Mr. Elya Joffe

Founder and President of “Elya Joffe – Electromagnetic Solutions”, Ltd.

On the Usage of Statistical Methods for Analyzing
the Significance of Sample Shielding Effectiveness Test Results

ABSTRACT: Shielding is over of the most common and useful EMI control techniques and is often used in the design of equipment enclosures. Verification of shielding is rarely carried out as metallic surfaces are assumed to provide sufficiently high suppression. However, in modern designs, two aspects cannot be neglected :
(a) the shielding effectiveness of composite materials
(b) the consistency of the measurement results.

In this lecture, the second aspect will be addressed. In particular, the uncertainty of the measurement results will be considered with respect to:

  • the IDEAL (i.e., “correct”) value
  • the uncertainty due to variations in the design parameters (i.e., parameter distribution vs. the ideal)
  • the uncertainty due to the nature of shielding measurement (i.e., measurement uncertainty.
In addition, when validation tests are carried out across a large frequency span, and  statistical data (e.g., average, standard deviation, confidence interval, etc.) is derived (per frequency) across the frequency band, “curve fitting” adds “statistical uncertainty”. This last item, often neglected, is the main focus of this lecture.