Issue link: https://iconnect007.uberflip.com/i/701250
34 The PCB Magazine • July 2016 This process is iterative and should be re- peated for all frequencies of interest to build a ɛ r versus frequency curve. In this effort this pro- cess was performed up to 110 GHz. Wideband circuit measurements at millimeter wave fre- quencies are very difficult to obtain accurately without building multiple designs and fine tun- ing for the frequencies of interest. Due to the lack of fine tuning in this study, some circuits demonstrated good wideband performance while others did not. Free Space Quasi Optical Extraction of ɛ r The free space quasi optical method is per- haps the most intuitive way to measure the di- electric properties of materials, as the method consists of projecting a transverse electromag- netic (TEM) wave through the material under test and recording the transmitted and reflected energy. The method is defined as quasi optical because the size of the optical components is small with regard to the wavelength and the design requires use of geometric optics [10] . De- spite this more obvious configuration, it is one of the most nontrivial due to the complicated mirror assemblies, frequency dependent beam size, and the non-ideal lossy mediums that are the unclad PCB materials under test. A typical configuration is presented in Figure 6. The free-space quasi optical system utilizes a two-port VNA connected to two corrugated feed horn antennas specifically configured for a particular frequency band (K-Band or W-Band for example). The horn antennas point toward mirrors which shape the radiated beam pattern into a Gaussian beam reflected toward the un- clad dielectric material under test. The anten- nas and mirrors are symmetric about the cir- cuit board material under test. These methods evaluates the change in magnitude and phase of the transmission (S21) parameters, and can yield in plane ɛ r and tan δ at frequencies within the band of interest. Note that any copper clad circuit board materials under test must have all copper removed before testing as this method only measures dielectric properties. Calibrating the VNA for these measurements required the following steps: 1) Isolation: blocking the beam propagation path with a metal plate to account for diffrac- tion effects at sample edges and multiple resid- ual reflections from the antennas. 2) Reference: measuring the through trans- mission (S21) parameters without the material under test placed in the sample fixture to ac- count for the permittivity contributions of air. 3) Time domain gating: mathematical elimi- nation of multipath signals in time domain us- ing the sum of the distance between the horn antennas and the dielectric sample (in this case +/- 2ns). Circuit board materials in this test were measured from 40 GHz to 60 GHz using the presented setup. The measurements and calcu- lations were accomplished with commercially available software and the resulting data is pre- sented in the results section. ɛ r was determined with relative ease but the determination of loss tangent was more difficult due to the thickness of the samples tested (~5 mil). Perturbation of Resonant Cavities to Measure ɛ r and tan δ The cavity resonator method is widely used as a way of characterizing dielectric properties of circuit board materials at lower frequencies. The nature of resonant methods makes them particularly useful for measuring both permit- tivity and tan δ with relative ease. Collecting data and calculating ɛ r is straight forward and requires only a suitable resonant cavity and a VNA. This method measures the in plane ɛ r and ROUND ROBIN OF HIGH-FREQUENCY TEST METHODS BY IPC-D24C TASK GROUP (PART 1) Figure 6: Quasi optical measurement system.