Room for a View:
A discussion on emissions measurement

Many companies, have by now concluded that verification of the Electromagnetic Compatibility of their products may most advantageously be conducted in-house. The reasons usually advanced may include, convenience, cost and time savings, on site refining of designs and elimination of EMC problems, together with the progressive development of in-house skills.

David Mawdsley

Such being the case, an appreciable sum will probably have been invested in the training of operatives and in the acquisition of EMC measurement equipment accurate to fractions of a dB. However there still remains the problem of a suitable test site. For many of the ensuing measurements an Open Area Test Site (OATS) will have been selected. The ideal OATS would be in the middle of a vast open area with no source of reflections and a complete absence of any ambient noise or spurious radiations. In practice typical sites may be, in a nearby field, in the company car park, or even on the roof of one of the buildings, but in most instances the chosen site will fall well short of the ideal.

Given this scenario, it is a fact that organisations having the best of intentions, with trained and experienced operatives equipped with costly and accurate test equipment, may still experience a complete lack of test repeatability and in truth may be subject to measurement variations of as much as 20dB.

A recent visit and discussion with David Mawdsley of Laplace demonstrated that the frequently overlooked but essential part of such measurements are uncertainties due to the site. These can be reduced to two quite separate constituents; ambient noise and test site distortion. These two problems require individual and separate solutions. David pointed out that EMC radiated levels are generally in the region of 30 to 50 dB mV/m but that background emissions are commonly up to 80 dBmV/m this represents a background noise level 40dB (or 100 times) higher than the potential signal of interest.

Although this would seem to make matters impossible, in practice background signals are invariably narrow band and therefore each will occupy only one 120KHz IF bandwidth slot. There are over 8,000 such slots in the 30 - 1,000MHz range, so even in the unlikely event that there were 80 such background transmissions only 1% of the total bandwidth would be masked by this background. Hence the 'clever' EMC system should be able to acquire a stable signature of the background and to use this to help identify emissions from the EUT. This may be more difficult than it sounds due to the unstable nature of the background in most locations.

Laplace Instruments

There are of course alternatives to the OATS, screened rooms for example, these will certainly reduce background noise but the multiplicity of reflections and consequent standing waves highlights the second of the two problems, test site distortion which will render them useless for emissions measurements. Other options include Anechoic and semi anechoic chambers, and the various versions of test cell. Among the best known of these is the G-TEM this cell has well defined characteristics but is obviously limited to testing smaller objects. All of the above however can introduce errors.

The Laplace solution is to calibrate the site each time it is to be used and to correct the results accordingly. If this can be achieved then any major errors due to reflections, distortions, EUT cables, etc. can be quantified and corrected. Such calibration can be completed using an Emissions Reference Source (ERS) which comprises a small comb transmitter having a precisely calibrated output covering the appropriate range of frequencies, i.e. 30MHz to 1GHz. Software generated comparisons made between the measured 'test site' output and the calibration data provide a ready indication of site errors. Thus enabling the user to realise acceptable results from a relatively poor site.

The Laplace SA1000 analyser is currently unique in it's ability to automatically calculate these site correction factors and to apply the results to subsequent measurements.

An example of this is depicted in Fig 1 which represents results taken from a typical laboratory environment. The black line represents the uncorrected display taken from an ERS, the red line is the corrected plot. Comparison of the two plots reveals just how great the errors can be, in this example indicating around 20dB for frequencies above 550MHz.

Figure 1.

The technique becomes particularly relevant with the inevitable reducing cost of screened chambers and subsequent increase in their use. Using an ERS could ensure that the chamber characteristics remain defined and consistent, and would ensure that the results are able to be referenced to OATS measurements consistent with the current standards requirement. But for the majority of organisations who simply are unable to afford the cost of any 'proper' test site, then the use of an ERS could at least provide a degree of credibility and integrity.

Laplace produce an excellent range of cost-effective solutions to EMC Testing.

For full information contact David Mawdsley on 01692 500777