E. Applied Systems

E.1 Engineering and Control

Course description

Engineering is where physics and mathematics meet cost, material limits, and human needs. The design process runs from requirements through prototype and test to iteration, rarely in a straight line. Henry Petroski's To Engineer Is Human is a readable entry point, built around failures: structures that collapsed because a design assumption turned out wrong under a load nobody tested for. A safety factor and a material choice hedge against that gap between an equation and the thing it describes.
If you already have engineering training, test out of this unit with the checkpoints and put the recovered hours into the build project instead of relearning circuits from scratch. Otherwise, learn to read and build simple circuits, reason about signals and digital logic, and separately learn statics and enough thermodynamics to see how an engine turns heat into work. Then build something that forces the disciplines together: a solar panel wired to charge a phone, or a small robot whose motors a microcontroller commands from sensor input. Tolerances and loose connections teach things no textbook chapter does.
Control theory is the formal core of making a system do what you want. Its first rigorous treatment is James Clerk Maxwell's 1868 paper "On Governors," explaining why the centrifugal governors regulating steam engines sometimes oscillated instead of settling. The modern textbook, Karl Åström and Richard Murray's Feedback Systems, is free from the authors and covers the same question with today's tools: differential equations, stability criteria, and the principle that negative feedback drives a system toward a set point. Work through the algebra for a simple case, then simulate it; a PID controller balancing an inverted pendulum is a standard exercise with interactive versions online.

Checkpoints

Each checkpoint is tagged with the stage of the cycle it tests.
StageCheckpointAttempts
formalism
Complete a build combining circuitry, code, and a mechanical part, and write a one-page failure log.Post the first attempt
formalism
Show algebraically that a proportional-only controller leaves a steady-state error for a first-order system, and that integral action removes it.Post the first attempt
skepticism
Take one documented structural failure, from Petroski or a public report, and state which design assumption broke down.Post the first attempt
abstraction
Map a thermostat, a cruise control, and mammalian temperature regulation onto the same feedback diagram, naming the sensor and actuator in each.Post the first attempt

Discussion

Proposed edits

Found an error, a missing idea, or a better source? Propose a change. A moderator reviews each proposal.