OATS Antenna Factors for Small Anechoic Chambers

Deviations in the correlation to open area test site (OATS) results recently have been reported mainly in the frequency range between 30 and 200 MHz. The question arises about how accurately the manufacturer’s OATS antenna factors can be applied in smaller test chambers because of antenna interactions with the nearby walls. This concern applies to biconical as well as bilog antennas.

To determine the in situ antenna factors for a small anechoic chamber, an individual, traceable calibration was performed using several independent procedures.

Experimental Setup

All experiments were carried out in a 7-m × 4-m × 3-m fully anechoic ferrite chamber with a Rhode & Schwarz Model HCT12 Automatically Controlled Turntable with a ferrite-coated base. The 1.2-m dia circular plate was elevated 0.8 m and supported on a pertinax tube. The removable plate consists of laminated, glued wooden layers and represents wooden test tables like those used in almost any lab.

The layout of the chamber is shown in Figure 1 (see the September 2001 issue of Evaluation Engineering.). The 3-m test distance has been arranged diagonally within the chamber to obtain better normalized site attenuation (NSA) results. The chamber has 5-mm × 100-mm × 100-mm EUPEN Cunico tiles mounted on 300-mm × 300-mm panels. The antenna mast is a shortened Model HCM from Rhode & Schwarz.

Two independent procedures were used to determine the antenna factors:

1. The field was generated by a Chase Blue Bilog Antenna at a 1.5-m height, the E-field was measured by a Holaday FM 2000, and power was provided by a Holaday FP 2000 Probe positioned over the center of the turntable at a 1.5-m height. Antenna factors are derived from measurements of forward power in the E field.
2. A battery-powered Schwarzbeck IGUF Impulse Generator fed a Schwarzbeck Precision-Tuned Dipole Transmitting Antenna. A Chase Blue Bilog Receiving Antenna was at a 1.5-m height and connected to a Rhode & Schwarz ESMI 28-GHz Receiver.

To cross-check the chamber characteristics, another set of two procedures was used. The NSA for fully anechoic chambers according to EN 50147-2 1996 was measured with Rhode & Schwarz Model HK116 Biconical Antennas and Model HL223 Log-Periodic Antennas. Both antenna types were calibrated by the manufacturer using the free-space method.

To correlate from 3-m fully anechoic to 10-m OATS emission results, a special algorithm was applied.¹ The measurement distance and the ground-plane reflections are considered as system factors.

Since full-size biconical antennas could potentially interact with the fairly close anechoic walls, a 40-cm-long small biconical antenna with a 10-MHz battery-operated comb generator (VSQ) was used to check the volumetric effects in the turntable region. Each of the electronic measurement systems was calibrated under EN 45001 (17025) and traceable to national standards.

The vertical NSA is particularly critical for OATS and some chambers.² This is why the vertical NSA was measured in the first place (Figure 2, see the September 2001 issue of Evaluation Engineering.).

Between 30 MHz and 60 MHz, Figure 2 shows that either the chamber or the antenna has a problem. For that reason, the vertical antenna factors have been measured in the chamber with the field probe and the tuned dipole procedure.

The tuned dipole procedure starts at 60 MHz and not at 30 MHz because of the physical length of the tuned dipole in that frequency range. It is important to realize that in the vertical and horizontal case the turntable always is in place during the measurements. This unwanted interaction of the dipole with the turntable and walls extends to 200 MHz. In each case, the manufacturer’s data was 10-m free-space OATS antenna factors, taken at a 4-m height in vertical polarization.

While all methods are in good agreement above 200 MHz, the most precise procedure down to 30 MHz obviously is the field-probe method. As a result, this data was used later as the actual antenna correction factor, and it agrees with the NSA.

Comparing the vertical NSA to the measured vertical antenna factors reveals at least 5-dB deviation from the ±4-dB criterion at 30 MHz (Figure 3, see the September 2001 issue of Evaluation Engineering). The overall averaged correction above 200 MHz ranges from 0 dB to about 2 dB and is far less profound than the low-frequency correction.

Problems at Higher Frequencies

In the frequency range above 200 MHz, the horizontal polarization seems to be more critical in a fully anechoic chamber with a ceiling height of about 3 m. This effect of a rectangular longitudinal cross section leads to a smaller angle of incidence with respect to the ferrite walls, so a test was performed with the E-field probe.

The findings are shown in Figure 4, see the September 2001 issue of Evaluation Engineering. This plot demonstrates maxima and minima values in the forward feed power to the Chase antenna over the frequency range from 100 MHz to 800 MHz. The E field was kept constant at 10 V/m.

A rough propagation analysis, involving the ceiling and floor reflections, leads to the following results for the geometry given in Figure 5, see the September 2001 issue of Evaluation Engineering. The length of the direct path is d = 3 m.

For h = 2.83 m, the length of the indirect path s is computed as:

The path length difference is computed as:

D = s – d = 1.12 m          (2)

Looking at the constructive and destructive interference conditions and considering the phase shift of 180° at the reflection points lead to the following power maxima (c = speed of light, n = 1,2,…):

and:

(4)

f_1,max = 268 MHz
f_2,max = 536 MHz
f_3,max = 810 MHz

The power minima occur at:

and:

(6)

f_1,min = 134 MHz
f_2,min = 405 MHz
f_3,min = 670 MHz

Power maxima leads to field strength minima for a constant transmitting power and vice versa.

To fix the problem of high-frequency maxima and minima, the emission method with the tuned dipole was used to confirm these findings. As a result, the dipole was tuned to the 810-MHz maximum, and the IGUF pulse generator used to feed the dipole was positioned in the center of the test volume. The results with and without additional absorbers are shown in Figure 6, see the September 2001 issue of Evaluation Engineering.

It is important to understand that only one or two lines of absorber blocks (60 cm × 60 cm), positioned halfway between the transmit and receive antennas on all four walls, have a marked effect on the chamber specifications. The exact position must be optimized experimentally. By no means is it enough just to cover the floor or only the ceiling, due to symmetry aspects. The fix, however, certainly is worth the effort and requires very little money.

Tabletop Reflections

Reflections are introduced by the wooden top plate of the turntable, the test table used to support EUTs. Although it is required by CISPR 11, 16, and 22, the wooden tabletop strongly interacts with the floor and ceiling effects. This has been confirmed by the VSQ emission test radiator as shown in Figure 7 and with the E-field probe procedure.

The setup with and without the wooden, laminated, 1.2-m dia top plate demonstrates 20-dB reduced field strength at the receiving Chase antenna around 660 MHz. This effect only recently has been reported in the literature and only tested for the immunity case.³

Such findings are not unexpected, since wood and plywood have considerable reflection coefficients at higher frequencies. However, these reflections may lead to major uncertainty in EMC tests for radiated emissions and immunity.

The standardization committees are advised to further consider and investigate wooden tabletop reflections at higher frequencies. Similar adverse effects also are present in larger chambers, both in semi-anechoic and fully lined facilities.

Conclusion

Small, fully anechoic chambers with ferrite lining may be improved for use in compliance testing by determining the in situ antenna factors and considering the two major reflection effects. In particular, the tabletop reflections must be evaluated very carefully. The additional investment cost is marginal, but the quality improvement is decisive.

References

1. Ristau, D. and Hansen, D., “Correlating Fully Anechoic to OATS Measurements,” Proceedings of the 13th Wroclaw EMC Symposium, 1996, pp. 402-405.

2. McConnell, R.A. and Vitek, C., “Calibration of Fully Anechoic Rooms and Correlation With OATS Measurements,” Proceedings of the IEEE EMC 96 Symposium, 1996, pp. 134-139.

3. Beggio, A., Borto, G., and Zich, R.E., “The Unwanted Effects on the Radiated (Emissions and) Susceptibility Measurements Due to the Introduction of a Wooden Table,” Proceedings of the 1999 International Symposium on EMC, 1999, pp. 252-255.

About the Authors

Diethard Hansen, Euro EMC Service (EES) Dr.-Ing. D. Hansen, Bahnhofstr. 39, CH-8965 Berikon, Switzerland, 011 41 566 337381, e-mail: [email protected].
Peter Lilienkamp, University of Applied Science FH Kiel, Grenzstr. 3, D-24149 Kiel, Germany, 011 49 4312104061, e-mail: [email protected].

Published by EE-Evaluation Engineering
All contents © 2001 Nelson Publishing Inc.
No reprint, distribution, or reuse in any medium is permitted
without the express written consent of the publisher.

September 2001