LSI technology—large-scale integration—pointed toward a major change in how digital systems could be designed. As semiconductor manufacturers learned to place more transistors and interconnected logic on a chip, computer functions once assembled from many components could be consolidated into compact arrays. For industrial automation, the important consequence was not simply smaller hardware: it was the possibility of using a programmable microprocessor for control tasks that might otherwise require custom logic.
This shift depended on advances in semiconductor fabrication, memory, and circuit design. It also changed the balance of work between hardware and software, creating a new way to think about digital control systems.
LSI technology changed the economics of logic design
Integrated circuits could make individual logic functions inexpensive when produced in large quantities. But custom arrays carried a different cost: designing and validating a particular pattern of interconnected transistors required substantial engineering effort. If production volume was limited, that nonrecurring development cost could outweigh the low per-unit manufacturing cost.
This presented a challenge for systems that needed specialized logic but not enough units to justify a custom array. Semiconductor manufacturers therefore had an incentive to make circuit structures more regular and reusable. Memory was a natural early focus because repeated cells could be arranged in predictable patterns. Read-only memory offered another route: a regular array could be configured to perform different functions by changing a mask during manufacture.
Microprogramming extended this idea. Rather than build every decision into a unique network of gates, designers could use stored instructions or control patterns to direct a system’s operation. The result was a more adaptable approach to implementing logic, though memory and chip economics remained important constraints.
From regular arrays to the computer-on-a-chip
Experience with memory arrays, programmed logic, and electronic calculators helped establish the feasibility of integrating more complex functions. The next step was the microprocessor: a nearly complete central processing unit (CPU) implemented on one or a small number of chips. Early devices were generally organized as 8-bit machines, but their significance lay in the architectural change they represented.
A conventional control design might use a collection of gates arranged specifically for one task. A microprocessor instead executes a program, allowing a designer to describe the required sequence of operations in software. The semiconductor manufacturer provides the core processing hardware; the system designer develops the program and connects the processor to the rest of the application.
This does not mean that hardware design disappears. The processor still needs appropriate interfaces, memory, and supporting circuitry, and the application must be engineered as a complete system. But for suitable tasks, a programmable device can reduce the need to create a new logic network for every function.
Why MOS integration mattered
The paper describes metal-oxide-semiconductor (MOS) devices as especially promising for complex arrays. Compared with bipolar logic, MOS inputs have very high impedance, so they can present less direct electrical loading to the circuits driving them. The source also notes that MOS switching speed can be constrained by parasitic capacitance—the unwanted capacitance associated with device structures and connections—rather than simply by carrier transit time.
In the period discussed, work on processing and circuit design aimed to improve MOS speed and switching behavior. Complementary MOS (CMOS), which pairs devices of opposite types, offered clearly defined logic states and low static power under standby conditions in the described circuit approach. The paper emphasizes that circuit performance and power depend on implementation, operating frequency, and switching activity; these observations should be understood in their historical technical context rather than treated as specifications for present-day components.
Integration density also had an economic dimension. Larger chips could hold more functions, but chip area affected the number of devices obtainable from a wafer and the likelihood of producing a working chip. Reducing the area needed per function and improving manufacturing yield were therefore central to making semiconductor memory and logic cost-competitive.
Implications for industrial automation
The paper identifies manufacturing automation and digital control as areas where low-cost microprocessors could be useful. Large computers, and even minicomputers, could be uneconomical for many individual control applications. A compact processor offered a way to apply computing to functions such as process control and self-diagnostics without requiring a general-purpose computer for every task.
For automation engineers and system integrators, the architectural implication is a choice between fixed-function logic and programmable control. A dedicated logic network may suit an application whose behavior is stable and whose production volume supports custom hardware. A microprocessor can be more attractive when the control sequence can be expressed in software and may need to be tailored or revised.
The shift also changes the skills involved in system development. The paper anticipates that logic designers would spend more effort writing and optimizing programs, while semiconductor manufacturers took on more of the underlying hardware design. In this model, efficient algorithms and careful use of memory become important design concerns—not just the number of gates.
Practical takeaway
LSI technology helped make programmable computing a practical option for a wider range of control applications. Its central lesson for industrial automation is architectural: when processing hardware becomes sufficiently compact and economical, designers can replace some application-specific logic with software running on a reusable processor. Selecting between those approaches requires attention to function, integration needs, development effort, and cost—not simply chip size or speed. The move from discrete components to integrated circuits, and then toward microprocessor-based systems, established a foundation for more flexible digital control design.