Kevin Hinckley, Signal Integrity Engineer, Sun Microsystems
Douglas Winterberg, Signal Integrity Engineer, Sun Microsystems
Mike Ballou, RF and Microwave Applications Engineer, Agilent Technologies
Gustavo Blando, Principal Engineer, Sun Microsystems
Jason Miller, Senior Staff Engineer, Sun Microsystems
Roger Dame, Member of Technical Staff, Sun Microsystems
Gregory Truhlar, Signal Integrity Engineer, Sun Microsystems
Alexander Nosovitski, Signal Integrity Engineer, Sun Microsystems
Shelley Begley, Application Development Engineer, Agilent Technologies
Istvan Novak, Distinguished Engineer, Sun Microsystems

Recent advances in printed-circuit board density and increase of interconnect speed have raised the mainstream need for low-loss dielectric materials. Even though the choice of available laminates with medium and low loss is exploding, it comes with a confusing array of possible test methods for dielectric loss (Df) measurement. One of the primary challenges is that dielectric loss mostly matters at high frequencies where the dielectric loss cannot be measured independent of other contributors. We can measure various signatures of overall loss, which is the result of not only dielectric loss, but also copper resistivity, copper volume, copper surface roughness, radiation and potential non-homogeneities and non-isotropy of materials. There are three broad categories of measurement techniques: a) direct impedance measurements, b) resonance-based methods and c) wide-band model-based signature tests, several of them are now IPC standards. This paper introduces the model-based Capacitance Gradient Method (CGM), which is largely independent of conductive and radiation losses. The paper will also provide a comparative overview with a few selected methodologies, such as Split-Cylinder Resonance and Parallel-Plate methods, describing their strengths and potential pitfalls, together with real-life test data.