Fred Weiss of President, Morgan/Weiss Technologies, Inc.

In days of old, microwave design was done using laborious hand calculations coupled with intuition and “cut and try” prototyping methods; the results were, and remain, impressive in their performance and their long-lasting impact on modern communication systems.

The 1980’s saw the advent of a number of commercial computer-based products for analyzing and simulating the behavior of electromagnetic structures, with the goal of easing and speeding the EM design task. Those first tools were able to handle structures of very limited complexity; they were difficult to use; and they were expensive. Fast-forward to 2018: Today we have a wide selection of such tools which are able to handle structures of considerably greater complexity than their earlier cousins, but so-called “big iron” EM tools can still be a) relatively difficult to use and b) typically quite expensive. (Has anyone priced an HFSS or Ansoft simulator seat recently?)

One technique the microwave community has used for many years is that of Mechanical Scale Modeling (“MSM”) in which electrically functional 3D-mechanical-scaled models of EM structures are used to predict the performance of the much smaller structures ultimately built into PCBs and hybrids. Typically used for relatively simple functional blocks, the technique is essentially an analog simulator that gives one the ability to quickly, accurately, and quite inexpensively develop microwave components using larger, more tractable dimensions and easier-to-measure scaled-down frequencies.

An interesting aspect of the MSM approach is that it can be applied to a broader set of problems than simple microwave component design. In particular, the author has found it to be a useful, inexpensive, and relatively quick-turn tool for a variety of PCB- and IC-related design tasks, including package modeling, inductor design, and even on-chip IC-level interconnect analysis. Although the author’s original work was done in the 1980’s, he has found the technique to be useful even up to the present day – particularly in budget-constrained environments.

This talk will focus primarily on the author’s experiences with the MSM technique as applied to IC and package modeling, including a brief review of the theoretical underpinnings of the method; its limitations; actual hardware models; and measurement techniques and test results. As time permits, we may also take a brief look at several “Lessons learned along the way” relating to the design and use of some familiar transmission line structures.

Biography:

Fred holds a Bachelor’s Degree in Physics from the University of Washington (Seattle), and MS and PhD degrees in Electronic Engineering from the University of California at Los Angeles. He has a lifelong interest in electricity in general, and data conversion circuit design in particular.

Over the course of his career he has had the pain and pleasure of designing high speed data convertor, RF, synthesizer, analog/mixed-signal, and RF power control products in Bipolar, CMOS and BiCMOS, SiGe, and GaAs MESFET IC process technologies. He was the originator and principal designer of the first commercial integrated gigasample digital-to-analog converter products, specifically the TriQuint GigaDACÔ line of 1 Gs/s 8- and 14-bit DACs built in GaAs MESFET technology and used for video and signal processing applications in civilian, military, and space borne systems. Also at TriQuint, he and Ty Bowman developed the second-known integrated 1 Gs/s, 6-bit ADC+Sample/Hold device at roughly the same time that John Corcoran and Ken Poulton at HP Labs reported theirs; however, since the TriQuint device was for a military application it couldn’t be published.

Along the way Fred has accumulated a fair amount of experience in applied electromagnetics, including among other things the use of 3D scale modeling techniques both to “simulate” and debug IC circuitry, as well as model package parasitics;.

Since finishing graduate school in 1981 he has been privileged to work in various design and management roles for Tektronix Laboratories, TriQuint Semiconductor (twice), Analog Devices NW Labs, Avnera, and most recently Keyssa (ex-WaveConnex). He has consulted for a variety of technology companies; is a cofounder of NIVasc, a local medical device company focused on non-invasive vascular monitoring; and although he tries to be retired he is president of Morgan/Weiss Technologies, a developer of IP for the high-speed memory module market.

At last count, he is named as the inventor or co-inventor of record on some 20 patents and has authored 16 professional papers or symposium presentations, as well as presented various seminar talks and short courses. He has been honored to teach data convertor and RFIC design short courses as a Visiting Professor at the Johannes Kepler University in Linz, Austria, and in times past served as an Adjunct Professor of Electrical Engineering at Oregon State University.

Fred is a Member of the IEEE and a former US Air Force Communications-Electronics Maintenance Officer.