P. Penfield

De-Embedding and Unterminating

De-embedding is the process of deducing the impedance of a device under test from measurernents made at a distance, when the electrical properties of the intervening structure are known. Unterminating is the process of deducing the electrical properties of the intervening structure from a series of measurements with known embedded devices. The mathematical steps necessary for de-embedding and unterminating with theoretically redundant measurements in order to minimize the effect of experimental errors.

Education via advanced technologies

Advanced technologies like the World Wide Web offer interesting opportunities for improving higher education. A study done at the Massachusetts Institute of Technology (MIT) focused on these matters and made several specific recommendations. Since this study was completed, a new center was established at MIT to coordinate and promote the use of advanced technologies in education, and a high-level council on educational technology was formed. After briefly peering into the future, this paper describes the study and the new center.

Noise in Negative-Resistance Amplifiers

Negative-resistance amplifiers that consist of a noisy negative resistance imbedded in a lossless three-terminal-pair linear network to make a two-port amplifier are analyzed. It is shown that Haus and Adlers noise measure M_{e} , is not constrained to lie between two eigenvalues of a noise matrix, as is the case for bona fide two-port amplifiers. Instead, it is always equal to its optimum value, and is independent of the (lossless) imbedding network used. As a corollary of this, the noise figure of such an amplifier fails to equal its optimum value only insofar as the exchangeable gain is not high. The single value of noise measure may be computed from any simple lossless circuit at hand, or else from the exchangeable noise power of the noisy negative resistance.

The force density in polarizable and magnetizable fluids

The force density in polarizable and magnetizable fluids in nonuniform motion is evaluated in a way that is consistent with the special theory of relativity. The derivation is based on a generalization of the principle of virtual work. For application of the principle, it is necessary to know the energy density and power-flow density in a local rest frame of the fluid element under consideration. These are obtained from thermodynamic information and an extrapolation of the Poynting theorem applied to a rigid, nondeforming medium. Two different formulations of electrodynamics of moving and deforming media are compared. It is shown that they lead to the same force density if the same thermodynamic information is used in each of them. A simple model for the force density is obtained, starting from the E-H formulation.

CAFE-the MIT computer aided fabrication environment

The computer-aided fabrication environment (CAFE) is a software system being developed at the Massachusetts Institute of Technology (MIT) for use in the manufacture of integrated circuits (ICs). CAFE is intended to be used in all phases of process design, development, planning, and manufacturing of IC wafers. While still under active development, CAFE presently provides day-to-day support to research and production facilities at MIT with both standard and flexible product lines. CAFE provides a platform for work in several active research areas at MIT, including technology (process and device) computer-aided design (TCAD), process modeling, manufacturing quality control, and TCAD tool integration. The overall architecture and characteristics of the CAFE system are described. The significant problems solved and the design decisions made are identified, and some of the most important components of the CAFE system as it currently exists are discussed. The CAFE architectural framework supports a wide variety of software modules, including both development tools and on-line applications. The key components of the CAFE architecture are the data model and data base schema, the process flow and wafer representations, the user interface, and the application programming and database interfaces. >

Force on a current loop

Abstract Two models for a microscopic current loop are examined. In each case the force, which is related to the acceleration through Newtons law, is found to agree with the force on a magnetic-charge dipole of the same magnetic moment.

Hamilton's Principle for Fluids

Hamiltons principle is used to derive the force equation for both nonrelativistic and relativistic flow of a compressible, one‐species, inviscid fluid. Rotational flow is predicted by use of Lins constraint. Lins constraint is given a physical interpretation and found to be considerably simplified by use of four‐dimensional notation.

Principal values and branch cuts in complex APL

Complex numbers are useful in science and engineering and, through analogy to the complex plane, in two-dimensional graphics, such as those for integrated-circuit layouts. The extension of APL to complex numbers requires many decisions. Almost all have been discussed in detail in a recent series of papers. One topic requiring further discussion is the choice of branch cuts and principal values for the primitive APL functions that require them. Conventional mathematical notation and the experience of other computer languages are of only moderate help. For example, one cannot find in the mathematics or computer-science literature a definitive value for the principal value of the arcsin of 3. The extension of APL to the complex domain presents a unique opportunity to define a set of choices that will best serve APL and other languages. This paper recommends locations of all branch cuts, directions of continuity of the branch cuts, and values at the branch points. It also recommends that comparison tolerance be used in the selection of principal values. The results apply to APL, other languages, applications packages, and VLSI hardware for complex calculations.

Fourier coefficients of power-law devices

Abstract A definite integral often used in evaluating Fourier coefficients of power-law devices under sinusoidal excitation is 1 2π ∫ 0 2π (1+β cos ωt) α cos nωtd(ωt) where α is an exponent characteristic of the device, and β is an amplitude factor.In many practical cases this integral can be evaluated in closed form, and when it cannot, there is a rapidly converging series representation. This representation can be used to derive differentiation formulas and recursion relations, and is also ideally suited for numerical calculations. We discuss this representation, give differentiation and recursion relations, and list several of the closed-form solutions

Thermodynamics and the Manley—Rowe Equations

An unproved hypothesis is made: a physical system obeys the Manley—Rowe frequency‐power formulas if, and only if, it is reversible in the thermodynamic sense. A general proof of this hypothesis is not available, and the exact conditions under which it is true are not yet known, but it is demonstrated for a three‐frequency upconverter pumped sinusoidally and exchanging noise at the other two frequencies.