Exploration of a More Conformal AUT Volume for Wide-Mesh PNF Sampling

Author: Scott T. McBride
Publication: AMTA 2025
Copyright Owner: NSI-MI Technologies

The approach of non-redundant near-field sampling has been available for many years. A general and automated approach that yields the expected time reduction for an arbitrary antenna volume, however, has been elusive. One of the more practical approaches is the “PNF wide-mesh” sampling, where the probe grid is separable in x and y, and this approach is the one explored in this paper.

A fundamental step in non-redundant sampling is to identify a volume that fully contains the AUT. Constraints imposed by theory have typically led this volume to be rotationally symmetric about a z-oriented line, and often also require that the volume be more spherical (less oblate) than a volume circumscribing the AUT. That larger volume generally results in more acquisition time than would a conformal volume, but allows those samples to be readily interpolated to the conventional half-wavelength PNF grid. This paper examines the impacts of relaxing those constraints in order to further reduce the required sampling time for a box-shaped AUT. It then looks for ways to reduce or remove those impacts.

The implementation of this algorithm involved a minor reformulation, specific to the PNF (or linear-axis) geometry, of the underlying non-redundant sampling theory. That reformulation is briefly described herein. A new family of tunable AUT-volume edge treatments similar to the existing “double-bowl” is also described.

The paper will show minor reductions in predicted acquisition time compared to non-redundant sampling with a circular double-bowl volume. Each non-redundant approach typically offers a 40-60% reduction with a rectangular AUT volume compared to a full conventional scan. A more notable advantage of the new approach is a significant reduction in preacquisition activity defining the several parameters that govern the non-redundant acquisition and processing.

View the paper