TEM Waveguides: An Evaluation of Their Applicability for EMI Testing

Depolarization angle determines the frequency range at which the TEM mode is dominant.

The testing of electromagnetic interference can often present problems due to the high costs of constructing and preparing test equipment. In recent years, investigations into how these very complex field tests could be carried out using simpler, less expensive equipment have been conducted.

Various proposed new test installations have been based upon far-field measurements of an antenna. The aim of these installations has been to test the emission or interference of a device under far-field conditions that are as realistic as possible. Transverse electromagnetic (TEM) cells and other TEM waveguide constructions have already entered the compliance standards. Well-enclosed housings no longer require testing in expensive shielded rooms. Therefore, a large demand for TEM waveguides has evolved. Recently, many test facilities with this type of inexpensive equipment have been promoted in the marketplace as offering alternative possibilities for field measurements. They all claim to fulfill the function of TEM waveguides.

This article reports an investigation into how it could be determined whether such facilities fulfill TEM-wave conditions. Relevant definitions presented in IEC 61000-4-203 are considered. This standard should be made normative as EN 61000-4-203. The general definitions in the IEC standard will be clarified for the norm in order to allow correlation of the test devices. The proposed limits for field homogeneity and depolarization angle are also justified in this article.

Under these assumptions, TEM, GTEM, G-Strip, and MAC cells were tested. The result was that the applicable frequency range for each type of test facility could be clearly identified.

TEM Wave Propagation

Any electromagnetic field can be described as a superposition of well-known wave types called field modes.¹ The following vector sums represent this.

E represents the total electric field, and H represents the total magnetic field. There is no component of the x- and y-fields in the longitudinal direction, that is, the direction of wave propagation. There are only transverse components of the x- and y-fields. The TEM wave does not have to have a longitudinal component.

The ratio of the absolute values

and is given by the field wave impedance:

for propagation in air.

The propagation velocity of the TEM mode is given by the material constant likewise:

This is the well-known speed of light.

Identification of fields as the TEM type depends on the following criteria.

• Field components are found only in the transverse plane.
• The ratio of the absolute values is given by 120p W.
• Propagation velocity is equal to c₀.

Another important aspect of the TEM mode is that it is independent of the radiator/antenna.

So much for scientific considerations. Attention now turns to actual applications of field measurements. A typical test setup as proposed by IEC 61000-4-3 is shown in Figure 1. An antenna illuminates the device under test (DUT) for the frequency range 10–1000 MHz. The distance between the DUT and the transmitting antenna has to be 3 m.

Figure 1. Immunity measurement in an anechoic chamber (MAZ, Hamburg, Germany).

It can be shown that the radiated field far away from the antenna fulfills the TEM-mode criteria. The rule of thumb for the minimum distance between device and antenna is somewhat on the order of the wavelength used. For this case that rule is not valid. The wavelength from 10 MHz requires a distance of 30 m. On the other hand, a high field strength is needed at the DUT. So, a compromise was found, and the minimum distance has been set at 3 m. Nevertheless, TEM-mode conditions are assumed at the DUT.

A defined field mode, that is, the TEM mode, is necessary for ensuring reproducible and well-defined field measurements. If several test facilities are to be comparatively evaluated, the main task is to determine how each test site generates a pure TEM field.

Evaluation Criteria

The calibration procedure for field immunity tests according to IEC 61000-4-3 is schematized in Figure 2. Some criteria that alternative test equipment must fulfill, extracted from the standard, are now discussed.

Figure 2. Field calibration according to IEC 61000-4-3.

Field Homogeneity. IEC 61000-4-3, Chapter 6.2, requires the field to be homogeneous over a surface in front of the test object. A surface perpendicular to the direction of the field propagation is selected. This square surface must be at least as large as the irradiated surface of the DUT, but not larger than 1.5 x 1.5 m. With the DUT not present, the field strength is measured at 16 evenly distributed points. For at least 12 of the points, the difference between the minimum and maximum values obtained must not exceed 6 dB (see Figure 3). The surface on which the field points are located must be as large as the irradiated surface of the DUT. The test points may not be situated on the conducting surface or beneath the DUT.

Figure 3. Definition of field homogeneity for irradiation testing according to IEC 61000-4-3.

A disadvantage of this calibration procedure is that 25% of the measured values are not taken into account in any way. For this reason, a new statistical approach was proposed.² Assuming a statistical distribution of the measured values, the 6-dB criterion can be written as

with a mean value calculated as

standard deviation as

and probability as

To obtain the required homogeneity of the field strength, first the mean value and the standard deviation for the n = 16 measured field values have to be calculated. If

is no more than 6 dB, the criterion is fulfilled. As an example, Figure 4 displays the field homogeneity of a GTEM cell at 100 MHz.

Figure 4. Field homogeneity of a GTEM cell at 100 MHz.

Defined-Field Polarization. Regarding defined-field polarization, IEC 61000-4-3, Chapter 7, says that the investigation "shall be carried out separately for vertical and horizontal fields." This requires that the observed field strengths must be differentiated. The total field strength is commonly represented as

However, the field-strength component in the direction of polarization is often investigated. For waveguides with dominant higher-order modes, a consideration of the total field strength is better. Both values were determined in the comparative analysis shown in Figure 5.

Figure 5. Definition of the field components for immunity testing.

This requirement should clarify any possible channels of radiation intrusion. The dependence of interference on the direction of polarization allows deductions to be made concerning, for example, intrusive radiation caused by slits or other apertures.

By determining the depolarization between the theoretical and actual polarization angles, one can gain evidence as to the presence of a TEM mode. The existence of a component in the direction of propagation gives an indication of the presence of a higher mode. If a longitudinal component is observed for either E or H, a pure TEM field does not exist. The depolarization (error) angle can be deduced from the following equations:

^and

where E_xpol is the field-strength component in the transverse direction perpendicular to the polarization vector and E_long is the field-strength component in the direction of propagation (see Figure 6).

Figure 6. Definition of the depolarization angle.

The ideal error angle of 0° can never be achieved. Assuming a point source, the propagation of the TEM wave is spherical. This introduces an angle of 14°. As already demonstrated in Figure 4, a value of 20° for both angles is realistic.² This has led to the requirement that 75% of the measured depolarization angles must be within ±20°.

Fully anechoic chambers and any other rooms must also fullfil the 14° criterion. If they fail to do so, there is a problem with the chamber itself. In most cases, the failure is caused by the absorbers.

Measuring the Field Components

Normally, a highly resistive dipole is used to measure the components of the electric field (E-field; see Figure 7a). The radiation pattern of such a dipole is shown in Figure 7b. It can be seen that through the diode the sign of the E-field components is missing. This leads to a possible misinterpretation of the direction of the actual vector as shown in Figure 8. For simplicity, only the x-y plane is shown. The negative sign of E_y was not measured.

Figure 7. A highly resistive dipole is used to measure the components of the E-field. (a) The dipole with a diode to measure root-mean-square values. (b) The isotropic radiation pattern of the dipole.

Figure 8. The effect on interpretation of vector direction when the sign of the E-field components is missing.

To reconstruct the sign, the measurement has to be done twice at each point. Figure 9 shows the procedure for a rotation of the coordinate system of 15°. Again, only the components with a positive sign are measured. At the first measurement, the vector could be at the points marked with a cross. The second time, it has to be at the points marked with a circle. From this we can exclude the shaded area.

Figure 9. Reconstruction of the missing sign of the E-field components.

TEM Waveguides: Comparative Test Results

Five types of alternative TEM waveguides were investigated: a TEM cell of the Crawford type, a GTEM 500, a GTEM 1250, a G-Strip, and a MAC cell.

The positioning of the measurement points in the TEM cell is shown in Figure 10. Figure 11 depicts the measurement equipment used for the investigation of a GTEM cell. Regarding nomenclature used for the coordinate system, z is the coordinate in the propagation direction or longitudinal direction, y is the coordinate in the direction of polarization, and x is the coordinate in the direction of the transverse plane perpendicular to the direction of polarization. Nine measurement points in one plane were chosen for each of the three coordinates, and this plane was further considered. For the TEM cell, GTEM cells, G-Strip, and MAC cell, the frequency range of 30–1000 MHz was employed. The input voltage was always set to 10 V. Table I gives reference dimensions for the area of field points for the five TEM waveguides investigated (see Figure 10).

Waveguide	a (mm)	b (mm)
TEM	490	580
GTEM 500	450 ± 40	460 ± 50
GTEM 1250	1140 ± 94	1160 ± 94
G-Strip	350	660
MAC	225	450
Table I. Reference dimensions for the positioning of field points as in Figure 10.

Figure 10. Field points in the TEM cell.

Figure 11. Measurement equipment used in the investigation of a GTEM cell.

The complete results of the testing, too voluminous to be printed here, are available from the author. Test setups for each of the five TEM waveguides are depicted and accompanied by waveform graphs for transverse depolarization angle, longitudinal depolarization angle, field strength at a 75% interval, and the 75% interval expressed in decibels. For each field point, the components of the electrical field strength are shown. Key results were as follows.

The Crawford-type TEM cell produced a depolarization angle of 0° up to the first resonance of higher-order mode. Clearly, the depolarization angle is a very good gauge by which to test the presence of higher-order modes. For GTEM cells the case was different. A depolarization angle of ±9° for the low-frequency range appeared. This derived from the fact that wave propagation in a GTEM cell is spherical, by contrast with the planar wavefront in a Crawford cell. A planar wavefront has to be assumed for a G-Strip cell, but a strong variation of the angle could be seen in the diagrams. A slower fluctuation than for the TEM cell was apparent from a comparison of the measurements. This resulted from the absorber-lined walls of the G-Strip. Finally, the MAC cell showed a pure TEM wave for the low-frequency range, but for the higher range, strong higher-order mode propagation was seen.

Conclusion

Proposed new test sites were investigated to determine their equivalence to the established open-area test site and anechoic chamber. The coming clarification of the general definitions from the IEC 61000-4-20 standard to allow correlation between the test devices results in the condition that a pure TEM wave must be generated on such an alternative site. A new statistical approach to characterizing field homogeneity has been introduced. The depolarization angle is the criterion that determines the dominant presence of the TEM mode. Test results for TEM, GTEM, G-Strip, and MAC cells clearly identified the applicable frequency range for each test facility.

References

1. M Koch, "Analytische Feldberechnung in TEM-Zellen" (PhD diss., University of Hannover, 1998).

2. H Garbe, M Koch, and H Haase, "Specification of Alternative Test Sites with Respect to Given EMC Field Standards" in Proceedings of EMC Zürich '97 (Zürich: EMC Zürich '97, 1997) 459–464.

Bibliography

Garbe, H. "Beurteilungskriterien für TEM-, GTEM-Zellen, und andere Wellenleiter." In Proceedings of EMV '97 (Stuttgart: Mesago, 1997), Workshop 7.

Garbe, H, and Koch, M. "Normgerechtes Testen mit einer G-Strip?" research report, University of Hannover, 1996.

Groh, C, et al. "TEM Waveguides for EMC Measurements," IEEE Transactions on EMC 41 (1999): 440–445.

Kärst, JP, Koch, M, and Garbe, H. "Vergleich der Feldhomogenität verschiedener TEM-Wellenleiter." In Proceedings of EMV '98 (Düsseldorf: VDE-Verlag, 1998), 291–298.

Wilson, P. "On Correlating TEM Cell and OATS Emission Measurements," IEEE Transactions on EMC 37 (1995): 1–16.

Dr.-Ing. Heyno Garbe is a professor at the University of Hannover in Germany. He can be reached at +49 511 7623760 or by e-mail at heyno.garbe@ieee.org.

Back to September/October Table of Contents