATS facilities are large, expensive, semianechoic chambers that permit antenna scan heights of 1–4 m and meet the site attenuation requirements of ANSI C63.4.

An effort is under way to develop alternative test facilities that provide measurement accuracy at a lower cost than semianechoic chambers. One such effort is detailed in prEN 50147-3:1998.⁵ This preliminary standard proposes the use of fully anechoic chambers (FACs) for certification measurements. Because the measurement antenna is kept at a fixed height, the hope is that chambers can be built smaller, and thus make this option more affordable.

The Comparison

A series of tests was performed in an effort to validate the proposed use of a fully anechoic chamber for certification measurements to semianechoic limits. This article examines the results of those tests. Data were gathered on two different noise sources. The first was a PC system (see Figure 1). The second was a battery-powered 30-MHz comb generator with an omnidirectional antenna from York EMC Services Ltd. (Castleford, UK) (see Figures 2 and 3). Both systems were tested in a large, FCC-listed, 3-m semianechoic chamber that meets the volumetric site attenuation requirements of ANSI C63.4-1992. Figures 4 and 5 show the respective horizontal and vertical polarities for a PC system.

Figure 1. PC system.

Figure 2. Comb generator with omnidirectional antenna—test setup for vertical polarity.

Figure 3. Comb generator with omnidirectional antenna—test setup for horizontal polarity.

Figure 4. Horizontal polarity for a PC system.

Figure 5. Vertical polarity for a PC system.

First, the chamber was set up in the semianechoic configuration prescribed by regulatory agencies around the world, with a reflecting ground plane and an antenna that is varied in height from 1 to 4 m (see Figure 6 for semianechoic configuration). Then, without touching or moving the PC system or noise source, the chamber was converted to fully anechoic by adding ferrite tiles with matching absorber to the floor (see Figure 7). The fully anechoic configuration met the proposed requirements detailed in prEN 50147-3:1998, including the use of free-space antenna factors to calculate field strength. There were some distinct advantages to performing the tests in this manner:

Figure 6. Semianechoic chamber (SAC) configuration.

Figure 7. Fully anechoic chamber (FAC) configuration.

An examination of the data shows that a fully anechoic chamber is not a reliable predictor of semianechoic test results. For both noise sources, and for both antenna polarities, the fully anechoic configuration measured too low at almost every frequency. The worst-case error was 7.3 dB. In a real-world application, this could mean that a product that passed in a fully anechoic chamber by 4 dB, could fail the certification test at a FCC-listed facility by more than 3 dB.

Isotropic Source

Prior to the study, it was suspected that measurements made on an isotropic source might correlate better than a PC system. Real-world unintentional radiators, such as the PC system, have radiating lobes in all three orthogonal axes. FCC-listed facilities scan the antenna height from 1 to 4 m and are therefore better able to detect the maximum field strength of these emissions. Surprisingly, the performance of the fully anechoic configuration was equally poor for both the PC system and the isotropic source. Figures 8 and 9 show the respective horizontal and vertical polarities for the isotropic noise source.

Figure 9. Vertical polarity for the isotropic noise source.

Figure 8. Horizontal polarity for the isotropic noise source.

Annex 2 of prEN 50147-3:1998 suggests new free-space specification limits based on CISPR 11 and CISPR 22 limits. The proposed limits imply that measurements in a fully anechoic chamber should be compared with a limit that is 5 dB more stringent than the current CISPR 11 and CISPR 22 limits used at semianechoic facilities. This is consistent with the measurement data in this study. The addition of 5 dB to the FAC data would improve correlation with the SAC data. However, depending on the equipment under test (EUT), it is important to note that a simple 5-dB offset will not always provide acceptable correlation.

Annex 2 of prEN 50147-3:1998 cites a study in which the comparison between an FAC and an OATS resulted in differences between 0.1 and 11.6 dB. The high cost of an FCC-listed semianechoic chamber has been one of the major arguments for fully anechoic chambers. In the past, it has been suggested that FACs could be built smaller and more cheaply than SACs because FACs use a fixed antenna height. However, prEN 50147-3:1998 specifies facilities whose size, performance, and attending cost rival that of semianechoic chambers.

To test an EUT with a maximum diameter of 1.2 m, a 3-m test distance is required. A maximum EUT diameter of 2 m requires a 5-m test distance, and a 4-m EUT diameter is tested at 10 m. In addition, the normalized site attenuation is difficult to achieve without the use of an excellent radio-frequency (RF) absorber and larger chamber dimensions.⁶

So far, no compelling evidence has been offered by proponents of fully anechoic chambers to prove that a free-space environment simulates real-world conditions. Regulatory agencies have not yet adopted specifications that allow the use of fully anechoic chambers for compliance testing. Therefore, until new free-space specification limits and methods are widely adopted, fully anechoic chambers should not be used for certification measurements or as a precompliance tool. However, should a free-space model become accepted by regulatory agencies, large semianechoic chambers could be easily converted to fully anechoic chambers.

Key Definitions Fully Anechoic Chamber (FAC). A shielded enclosure like the SAC (see below), except the ground plane is also RF-absorbing material. The intent is to simulate a free-space environment. The antenna height is fixed—usually centered on the test volume at a height greater than 1.5 m. Open-Area Test Site (OATS). A large, obstruction-free, outdoor area with a flat metallic plane covering the ground. By design, there are no objects above the ground plane that reflect radio-frequency (RF) energy. The equipment under test is placed 0.8 m above the ground plane, and an antenna is placed 3 or 10 m away. For each emission, the height of the antenna is varied from 1 to 4 m to measure the maximum field strength. The measured field strength is the vector sum of the incident electromagnetic wave and the reflected wave. Semianechoic Chamber (SAC). A shielded enclosure whose internal walls and ceiling are lined with material that absorbs electromagnetic energy in the frequency range of interest. In semianechoic chambers listed with FCC for radiated emissions testing, the RF absorber is effective from 30 MHz to 1 GHz (and typically up to 18 GHz). A sufficiently large semianechoic chamber with this type of absorber is intended to simulate an OATS with the added benefit of no ambient RF interference. The antenna height is varied from 1 to 4 m. Unintentional Radiator. A device that intentionally generates RF energy for use within the device but is not intended to emit electromagnetic. For example, most digital devices such as personal computers are considered to be unintentional radiators.

References

1. RE Boyd, JA Malack, and IE Rosenbarker, "EMI Control for Data Processing and Office Equipment," in Proceedings of the First Symposium and Technical Exhibition on Electromagnetic Compatibility (Montreux, Switzerland, 1975).

2. Limits and Methods of Measurement of Electromagnetic Emanations from Electronic Data Processing and Office Equipment, (Washington, DC: Computer and Business Equipment Manufacturer's Association (CBEMA), Environment and Safety Committee Five [ECS5], 1977).

3. RF German and R Calcavecchio, "On Radiated EMI Measurement in the VHF/UHF Frequency Range," in Proceedings of the IEEE International Symposium on Electromagnetic Compatibility (Baltimore, MD: IEEE EMC Society, 1980), 91–97.

4. ANSI C63.4-1992, "American National Standard for Methods of Measurement of Radio-Noise Emissions from Low-Voltage Electrical and Electronic Equipment in the Range of 9 kHz to 40 GHz," American National Standards Institute, New York.

5. prEN 50147-3:1998, "Electromagnetic Compatibility Basic Emission Standard, Part 3: Emission Measurements in Fully Anechoic Rooms", TC210-WG4-9905, CENELEC, Brussels, January 1999.

6. MAK Wiles, W Müllner, "Conversion of Semi to Fully Anechoic Rooms per CENELEC prEN50147-3," in Proceedings of IEEE International Symposium on Electromagnetic Compatibility (Montreal, QC, Canada: IEEE EMC Society, 2001), 268–273.

Greg Kiemel is director of engineering for Northwest EMC, Inc. He can be reached at gkiemel@nwemc.com.