The Measurement of EMC Characteristics

in Small Fully Anechoic Rooms

(FAR Project final report SMT4-CT96-2133)

Dave Fynn Reports

The above is the subject of two-year feasibility study funded by the European Community in support of a proposed CENELEC standard prEN50147-3. The remit was to investigate the viability of utilising relatively low cost Fully Anechoic Rooms (FARs) as an alternative to the existing 10m Open Area Test Site (OATS) or 10m Semi Anechoic Room (SAR). A parallel objective was to support the work of CENELEC TC210 to introduce a standard to enable FARs to be accepted as an alternative to OATS for full compliance emission testing.

The project was undertaken by a partnership of eight European EMC specialists and was co-ordinated by Martin Alexander of the National Physical Laboratory.

There is a growing awareness that testing on an Open Area Test Site in the presence of worsening ambient RF interference is yielding increasing measurement uncertainties thus becoming increasingly time consuming and consequently more costly. This, in some cases, has led to the use of 10m Semi Anechoic Rooms. Such facilities, however, are quite expensive, in excess of £1m. A 10m SAR is at least four times more costly than a 3m FAR. The object, therefore, was to establish acceptable radiated emission limits for the appreciably less costly 3m FAR.

The work undertaken consisted of a series of theoretical and experimental studies; a program of modelling and simulation supported the tests. Comparisons were made of various types of Equipment Under Test (EUT) on test sites including OATS SARs and FARs. The EUTs were specially designed to resemble real apparatus and to be stable thus ensuring consistency of results at the various test sites. For reference purposes some tests were performed on a modelled “ideal” OATS. Modelling also enabled many more variations of EUT to be examined than would otherwise have been practical. The work also included a study both on the improvement of the quality of absorber and the amount of RF Absorbing Material (RAM) required.

 The purpose of the theoretical part of the investigation was to quantify the measurement differences between OATS and FARs. The test results of simulated EUTs on both were compared in order to establish a limit level which ensured an identical pass rate for the two types of site.

A significant point of contention to arise from the study was the use of hybrid bicone/log antennas on 3m sites, the antenna being half as long as the test distance. Site evaluation could be measured using separate biconical and log periodic arrays which are considered to yield better accuracy, but this conflicts with the standard which specifies the same antenna for both site evaluation and emissions testing.

The now familiar issue of cable layout was addressed at some length. The use of ferrite clamps remains under consideration pending more practical experience. Simply attaching clamps on cables outside the volume did not prove to be satisfactory. It was concluded that the cables within the volume should either be adjusted taking the maximum of many positions, or preferably electronically switched using switchable clamp-on ferrites to effect a rapid variation of the termination to effectively eliminate any resonance nulls.

 The study identified a further potential for time saving associated with the FAR. This being that the absence of a ground plane eliminates the need for height scanning. It was recognised that emissions testing above 400 MHz could, depending upon the height of the product, require more than one elevation. The general perception was that for a product of 1.2m height two elevations differing by 0.35m should suffice. The question remains, however, as to whether the relevant site evaluation would need to be conducted at two elevations.

In general the project revealed close agreement between measurements completed at 3m in a FAR and on a 10m OATS. It has been demonstrated that the worst results are produced in a 3m SAR. In the absence of available land the 3m FAR was less costly than a 10m OATS and considerably less costly than a 10m SAR. Moreover any cost comparison should consider the fact that the FAR could, without modification, be used for immunity testing.  A cost comparison should not be made on the capital outlay alone, considerable savings can be made on running costs.  For example reduced measurement time and avoidance of repeat measurements because there is no ambient interference. Because of time savings it may be possible to dispense with a pre-compliance facility and go straight to full compliance testing.

Significant improvements over the existing EN50147-2 which have resulted from the prEN50147-3study include:

·   Removal of artificial ground plane.

·   Ability to accurately measure at 3m replacing the requirement for 10m.

·   Using an omni directional antenna to simulate possible EUT emissions in the test volume.

·   No longer having to height scan to avoid signal nulls. At higher frequencies where the EUT could give a null there is a possible requirement for two elevations.

·   Simulating the emission test by completing the site evaluation with the antenna in a fixed position overcomes a deficiency of EN50147-2. Whence the receive antenna is artificially moved in the horizontal plane during the site evaluation but not during the emission measurement.

The group have concluded that whilst a very good 10m OATS is expected to have a measured Normalised Site Attenuation (NSA) to within ±0.5 dB it is currently not possible to achieve this in a small 3m room especially below 150MHz. In the absence of a better absorber the room size would have to be increased. It was considered that this may, in future, be avoided when the real advantages of the FAR have been fully appreciated, but this would depend upon a relaxation of the existing NSA requirement of ±4dB. Another possibility is to allow the use of chamber correction factors below 150 MHz, dependent on product size. Within these limits the 3m fully anechoic room provides a versatile and relatively low cost EMC facility when compared to a 10m SAR or 10m OATS. Testing at a 3m range on either OATS or SAR was not considered as a viable alternative because of the high level of uncertainties. It is readily apparent that a FAR evaluated to the requirements of EN50142-2 would remove most of the advantages identified in this study. If the 3m FAR is to be accepted and the allied cost savings enjoyed then the recommendations and suggested limits of the prEN50147-3:2000 will need to be sympathetically examined.

At the last CENELEC SC210/A/WG4 meeting on 9 July it was voted to publish EN50147-3 as a Technical Report. Abstracts from this document form the basis for a standard being developed by CISPR sub-committee A, which it is hoped will be ready for publication in 2003.