Building an Open Area Test Site (OATS) for Antenna Calibration

 

By Malcolm Rich and Ian Russell, QinetiQ Aquila

 

 

Introduction

The first question any sane person would ask is “why build one?”; especially when there are a number of other antenna calibration facilities in the UK. The justification for it was purely financial!  There is a large Electromagnetic Facility at QinetiQ Aquila, with a significant antenna calibration requirement, in addition to those from the rest of QinetiQ.

 

Additionally, we were developing new test methods for the UK military EMC Defence Standard that required the use of an OATS as well as enabling us to expand into the field of commercial EMC.

 

This gives us the why, the how was of course a lot more difficult, and to lead this project we recruited Luke Turnbull (Now PhD) from York University. It was his particular experience and use of novel techniques that enabled us to get where we are today, which is an OATS designed for use up to 5GHz.

 

OATS design

The first task was to investigate the performance of different sizes of OATS ground-planes. An important part of the construction of an OATS is the metallic surface, which is laid onto the ground to stabilise the grounds reflection co-efficient. Advice on the construction of OATS can be found in a number of standards documents such as ANSI C63.7 [1] or EN55 022 [2]. The size of ground-plane needed to give the required reflection characteristics is described in both [1, 2].  [1] in particular goes into much detail in assessing the size needed for a good reflecting surface by using the theory of Fresnel Zones.

 

The first Fresnel Zone is the area of reflecting surface for which the path length for a reflection from the edge of the Zone is half a wavelength longer than a reflection from the centre of the Zone.

 

Referring to Figure 1, the mathematical description of the Fresnel Zone is:

d2 - d1 = l/2(1)

where d2 describes the path taken when a reflection occurs from a point on the edge of the Fresnel.

 

 

 

Figure 1: Ray tracing diagram for Fresnel Zones problem

Figure 2: Fresnel Zones schematic diagram for nearly parallel rays

 

Zone, and d1 is the ray-tracing path of reflection.  [1] considers that the first Fresnel Zone is the minimum area needed to provide a good reflection but gives no indication as to how the reflection characteristics of this surface compare with the characteristics of the ideal case of an infinitely large surface. The use of the Fresnel Zone is also lacking in that only the phase difference between the reflections from various parts of the reflecting surface are used, and that the path length differences are not.

 

The investigation attempted to find a more detailed analysis of the required size of reflecting surface than that in [1].

 

Fresnel Zones

The theory of Fresnel Zones is taken from books on optical theory, for example [3]. The problem from the point of view of optical theory is shown schematically in figure 2.  The light from the source of radiation, which reaches the sink of radiation, travelling through the aperture, is treated using the Huygens principle.  This states that each element of area at the aperture acts as a point source of radiation. The following equation is the description of light travelling from source to sink:

 

               (2)

 

with:

   —    field strength

λ     —    wavelength

U0    —    intensity of source

k     —    wave-number

    —    vector from source to point on aperture

s     —    vector from point on aperture to sink

θ     —    angle between r and s

dσ   —    element of area at point on aperture

 

 

Equation 2 is Fresnel’s equation.  The term  accounts for the phase difference between the rays passing through different points of the aperture.  Similarly the term 1/rs describes a decay in intensity due to the path length difference between rays passing through different points of the aperture.

 

The two objectives were to account for the path length difference between the reflection from various points on the ground-plane and to give an indication of the quality of the reflection from particular sizes of ground-plane.

 

Investigation method

The method used for the investigation was to calculate the field intensity by evaluating equation 2 numerically. The height of the source and sink and their separation were set to typical values. The ground-plane was segmented into areas with linear dimensions less than a tenth of a wavelength and the reflection from each of these elements of area was calculated. In addition to the propagation via the ground-plane, the propagation due to the direct path between source and sink was also evaluated. The effect of both paths of radiation were summed, resulting in a value of propagation loss which was a close approximation to the type of measurements that the OATS is used for. Finally, the above calculation was repeated for different sizes of ground-plane, and the propagation loss plotted as a function of ground-plane length.

 

Validating the method

The method has been validated against the method of images for calculating the reflected field over a perfect ground-plane.  For the geometry of a source and sink height of 2m, and a separation of 10m, the method of images can be used to predict the field strength at the sink position.

 

The results of the computational prediction of propagation loss for this geometry are compared with the method of image prediction in figure 3.  The agreement between the two methods can be seen to be excellent for the higher frequencies. At lower frequencies the methods start to diverge at around 300MHz, and the agreement is poor below about 80MHz.  It should be noted that the theory used assumes far field conditions and as such would not be representative of OATS measurements where the measurement distance is similar to a wavelength.

 

 

Figure 3: Validation of the computed propagation with image theory. Source and sink height 2m, separation 10m.

 

Investigation

The computation described in the above sections was used to investigate the relative performance of different sizes and shapes of OATS ground-plane.  Firstly, a comparison between the performance of an elliptical ground-plane and a rectangular ground-plane of the same overall length was made.  The elliptical ground-plane had the same proportions as the Fresnel Zones for that measurement geometry. In all of the investigations, the length and width of the ground-plane were varied simultaneously and the proportions of the rectangle characterised by their width to length ratio. The comparison is shown in figure 4.  It can be seen that the propagation loss for the rectangular ground-plane converges to the ideal value a lot more quickly than that for the elliptical ground-plane. This result is convenient when considering the practical advantages of constructing a rectangular ground-plane.  Secondly, propagation losses were calculated for rectangular ground-planes with various measurement geometries, various width to length ratios and at a number of frequencies.

 

 

Figure 4: Comparison between elliptical surface and rectangular surface (Width to Length ratio 0.75). Source and sink height 2m, separation 10m. Frequency 400MHz.

 

Conclusions

The method used is an improvement on the advice previously available, because there are fewer assumptions made.  In addition, the method presented gives an indication of the likely performance of a certain size of ground-plane, which was also lacking in previous advice. The results presented show the relative performance of elliptical and rectangular shaped ground-planes. The relative performance of differently shaped rectangular ground-planes was also investigated.  Because the predicted propagation losses presented are a function of five variables, it has only been possible to show a few representative calculations to illustrate the technique.  The method was used to make an informed decision on the design of an OATS.

 

The final product

We eventually decided on building a 15m by 9m ground plane surrounded with 3m of conductive mesh. This results in a total ground plane area of 21m by 15m. The mesh used has a pitch smaller than the shortest wavelength we were interested in. Construction was started early in 1998, with the work being completed by July of the same year. This included a dedicated brick building with A/C where all of the test equipment was installed.

 

The hard work of deriving test methods for various antenna types was started, with those for biconical and log-periodic antennas completed first.  Our first UKAS visit was 20 December 1999, an interesting event because we had to use a blowtorch to melt the ice off of the stainless steel hatch for the cable connectors! Soon after, we added loop, bilog and horn antennas to our scope of accreditation.

 

Problems and solutions

There were a number of problems encountered with the process of developing out test methods, one of them related to the bilog antenna.  The problem we encountered was the difficulty in achieving clean measurement from 20MHz to 1GHz. The source of the problem was traced to a grass bank near the location of the OATS.  We could not move the bank, so our solution was to move the antenna by raising it to 6.3m above OATS surface.

 

Future plans

The QinetiQ Aquila facility complements the 500m far field antenna range at QinetiQ Funtington used for antenna and radar signature measurement up to 60GHz. QinetiQ intends to develop both these facilities using their combined experience and capabilities to meet the demanding requirements of both military and commercial antenna calibration in addition to R&D projects.

 

References

[1] ANSI C63.7-1992 American National Standard Guide for Construction of Open Area Test Sites for Performing Radiated Emission Measurements. Published by IEEE.

[2] BS EN55 022 : 1995 and CISPR 22 : 1993. Limits and methods of radio disturbance characteristics of information technology equipment. British Standards Institution.

[3] E. Hecht Optics Addison Wesley. ISBN 0-201-11611-1.

 

For further information contact: igrussell@qinetiq.com