# @stem_antics on Instagram

- **Type:** Video
- **Original URL:** https://www.instagram.com/p/DbgNnQAAFvz
- **Gondola URL:** https://gondola.cc/posts/68532550-stem-antics-instagram
- **Thumbnail:** https://img.gondola.cc/tr:w-,h-,fo-auto/postThumbnails/60afb4825b.jpg
- **Posted:** 2026-08-01T15:59:11.000+00:00
- **Account Owner:** Stem Antics (@stem_antics) — https://gondola.cc/stem_antics

## Caption

Every stable machine you have ever touched is running an argument between three opinions about error.

That argument is PID control. Proportional, Integral, Derivative. It is the most widely deployed control algorithm on Earth, governing cruise control, thermostats, quadcopter attitude loops, reactor temperature and hard drive head positioning.

The setup is simple. You have a setpoint, the value you want. You have a process variable, the value a sensor actually reports. The difference between them is error. A PID controller converts that error into a command signal:

u = Kp e + Ki integral of e dt + Kd de/dt

Three terms, three tenses of time.

- Proportional acts on the present. Larger error, harder push. Used alone it always settles short of target, because zero error would mean zero output, and a heater holding temperature needs a non zero output to exist.
- Integral acts on the past. It accumulates residual error until the offset is driven out. This is what actually delivers accuracy, and it is also what makes the loop sluggish, because accumulation adds phase lag.
- Derivative acts on the future. It responds to how fast error is changing, applying brakes before overshoot arrives. It also multiplies sensor noise, which is why a great many industrial loops run as PI with the derivative gain set to zero.

Tuning is a trade between speed, overshoot and stability margin. Push the gains up and response sharpens until the loop oscillates. Back them off and the system turns docile and slow. Ziegler and Nichols published their empirical tuning rules in 1942, and engineers still reach for them today.

The elegance is that none of this requires a model of the plant. PID does not know whether it is steering a ship or heating a vat of resin. It only knows the error, and the three tenses in which that error can be read.

Good control is not brute force. It is knowing precisely how much to hold back.

#stemantics #controlsystems #engineering #automation #pidcontrol

## Stats

- **Views:** 2,560
- **Likes:** 82
- **Shares:** 0
- **Comments:** 1

## Tags

pidcontrol, automation, engineering, controlsystems, stemantics

---
Copyright (c) Gondola