Midterm Review Comments - Lectures 17-21

Lecture 17: pp. 1-4 contained design and test strategies, mostly for the labs.

Basic Concept: p. 5 Equivalent circuit including wires (interconnections) allows us to compute RC time constants

Basic Concept: p. 5 rise/fall time and delay definitions are given here

Basic Concept: p. 6 Difference between lumped and distributed RC time constants. Common misconception discussed- RC time constants do not compute actual delays

Basic Concept: p. 7-7a Fringing fields greatly increase interconnection capacitance, and vary depending on the ratio of wire width to length, the ratio of metal thickness to oxide thickness and the ratio of metal width to oxide thickness.

Basic Concept: p. 8 Computation of parallel plate capacitance

Basic Concept: p. 8 Components of Gate capacitance include gate to substrate, gate to channel/depletion region and depletion region to substrate. A common misconception is that gate to source diffusion and gate to drain diffusion are significant and must be included. It is the gate to source side of the channel and gate to drain side of the channel that are important.

Basic Concept: p. 9 Gate capacitance varies with gate voltage and also with V

Lecture 18: This lecture continues capacitance discussion

Basic Concept: p. 1 capacitance computation in the linear region (different from previous semesters)

Basic Concept: p. 1 capacitance computations in the saturation region are two part - the source side of the channel and the drain side.

Basic Concept: p. 2 The junction capacitance depends on depletion region thickness and biasing voltage. We can compute charge stored

Basic Concept: p. 3 From charge and voltage we can compute capacitance. Calculations depend on whether the junction between p and n type material is abrubt (ideal) or graded (actual case).

Method: p. 4 diffusion capacitance formula depends on diffusion area, perimeter, and some K constants derived in the text which we give you.

Basic Concept: Capacitance depends on doping density so if channel stops are included around transistors, the capacitance of the diffusion next to the channel and to the substrate are computed differently than to the other side walls of the diffusion.

Method: p. 5 Resistance is usually computed by using resistance in ohms/square, along with the width and length of the material.

Basic Concept: p. 6 Setup, hold and clock to Q time are all important time parameters for the flip flop. We compensate for early and late clocks by adjusting setup and hold times.

Basic Concept: p. 6 The duty cycle is the % of time a signal is held high. The clock period is usually determined by setup time, propagation delay through and logic and clock to Q time.

Lecture 19: Basic Concept: p. 1 flip flop timing

Basic Concept: p. 2 Contamination Delay and a way to avoid it

Method: p. 4-6 Elmore delays (really Elmore RC time constants)

Method: p. 8 Using Elmore delays to check impact of resizing devices and interconnections. Common misconception: Widening devices and interconnections does not always make circuits faster

Lecture 20: Method: p. 2 -4 When to break long wires and insert buffers.

Method: pp. 4-5 Using RC time constants to find critical paths through a logic network

Method: pp. 6-7 Applying RC time constants to an unconventional gate - XOR gate

Lecture 21: Method: pp. 1-3 Using superbuffers to drive large capacitive loads. Sizing of superbuffer devices depends on ratio of Cd to Cg and whether there are long wires.

Method: p. 3 Some interconnection models

Method: p. 4 when to use inductance in delay computations

Basic Concept: p. 5 what causes crosstalk?

Method: pp. 5-6 ways to reduce crosstalk

Method: p. 6 Computing wire delays using lumped model

Method: p. 7 computing wire delays using distributed model. A common misconception is using either of these when a wire is embedded in a circuit. These formulae only work when wires are floating.

Basic Concept: p. 8-9 full scaling and how it affects all circuit parameters

Basic Concept: p. 8 problems with constant voltage scaling