Why Analog Is a Different Engineering Problem Continuous Fidelity Not Binary Switching

# Why Analog Is a Different Engineering Problem: Continuous Fidelity, Not Binary Switching

## 1. Same Era, Same Fabs, a Different Problem

Part B opens with the connective tissue Part A set up: linear integrated circuits were built in the same era as the first commercial MOS logic parts, often inside the same companies — Fairchild Semiconductor's analog group was doing its own landmark commercial work through the mid-to-late 1960s while the company's own MOS alumni were out founding General Microelectronics and, soon, Intel. Same era, same fabs, often the same companies, genuinely different product categories, and this step states that relationship explicitly before describing what actually makes the category different.

A digital logic gate only has to land reliably on one of two states, and the entire discipline of noise-margin engineering exists to make that landing tolerant of real-world component variation. A linear amplifier has no such luxury: its output is supposed to track a continuously varying input signal with high fidelity, which means every transistor mismatch, every bit of thermal drift, and every stray capacitance shows up directly in the output rather than being absorbed by a margin built for exactly that purpose.

## 2. What This Means for Circuit Design

Because a linear circuit's output is judged on a continuous scale rather than a pass/fail threshold, analog designers through the 1960s spent much of their effort on problems digital logic designers rarely faced at all: matching pairs of transistors closely enough that their small differences canceled out, biasing circuits to stay stable across a wide temperature range, and compensating amplifiers so they did not oscillate when fed back on themselves. Those problems, not switching speed or density, are what the rest of Part B is actually about.