In this issue you will find two practical papers that should be of interest to the EMC community. The first paper is entitled, "Electromagnetic Attenuation with Stainless-Steel Fiber Mesh" and is by several authors from the medical community in Japan. They were specifically concerned about the shielding properties of flexible "metal mesh" to waves at different polarizations at frequencies higher than 1 GHz. I think that you will find the results presented here very interesting. The second, "Time and Frequency Domain Analysis of EMC Test Facilities" is about a methodology to determine the quality of an EMC test facility using equipment generally available to RF testing services and has been written by several members of the NIST staff. As one reviewer said, "this is not new, but it is well worth publicizing among EMC engineers." The purpose of this section is to disseminate practical information to the EMC community. In some cases the material is entirely original. In others, the material is not new but has been made either more understandable or accessible to the community. In others, the material has been previously presented at a conference but has been deemed especially worthy of wider dissemination. Readers wishing to share such information with colleagues in the EMC community are encouraged to submit papers or application notes for this section of the Newsletter. Click here for my e-mail. While all material will be reviewed prior to acceptance, the criteria are different from those of Transactions papers. Specifically, while it is not necessary that the paper be archival, it is necessary that the paper be useful and of interest to readers of the Newsletter. Comments from readers concerning these papers are welcome, either as a letter (or e-mail) to the Associate Editor or directly to the authors.

Electromagnetic Attenuation with Stainless-steel Fiber Mesh
Eisuke Hanada¹), Kyoko Takano²), Yasuaki Antoku³), Kouji Matsumura³), Yasushi Hoshino⁴), Toshiya Nishimura⁴), Kennichi Hyoudou⁴), Yoshiaki Watanabe⁵), Yoshiaki Nose²)
1) Department of Medical Informatics, Shimane Medical University School of Medicine
2) Department of Medical Information Science, Graduate School of Medical Sciences, Kyushu University
3) Department of Medical Information Science, Kyushu University Hospital
4) Department of Electromagnetic Wave Shielding, Clean Shield Division, Nippon Sheet Glass Environmental Amenity Co., Ltd.
5) Department of Information Science, Faculty of Science and Engineering, Saga University
Corresponding address:
Eisuke Hanada
Department of Medical Informatics,
Shimane Medical University School of Medicine
Izumo, 693-8501, Japan
E-mail: e-hanada@shimane-med.ac.jp

Abstract
Shielding from radio waves to prevent electromagnetic interference using "metal mesh," mesh made of metal fibers, is well documented. However, in shielding material catalogs, only attenuation against radio waves at frequencies less than 1.0 GHz is mentioned. In addition, only attenuation against waves from one polarization is mentioned, even if the mesh windows are not square or the metal content of the vertical and horizontal fiber differs.
Therefore, we measured the attenuation of metal fiber mesh, made by twisting short stainless-steel threads, of various shapes and sizes at a wide range of frequencies with two polarizations and two measurement methods. To test radio waves at frequencies of 1.0 GHz or lower, we adopted the method developed by the Kansai Electronic Industry Development Center (KEC). The attenuation against radio waves at frequencies over 300 MHz was measured by the loss insertion method using an electromagnetic anechoic chamber. Attenuation differed by method, even in the same frequency range. Attenuations also differed with changes in polarization when the mesh windows were rectangular. It was also confirmed that the length of the minimum composition unit of the metal fiber influenced attenuation.
Keywords: Mesh, metal fiber, electromagnetic shielding, KEC method, loss insertion method

I. Introduction
The use of mobile telephones in hospitals is often restricted because of electromagnetic interference (EMI) with medical electric equipment by radio waves transmitted from these devices. A good prevention method for combating EMI is electromagnetic shielding. Shielding from radio waves is possible using materials containing metal fibers. Mesh shielding materials, in which the fibers contain metal or that are made totally of metal (hereafter, metal mesh), are marketed and widely available. Metal mesh is flexible and lightweight. In addition, it has higher optical permeability than woven and non-woven fabrics made with metal fiber. However, in shielding material catalogs, only attenuation against radio waves at frequencies less than 1.0 GHz is mentioned. In addition, only attenuation against waves from one polarization is mentioned, even though metal mesh windows are not always square. Radio waves at various frequencies and with various polarizations are transmitted in urban areas [1]. The fundamental characteristics of electromagnetic shielding with a metal grid have been reported [2-5], but the shielding characteristics of mesh made of twisted metal threads are unknown. We measured the effect of a wide frequency band on the attenuation of metal mesh of various shapes and sizes.

II. Methods
1) Metal Mesh
The fibers used were made by twisting short stainless steel 32/2 strings. The characteristics of the fiber are shown in Table 1. The metal mesh was knit using these fibers to eliminate problems with fiber constitution. The mesh structure and sizes are shown in Figure 1. The mesh structures were square and rectangular. For square meshes, the intervals between the centers of warps (vertical pitches, "a" in Figure 1) and the centers of woofs (horizontal pitches, "b" in Figure 1) were 3 mm, 5 mm, and 7 mm. For rectangular mesh, the warp and woof pitches were 3 mm and 5 mm. Hereafter, any mesh for which the horizontal pitch was 3 mm and the vertical pitch was 5 mm is called "3 x 5 " mesh. However, the woof of the mesh tested was made with three parallel fibers and the warp had two threads. Therefore, although the mesh was square, the size of the window ( "c " and "d " in Figure 1) was not. It was rectangular, with a length to width ratio of about 5:7. The size of the window of each mesh is shown in Table 1.


2) Measuring Methods and Frequencies Investigated
The total frequency range of the measurement was from 10 MHz to 10 GHz. The attenuation against radio waves at frequencies over 300 MHz was measured by the loss insertion method [6] using an electromagnetic anechoic chamber. Against radio waves at frequencies of 1.0 GHz or lower, attenuation was measured by a method developed by the Kansai Electronic Industry Development Center, Osaka, Japan (hereafter called the "KEC" method) [7,8]. For frequencies from 300 MHz to 1.0 GHz, attenuation was measured by both methods.

a) Measurement by the KEC Method
The KEC method has often been used for measuring the attenuation of electromagnetic shielding materials in Japan because this method does not require an anechoic chamber, which is too expensive for most researchers. With the KEC method, measurement is performed in a small shielded box to eliminate influence by radio waves other than the desired object. A block diagram of this system and details of the shield box used in this method are shown in Figure 2, and the equipment used is shown in Table 2. Each mesh was cut into a 230 mm square and inserted into the measurement equipment.
The attenuation of electric field intensity (Ae) was calculated as the difference between the receiving levels when a sample was set (E1) or not set (E2) in the chamber. The equation (1) is as follows [8,9].

(1)

The transmitted radio wave frequencies were every 10 MHz in the range from 10 MHz to 100 MHz and every 100 MHz from 100 MHz to 1.0 GHz. Radio waves of 15 MHz and 150 MHz were also transmitted. A 13 dBm sin wave was transmitted at right angles to the metal mesh shielding material sample at all frequencies.


b) Measurement by the Loss Insertion Method
The loss insertion method differs from the KEC method in that it allows not only large pieces of material to be measured, but also allows the antennas to be set freely over a range of distances. In this study, by use of the loss insertion method it was possible to imitate the walls and windows on which shielding material would actually be installed [10].
A block diagram of the measurement system used in the loss insertion method is shown in Figure 3. Shielding material (mesh) was fixed, using metal plates, to an open 560 mm square window on the door of the anechoic chamber. A wave generator was set on the outside of the anechoic chamber and connected to the transmitting antenna. A double-ridged guide antenna was used for frequencies of 1.0 GHz or higher, and a log-periodical antenna was used for frequencies less than 1.0 GHz. The receiving antenna, an antenna of the same style as the transmitting antenna, was placed in the anechoic chamber and connected to a spectrum analyzer. Equipment used in the measurement is shown in Table 2. The antennas were adjusted so that the transmitted waves were radiated to the mesh at right angles.
The attenuation was calculated as the difference of the receiving levels (attenuation of electric field intensities) between when a sample was set (E1) or not set (E2) in the window of the anechoic chamber. The equation for attenuation (Ae) calculation was the same as for equation (1).
The transmitted radio wave frequencies were every 100 MHz in the range from 300 MHz to 1.0 GHz, and every 1.0 GHz from 1.0 GHz to 10 GHz. In addition, radio waves of the following frequencies were also transmitted: 1.5 GHz, which is used in all Japanese cellular phone systems, 1.9 GHz as used in Japanese personal handy-phone systems [11], and 2.45 GHz as used in wireless LANs and in microwave ovens. At each frequency, a sin wave was transmitted at 13 dBm output power.

Figure 1 Structure of Stainless Mesh

III. Results
The measurement results by the KEC method are shown in Figure 4.1. When the mesh window was square, the attenuation increased or stayed almost the same in response to radio waves at frequencies up to 150 MHz for 3 mm pitch, up to 300 MHz for 5 mm pitch, and up to 500 MHz for 7 mm pitch. (Hereafter, the frequency at which the maximum attenuation was found is referred to as the "peak frequency".) When the mesh was rectangular, the peak frequency of the "5 x 3" mesh was 30 MHz and that of the "3 x 5" mesh was 150 MHz. When the transmitted radio wave exceeded these frequencies, attenuation decreased at a rate of 20 dB with each ten-fold frequency increase, except in the range from 700 MHz to 1.0 GHz.
The measurement result by the loss insertion method is shown in Figure 4.2. Against radio waves at frequencies of 300 MHz and higher, the decrease of attenuation was almost linear to the log scale of frequency. In addition, both measurements showed attenuation decreases at frequencies from 700 MHz to 1.0 GHz that did not follow the pattern found in the other frequency ranges. The decrease in the attenuation effect from 1.0 GHz to 1.9 GHz also did not follow the pattern found in the other frequency ranges.
Measurement of radio waves in the frequency range from 300 MHz to 1.0 GHz was done with both the KEC method and the loss insertion method. Attenuation by the KEC method was lower than the attenuation by the loss insertion method at each frequency. The difference in attenuation between the two methods at the same frequency differs with the frequency and the pitch of the mesh.

Figure 2 Structure of the KEC Method, Overview of the shield box and equipment used in the KEC method, Block diagram of equipment used with the KEC method, Subject set on the lower half of the shield box, Structure of the shield box.

IV. Discussion

Figure 3 Block Diagram of Equipment Used with the Loss Insertion Method.

1) Attenuation by Metal Mesh
The attenuation was almost the same when the frequencies were lower than the peak frequency, as shown in Figure 4.1. Because the subject meshes were made of fibers consisting of only one material and were of the same constitution, the quantity of the electromagnetic wave (the amount of energy) that could be absorbed by one metal thread did not vary with the pitch of the mesh, but was always the same. As the frequency of the radiated radio wave becomes higher, the amount of absorbable electromagnetic energy of the thread-like metal becomes smaller [12]. Therefore, when the frequency of the radiated radio wave becomes higher, the attenuation becomes lower. On the contrary, since the energy that can be absorbed increases as the number of fibers increases, the attenuation per unit area will be higher as the pitch of the mesh becomes smaller (there are more threads per unit). As the frequency of the radiated electric wave becomes lower, the metal fiber seems to become saturated, and the attenuation at lower frequencies will not increase. Therefore, mesh with larger pitches probably became saturated at higher frequencies. Although this phenomenon has been hypothesized, no proven reports were found.
In our measurements, the decrease of attenuation at two frequency bands near 800 MHz and near 2 GHz was not fixed and the value was higher than the expected attenuation. One of the reasons could be that each string took on the function of an antenna. When the length of a staple is half or a quarter the wavelength of the radiated radio waves, the staple acts as an antenna that shows the highest receiving characteristic.

Figure 4 Electric Field Intensity Attenuation of Mesh.

2) Differences of Attenuation Between the Two Measurement Methods
In the frequency range from 300 MHz to 1.0 GHz, the results by the two measurement methods were different. One possible cause of this difference is that measurement by the KEC method is always in the near region and in the far region by the loss insertion method. When measuring with the KEC method, radiation from the mesh, or reflection between the sample and spectrum analyzer, seems to have occurred. Therefore, rather than determining the electromagnetic wave that actually passed, the possibility exists that the receiving antenna received stronger electric field intensity.

3) Polarization in the Attenuation of Mesh Used as a Shielding Material
Both the KEC and the loss insertion methods showed differences in attenuation at different polarization when mesh windows were not square. This indicates that if the mesh is not square or the metal content of the warp and of woof differs, catalogs should list the attenuation effect of mesh-like shielding material against waves from at least two polarizations, horizontal and vertical.

V. Conclusion
We surveyed the attenuation of metal mesh of various shapes and sizes at a wide range of frequencies at two polarizations using two measurement methods. We found that the length of the minimum composition unit of a twisted metal fiber influences attenuation. We also found that when evaluating the attenuation of commercial mesh shielding material, differences in attenuation at different polarizations should be verified.

Acknowledgement
The authors wish to heartily thank Meiko Trading, Inc. for providing us the metal mesh materials for this study. This study was supported by grants-in-aid from the Japan Society for the Promotion of Science (No.14370771).

References
1. E. Hanada, K. Kodama, K. Takano, Y. Watanabe and Y. Nose, "Possible Electromagnetic Interference with Electronic Medical Equipment by Radio Waves Coming from Outside the Hospital," Journal of Medical Systems vol. 25 No.4, pp.257-267, 2001
2. T. Larsen, "A Survey of the Theory of Wire Grids," IRE Trans. MTT, vol. 10, pp.191-201, 1962
3. G.A. Otteni, "Plane Wave Reflection from a Rectangular Mesh Ground Screen," IEEE Trans. AP, vol. 21, no. 6, pp.843-851, 1973.
4. K.F. Casey, "Electromagnetic Shielding Behavior of Wire-Mesh Screens," IEEE Trans. EMC, vol. 30 No.3, pp.298-306, 1988.
5. S. Chrisatine and A. Orlandi, "An Equivalent Transmission Line Model for Electromagnetic Penetration Through Reinforced Concrete Walls," IEICE Trans. Commun., vol. E78-B, no. 2, pp. 218-229, 1995.
6. Department of Defense (U.S.A.), "MIL-STD-285 Military Standard Attenuation Measurements for Enclosures, Electromagnetic Shielding, for Electronic Test Purposes, Method of," 1956 June.
7. E. Hariya, and M. Umano, "Instruments for Measuring the Electromagnetic Shielding Effectiveness," 1984 IEEE International Symposium on Electromagnetic Compatibility, 2, 1984, Tokyo, Japan.
8. E. Hariya, M. Umano, N. Morinaga and Y. Nagasawa. "Experimental Investigation on the Electromagnetic Shielding Effectiveness" (in Japanese), IECE Technical Report, MW85-68, pp.9-16, 1985.
9. A. Manara, "Measurement of Material Shielding Effectiveness Using a Dual TEM Cell and Vector Network Analyzer," IEEE Trans. EMC, vol.38 No.3, pp.327-333, 1996
10. E. Hanada, Y. Watanabe, Y. Antoku, K. Kenjo, H. Nutahara and Y. Nose, "Hospital Construction Materials: Poor Attenuation Effect with Respect to Signals Transmitted by Mobile Telephones," Biomedical Instrumentation & Technology vol.35 No.4, pp.489-496, 1998.
11. E. Hanada, Y. Antoku, S. Tani, M. Kimura, A. Hasegawa, S. Urano, K. Ohe, M. Yamaki and Y. Nose, "Electromagnetic Interference to Medical Equipment by Low Power Mobile Telecommunication System," IEEE Trans. EMC, vol.42 No.4, pp.470-476, 2000
12. DDL. Chung, "Electromagnetic Interference Shielding Effectiveness of Carbon Materials," Carbon, Vol.39, pp.279-285, 2001.

Eisuke Hanada was born in Tokyo, Japan, in 1963. He received his B.Eng. and M.Eng. degrees from Kyushu University, Fukuoka, Japan, in 1985 and 1987, respectively. He received his D.Eng. degree from Saga University, Saga, Japan, in 2001. Since 1992, he has worked at Nagasaki University Information Science Center for four years managing the campus LAN and information servers, and at the Department of Medical Information Science, Kyushu University Graduate School of Medical Science. Since 2002, he has been working at the Department of Medical Informatics, Shimane Medical University Hospital. His research involves the wired/radio communication environment and information processing systems in hospitals.
Dr. Hanada is a member of the Japanese Society of Medical Informatics, the Information Processing Society of Japan, and the Acoustical Society of Japan.

Kyoko Takano was born in Fukuoka, Japan in 1970. She received her B.Econ. and M.Info. degrees from Teikyo Heisei University, Ichihara, Japan, in 1996 and 1998, respectively. She has been a research resident at the Department of Medical Information Science, Kyushu University Graduate School of Medical Sciences since 1998. Her research interest is the electromagnetic environment of hospitals and educational methods for nursing school students.
Ms. Takano is a member of the Japanese Society of Medical Informatics.

Yasuaki Antoku was born in Fukuoka, Japan, in 1970. He received his B.Eng. from Kyushu Institute of Technology, Fukuoka, Japan, in 1996. He is currently working at the Department of Medical Information Science, Kyushu University Graduate School of Medical Sciences. He is researching hospital LAN systems and is managing Internet servers.

Kouji Matsumura was born in Kochi, Japan, in 1973. He received his B.Eng. from Osaka University, Osaka, Japan, in 1996. He is currently working at the Department of Medical Information Science, Kyushu University Graduate School of Medical Sciences. He is researching hospital information systems and is managing Internet servers.

Yasushi Hoshino was born in Tokyo, Japan, in 1973. He received his B.Eng. and M.Eng. degrees from Science University of Tokyo, Tokyo, Japan, in 1997 and 1999, respectively. Since 1999 he has been working at the Department of Shielding Electromagnetic Waves, Nippon Sheet Glass Environment Amenity Co. Ltd.
Mr. Hoshino is a member of the Architectural Institute of Japan.

Toshiya Nishimura was born in Kyoto, Japan, in 1973. He received his B.Eng. degree from Shibaura Institute of Technology, Tokyo, Japan, in 1996. Since 1996 he has been working at the Department of Shielding Electromagnetic Waves, Nippon Sheet Glass Environment Amenity Co. Ltd.

 

Kennichi Hyoudou was born in Ehime, Japan, in 1966. He received his B.Eng. degree from Oita University, Oita, Japan, in 1996. Since 1996 he has been working at the Department of Shielding Electromagnetic Waves, Nippon Sheet Glass Environment Amenity Co. Ltd.

Yoshiaki Watanabe received his D.Eng. degree from Kyushu University, Japan, in 1978. He has been an assistant at Kyushu University, a lecturer at Kyushu University Hospital, and an associate professor at Saga University, Japan. Since 1990, he has been professor of the Department of Information Science at Saga University. His research interests are in the fields of neural networks and biomedical engineering.

Yoshiaki Nose was born in Fukuoka, Japan, in 1944. He received his MD and Ph.D. degrees from Kyushu University, Fukuoka, Japan, in 1969 and 1973, respectively. He is currently professor and chairman of the Department of Medical Information Science, Kyushu University Graduate School of Medical Sciences and the Medical Information Center at Kyushu University Hospital.

 

---

ULTRAWIDEBAND TIME- AND FREQUENCY-DOMAIN CHARACTERIZATION OF EMC FACILITIES
David R. Novotny, Robert T. Johnk, Claude Weil, Nino Canales

National Institute of Standards and Technology
RF Technology Division
325 Broadway
Boulder, CO 80305-3328
www.boulder.nist.gov/div813

Abstract
We have developed a methodology to determine the quality of an EMC (Electromagnetic Compatibility) test facility using equipment generally available to RF testing services. By utilizing the time- and frequency-domains, an accurate picture of the scattering and modal properties of a facility can be determined. This gives more information about performance than traditional scalar, swept frequency measurements of the facility. While the frequency information given from traditional Normalized Site Attenuation (NSA) type measurements can indicate facility performance issues, this dual-domain method can highlight the causes of facility irregularities. This can help eliminate guesswork and focus remediation efforts on a facility that may be out of compliance.
Keywords: Chamber, EMC, facility evaluation, fully anechoic room, FAR, normalized site attenuation, NSA, OATS, Time-Domain, ultra-wideband, UWB.

1. Introduction
Most current EMC and RF testing facilities have components that make the facility diverge from an ideal testing environment. Finite ground planes, shelters, obstructions, and positioning equipment can cause open area test sites (OATS) to vary from the idealized model of a free-space antenna over an infinite ground plane. These irregularities can be manifested as deviations in the NSA and can affect confidence in measurement results or increase measurement uncertainty. Similarly, unoptimized absorber, ventilation and access pathways, surface discontinuities, and obstructions such as mounting and fire-suppression hardware can cause anechoic chambers (and their semi-anechoic counterparts) to exhibit a less than desired complex scattering environment as opposed to the desired free-space operation. This can cause anomalies in measurements taken in the facility.
We have developed a self-referencing technique [1-3] that illuminates the facility with a short-time, ultra-wide frequency-band impulse using a broad-spatial beam width antenna system. By using both the time- and frequency-domains, a fault or scattering center can be located in time and space and, through gating the frequency effects on the overall system, can be determined.
We show data taken at several facilities and the frequencies at which the facilities are experiencing less than optimal operation. The time-domain analysis shows where faults are located. By utilizing both domains in our analysis, corrective solutions can be suggested and tests quickly retaken. While this is not a totally comprehensive test method, we believe that the technique shows great promise as a pre-compliance test and a fault locator.

Figure 1. The 1.5 GHz Impulse Response of the TEM Horn Antenna System. This reference measurement yields an effective range resolution of 30 cm and a baseline for measuring the frequency response of the facility.

2. Measurement Setup
In order to discriminate between scattering centers, fine spatial resolution over a large frequency span is required. Additionally, an ultrawideband, short-impulse antenna system is needed to transmit and receive the stimulus with a minimum of dispersion and antenna ring-down. Most current EMC standards mandate the testing of radiated emissions in the frequency range of 30-1000 MHz. This paper deals with testing in this range and along with the associated setup and procedure to ensure the acquisition of U.S. Government work, is not subject to copyright.
These measurements can be performed using either an impulse generator and oscilloscope with suitable frequency coverage or a vector network analyzer (VNA) to analyze the facility. The oscilloscope offers an unaliased transform and better ambient signal rejection; however (using normally available pulse systems), it lacks the dynamic range of a VNA. If a VNA is used, care must be taken to have a suitable number of frequency points and coverage to have an unaliased time window long enough and time resolution fine enough to discriminate between spatial events.
We have chosen to use a VNA for this series of measurements because of stability, availability and overall dynamic range. We took 802 data points evenly spaced from DC to 2000 MHz in 2.5 MHz steps. This resulted in an unaliased time window of 400 ns and an impulse response corresponding to a range resolution of approximately 15 cm, shown in Figure 1. The DC point was interpolated from the S21 data with a least-mean-squares circle fit to a real value using the initial complex data values taken with the VNA. Generally, a measurement window of 400 ns is adequate for a reasonably sized EMC facility. Our measurements confirm that most facility ringing has been attenuated well before 400 ns, so aliasing is not a problem. While the upper frequency limit of 2000 MHz is beyond the 3 dB bandwidth of the antennas and above the general usage of these facilities, the antenna falloff can be calibrated out and corrected. The extension in frequency beyond 1000 MHz and facility performance is of interest to the facility operators.
The antenna system consists of a pair of matched TEM horns. Several NIST researchers have constructed different TEM horns covering a variety of frequency ranges [4, 5]. They utilize a resistive termination to prevent aperture reflections and longitudinal moding in the horn. The horns we used have a nominal antenna factor of 25 dB from 25 to 1200 MHz and with their size of 1.2 x 0.3 x 0.3 m, offer a relatively small scattering cross-section and little facility loading. This allows modal structures to be accurately measured. The TEM horn design offers a broad frequency range of operation and a constant electrical phase center with a very short impulse response, which allow for minimal dispersion of the stimulus. The small aperture (0.3 x 0.3 m) creates a broad pattern that illuminates a large portion of the facility under test and provides fine spatial sampling.

Figure 2. Basic Data Processing Diagram. If the direct path between the horns is unobstructed, and the impulse response is time-separable from the ground-bounce and facility ringing, then only one self-referencing measurement is needed.

3. Measurement Methodology
These measurements are self-referencing; they do not rely on calibrated antennas or other reference artifacts to analyze the facility. By spacing the antennas in close proximity to each other (or over a small, well-defined, reflector for OATS evaluation) the horn-to-horn coupling can be isolated from any external interference by appropriate time gating. This gives an accurate reference that can be compared to the facility measurements. Often these reference measurements can be performed either as a part of the facility measurements or as a special setup at the facility. To ensure that the entire facility is subjected to testing, a number of spatial positions and polarizations are generally taken to ensure full illumination of the test site. Special care is taken to cover typical or regulatory dictated test volumes or antenna and device under test (DUT) configurations. By comparing the various spatial measurements, faults can be analyzed for more than one aspect angle, and triangulated so that they can be directly identified.
The data processing model is shown in Figure 2. Once the facility and reference measurements are taken, the time-domain behavior is analyzed and time gates can be applied around specific events or sections of the data. The resultant gated waveforms are transformed back to the frequency-domain and normalized to the antenna reference. By not applying a time gate to the facility data, the overall response of the facility can be determined. Time-gate applications can identify and separate out the quality of the OATS surface, sidewall reflections in anechoic chambers, or modes and late-time ringing in the facility.

Figure 3. Relative Antenna and Shelter Positions Used in the Measurement of the Partially Covered OATS Facility.

4. OATS Evaluation
A commercial OATS was measured to assess the effects of a limited ground plane and a fiberglass shelter built to cover the DUT, but not the measurement antenna. The general layout is shown in Figure 3. To reduce edge reflections, the long, narrow ground plane is electrically connected to a tapered hill. The structure is a large fiberglass and foam building meant to protect equipment from the elements. RF energy penetrates the structure's walls before the outside antenna receives it. Additionally, there was concern that the structure may have some resonances that could attenuate or enhance emission or immunity characteristics. This could cause equipment to needlessly fail regulatory tests, be unknowingly susceptible to interference, or pass a failing device (a pre-compliance facility operator's constant headache).
The measurements were taken in two phases, as shown in Figure 4. The reference measurements were performed outside the equipment shelter at horn aperture separations 2 m and 10 m so a clean horn-to-horn measurement, without the shelter, could be made. This gave a reference with which to compare the obstructed measurement and a horn-to-horn reference to assess the quality of the ground plane. Then measurements were made of the shelter effects by placing the antennas at typical antenna and DUT positions.
The results, in Figures 5 and 6, showed a slight (<0.5 dB) attenuation through the shelter and a delayed ground bounce (this caused a shift in the 900 MHz null, which appears as a spike in Figure 6). The more important effects were secondary scattering from the shelter roof and ringing in the shelter that caused a variation of 1 to 2 dB across the measurement band.

Figure 4. The Time-Domain Response Plots Showing the 10m OATS Reference.

5. Ferrite-Lined, Pre-Compliance Chamber Evaluation
The NIST system was used to determine the effects of an upgrade to a ferrite-lined pre-compliance chamber. This facility was smaller than traditional chambers and reliable low-frequency performance was a concern. The operator was adding additional absorber to the chamber and was interested in the performance before and after the modifications. Due to initial budget limitations at the time of installation, a chamber of smaller than optimal size was procured, and the operator wanted performance data, information on any chamber imperfections, and possible remedies for them. There was exposed metal around the ventilation intakes, floor and access door. Retrofits attempting to cover these scatterers were planned with the measurement effort, and the operator wanted to know how the chamber tested before and after the changes.
The initial reference measurements were made in the center of the facility well away from the walls and floor at an aperture separation of 2 and 3 meters.
The time-domain analysis in Figure 7 showed that there were large reflections from the floor. The ferrite tiles on the floor were covered with a protective ESD carpet to protect the tiles and to avoid static buildup that could damage the equipment under test. This had the effect of reducing the effectiveness of the floor absorber. There were also reflections from the sidewalls of the chamber that were indicative of unoptimized absorber performance. By using the relative timing differences in the received signal and mapping the timing information with antenna position, triangulation of the location of the unknown effects was possible. The analysis showed that reflections were coming from the metal-ferrite edges and the absorber on the walls. The wall scattering was mainly due to the absorber being illuminated at large oblique angles, for which it was not optimized. The operator used this data to ameliorate the effects of these scattering centers.
The frequency-domain analysis, in Figure 8, showed large deviations due to chamber moding, scattering and unoptimized absorber usage. Without the dual-domain analysis, failure mechanisms would need to be guessed using intuition. While the retrofits the operator performed had some success, the results of our measurements showed that, other than full replacement of the absorber system, treatment of the specular reflections from the sidewalls and floors would have been a better improvement and could have been a more optimal use of retrofit funds.

Figure 6. The Raw, Frequency-domain Transformed Data and Normalized Performance of the OATS Facility in Horizontal-polarization at the Test Distance of 10 Meters. The excursion around 900 MHz is due to propagation delays through the shelter.

Figure 5. The Time-Domain Response of the OATS Facility. These measurements were taken through the shelter with the TEM horns in the typical DUT and measurement antenna locations (10m antenna separation, horizontal polarization).

Figure 8. Frequency Response and the Normalized Analysis of the Anechoic Chamber. The deviations in the data are due to strong sidewall reflections, absorber discontinuities, and long-term ringing in the facility.

Figure 7. Time-Domain Plots Inside a Ferrite-lined, Anechoic Chamber. At two separations, this allows for spatial location of the "unknown effect" by correlating the antenna locations and the relative timing of the effect. The distance separation increased by 1m, which accounts for the delay in the separation 2 and the delay difference between the fault and the direct signal between the measurements locates the fault 2.2 meters off the centerline of the antennas.

6. Summary
The use of time-and frequency-domain analysis shows great promise in facility evaluation. It provides the same information as the current NSA and chamber qualification tests; however, it also provides very useful insight into a facility's faults and whether they can be mitigated. The time-domain can show the presence of singular scatterers and modal buildups and discriminate between them. Then the information can be transformed into the frequency domain to determine the overall effects on the measurement system.
This is not a totally comprehensive test; only modes excited by the antennas can be measured, so it is limited by the scope of the antenna positions at which the measurements are taken. The effects of large DUTs and DUT/facility interactions also may affect facility performance not addressed by these tests. Yet we believe that this method is an improvement over current methods of standard facility testing and certification.

7. References
[1]. R. Johnk, D. Novotny, C. Weil, M. Taylor, T. O'Hara, "Evaluation of a Fully Anechoic Chamber Using an Ultra-Wideband Measurement System", Antenna Measurement Techniques Association 2001, Proceedings of the, Denver, CO, pp. 321-26, Oct 2001.
[2] Johnk, R.T.; Novotny, D.R.; Weil, C.M.; Taylor, M.; O'Hara, T.J, "Efficient and Accurate Testing of an EMC Compliance Chamber Using an Ultra-Wideband Measurement System" 2001 IEEE International Symposium on Electromagnetic Compatibility, Volume 1, page(s): 302 -307.
[3] Johnk, R.T.; Novotny, D.R.; Weil, C.M.; Medley, H.W. "Assessing the Effects of an OATS Shelter: Is ANSI C63.7 Enough?" 2000 IEEE International Symposium on Electromagnetic Compatibility, Volume 2, page(s): 523 -528.
[4] C. Grosvenor, D. Novotny, J Veneman, N. Canales, "Ultra-Wideband TEM-Horn Antenna Design Using Numerical Methods," Antenna Measurement Techniques Association 2002, Proceedings of the, Cleveland, OH, Nov 2002.
[5] J. Veneman, D. Novotny, C. Grosvenor, R Johnk, N. Canales, "Ultra-Wideband Antenna Pattern Characterization in a Non-Ideal EM Facility," Antenna Measurement Techniques Association 2002, Proceedings of the, Cleveland, OH, Nov 2002.

8. Biographies
David Novotny has worked in the RF Technology Division of NIST for over 12 years. Mr. Novotny earned his bachelor degree in Electrical Engineering from the University of Colorado at Boulder in 1990 and his M.S. degree in 1996. His research interests include time-domain measurements and modeling, especially ultra-wideband antenna and systems analysis.

Robert Johnk is the project leader of the NIST Time-Domain Free-Field Metrology program, where he has worked for 13 years. Dr. Johnk obtained his Ph.D. degree in electrical engineering at the University of Colorado. Dr. Johnk's research interests are time-domain scattering, sensor calibrations, shielding performance measurements, antenna and sensor calibrations, numerical electromagnetics, and ultra-wideband emissions metrology. Dr. Johnk has received IEEE EMC Symposium and NIST best paper awards. Dr. Johnk is a technical advisor to the United States Delegation to CISPR/A and he is active in several CISPR and ANSI working groups.

Claude M. Weil (M'64-SM'95-F'00) was born in Newcastle-on-Tyne, U.K., on June 26, 1937. He received the B.Sc. degree from the University of Birmingham, Birmingham, U.K., in 1959, the M.S.E. degree from George Washington University, Washington, D.C., in 1963, and the Ph.D. degree from the University of Pennsylvania, Philadelphia, in 1970, all in electrical engineering. Prior to 1964, he worked as a Navy Systems and Instrumentation Engineer and also designed microwave components and antennas. From 1971 to 1983, he was with the Environmental Protection Agency's research program on the health effects of RF radiation. From 1983 to 1985, he was with Boeing Military Aircraft Company, where he was involved with radar cross-section (RCS) measurements and analysis. Since 1986, he has been with the Radio-Frequency Technology Division, National Institute of Standards and Technology (NIST), Boulder, CO, where he developed millimeter wave six-port systems and power standards. He has served as a Senior Project Leader in the NIST Electromagnetic Properties of Materials Program and is currently a senior staff member in the NIST Time-Domain Free-Field Metrology program. Dr. Weil is member of Sigma Xi and a fellow of the IEEE Microwave Theory and Techniques Society (IEEE MTT-S) and the Instrumentation and Measurements Society. He served as general chairman of the 1997 IEEE MTT-S International Microwave Symposium, Denver, CO.

Nino Canales joined NIST in 1989 after serving 20 years in the US Army in the calibration field. Nino has a two-year science degree from New York University and an under-graduate degree in Business from Columbia College. Nino has over 30 years experience in antenna measurements.