Micrologic 1960 Step 16 Circuit Design for Monolithic Integration

# Design Circuits for Monolithic Integration, Not for Discrete Parts: What a Circuit Designer Had to Unlearn

## 1. Why a Habit That Served Discrete Design for a Decade Now Costs a Designer Yield

This step names the methodology shift every circuit designer working with this series had to make: a habit built around discrete parts, where a resistor or transistor is bought to the tightest value and tolerance a design calls for, actively works against a monolithic layout, where matched components and ratio-based design beat any single diffused value's absolute accuracy. A discrete designer, free to order any resistor value from a catalog, specifies absolute values and expects them delivered. A monolithic designer, bound to Step 1's rule that a resistor is a shape rather than a part, cannot get an arbitrarily precise absolute value out of one diffusion run — but two resistors drawn side by side on the same die, sharing the same diffusion depth and the same sheet resistance, track each other far more precisely than either one tracks any fixed catalog number, because whatever varies run to run varies the same way for both of them at once.

where the ratio on the left cancels $R_{\text{sheet}}$ entirely, leaving only a geometric ratio a layout tool controls precisely, while the absolute value on the right still carries the full uncertainty of whatever that run's diffusion actually did — a designer who keeps reaching for the right-hand quantity is fighting the process; a designer who redesigns around the left-hand one is using it.

## 2. Real Diagram: Placing Matched Devices Next to Each Other on Purpose

A monolithic designer who wants two components to track together places them close together and oriented the same way on the die, because the diffusion and oxide variations this series has lived with since 1954 vary gradually across a wafer's surface — devices far apart see different local conditions, while devices placed adjacent see nearly identical ones.

## 3. Why This Methodology Shift Had No Equivalent in Any Earlier Series

Every series before this one, back to 1954, optimized a single device's own properties in isolation — the discrete designer's world this step describes is simply what the rest of the industry, outside this series, was still doing with the devices those earlier series produced. Step 11's decision to standardize on one NOR primitive made ratio-based matching especially valuable, since every gate's threshold now depends on the same load-resistor-to-input-transistor ratio repeated across a whole die, and getting that one ratio right once pays off in every copy of the gate rather than in one component at a time.

A circuit designer who kept specifying absolute values into this series' chips was not making a small error; they were applying a rule that discrete components rewarded and monolithic ones actively punish, and this step is the first place this project has had to say so directly, because it is the first series where the two design philosophies actually came into conflict.