Control Systems
Feedback, transfer functions, time and frequency response, PID tuning, stability, digital control, anti-windup, feedforward and cascade, and troubleshooting, with simulated worked examples.
A control system keeps a quantity (a temperature, a speed, a position, a pressure) on target in spite of disturbances. Almost all of it rests on one idea, feedback: measure, compare with the target, correct. This guide takes that idea from the block diagram to a tuned loop in the plant: transfer functions and poles, the time and frequency responses, PID control and tuning, stability and margins, digital control and sampling, practical additions (anti-windup, feedforward, cascade) and troubleshooting. An oven with a dead time and a mass-spring servo run through the pages, and every response is simulated from the equations. For the hardware side see Motion Control and Actuators and PLC and Logic Control.
Start here: what do you need?
| I need to… | Go to |
|---|---|
| understand open loop, closed loop and steady-state error | Feedback Control Basics |
| turn a differential equation into a transfer function | Laplace Transforms and Transfer Functions |
| predict overshoot, rise time and settling time | Time Response and Specifications |
| tune a PID controller | PID Control and Tuning |
| read a Bode plot or check phase margin | Frequency Response and Bode Plots |
| check stability, or use Routh and root locus | Stability: Routh, Root Locus and Margins |
| choose a sample time or program a PID | Digital Control and Sampling |
| fix windup, add feedforward or cascade, handle a valve | Practical Control: Windup, Feedforward and Cascade |
| run a step test or find why a loop oscillates | Tuning and Troubleshooting Control Loops |
A suggested order
Read Feedback Control Basics, Laplace Transforms and Transfer Functions and Time Response and Specifications for the language. PID Control and Tuning is the working core, and Frequency Response and Bode Plots and Stability: Routh, Root Locus and Margins explain why it works and when it fails. Digital Control and Sampling and Practical Control: Windup, Feedforward and Cascade bring it into a real controller, and Tuning and Troubleshooting Control Loops is for the plant floor.
- Feedback Control Basics: Open and closed loops, the block diagram, error, steady-state error and the shared oven example.
- Laplace Transforms and Transfer Functions: From differential equations to transfer functions, poles and zeros, first- and second-order systems and block algebra.
- Time Response and Specifications: Rise time, overshoot, settling time and damping, and how poles set the shape of the step response.
- PID Control and Tuning: The three terms, Ziegler-Nichols and lambda tuning, derivative filtering and a complete oven-loop example.
- Frequency Response and Bode Plots: Gain and phase against frequency, bandwidth, gain and phase margins and loop shaping.
- Stability: Routh, Root Locus and Margins: Poles in the left half-plane, the Routh test, root locus and how delay erodes margin.
- Digital Control and Sampling: Sample time, aliasing, delay, the Z-transform, discrete PID and choosing a sample rate.
- Practical Control: Windup, Feedforward and Cascade: Saturation and anti-windup, feedforward, cascade, filtering, deadband and nonlinear actuators.
- Tuning and Troubleshooting Control Loops: A tuning procedure, common loop faults, stiction and oscillation, and a table of symptoms.
One oven, from a step to a stable loop
Worked example: K = 2 °C/%, τ = 120 s, θ = 15 s
- Model (Laplace Transforms and Transfer Functions): G(s) = 2 e−15 s ÷ (120 s + 1); a step test (Tuning and Troubleshooting Control Loops) recovers K, τ and θ.
- Tuning (PID Control and Tuning): the reaction-curve PI (Kp 3.6, Ti 50 s) overshoots by 18 percent; the lambda PI (Kp 2, Ti 120 s) by 5 percent and settles first.
- Margins (Frequency Response and Bode Plots): phase margin 61° for the lambda tuning against 26° for the other.
- Limits (Stability: Routh, Root Locus and Margins): the loop with P-only control oscillates at a gain of about 6.6 with a period of 57 s.
- Practice (Practical Control: Windup, Feedforward and Cascade): anti-windup cuts a 24 percent overshoot to 1 percent when the heater saturates, and 80 percent accurate feedforward cuts a disturbance dip by 80 percent.
Quick reference: the equations
| What you need | Equation |
|---|---|
| Closed loop | T = C G ÷ (1 + C G), S = 1 ÷ (1 + C G) |
| First-order step | y = K (1 − e−t ÷ τ); settling ≈ 4 τ |
| Second-order | overshoot = exp(−π ζ ÷ √(1 − ζ²)); ts ≈ 4 ÷ (ζ ωn) |
| PID | u = Kp [ e + (1 ÷ Ti) ∫ e dt + Td de/dt ] |
| Lambda tuning | Kp = τ ÷ (K (λ + θ)), Ti = τ |
| Ziegler–Nichols (reaction curve, PI) | Kp = 0.9 τ ÷ (K θ), Ti = 3.33 θ |
| Phase margin | PM = 180° + ∠L(jωgc); ζ ≈ PM ÷ 100 |
| Delay margin | PM (rad) ÷ ωgc |
| Nyquist | fs > 2 fmax; sample at 10 × per rise time |
| Steady-state error (type 0) | ess = r ÷ (1 + Kp K) |
Related guides and tools
- PID Tuner: four tuning methods and a response simulation.
- Motion Control and Actuators, PLC and Logic Control and Sensors and Field Devices.
- Electrical Power for the motor and the supply, and HVAC and Psychrometrics for building controls.
- Programming Languages for numerical methods (simulation of a loop is a few lines of code).
- The Downloads library (a controls formula sheet).
Safety first
- A control loop is not a safety system. Keep limits, interlocks, relief devices and emergency stops independent of the controller, to the applicable standards (IEC 61508, IEC 61511, ISO 13849).
- Test changes carefully: an aggressive tuning or a wrong sign can drive a process, a motor or a heater to its limits in seconds. Start with low gains and a safe output limit.
- Fail safe: decide what the actuator does if the signal is lost (a heater off, a cooling valve open) and set the controller’s sensor-failure behavior.
- Manual and bypass must work, and a controller left in manual hides a fault.
- The numbers here are for study and preliminary design; check against the real process and the plant’s procedures.
FAQ
Where do I start?
Read Feedback Control Basics, then go to PID Control and Tuning if you need to tune a loop now.
Do I need the mathematics?
To tune a loop, a step test and a rule are enough. To understand why it works and fails you need poles, margins and delay, which these pages explain with numbers.
Why use an oven for the example?
Because it has the features that make real loops hard: a lag, a dead time, saturation and noise. The same method applies to flow, pressure, level and speed.
Where are materials and machine design?
They are the next guides in this series.
References
- Ogata, K., Modern Control Engineering, Pearson; Franklin, G. F., Powell, J. D., Emami-Naeini, A., Feedback Control of Dynamic Systems, Pearson; Nise, N. S., Control Systems Engineering, Wiley.
- Åström, K. J., Murray, R. M., Feedback Systems, Princeton University Press (free online); Åström, K. J., Hägglund, T., Advanced PID Control, ISA.
- Seborg, D. E., et al., Process Dynamics and Control, Wiley; Franklin, G. F., et al., Digital Control of Dynamic Systems; IEC 61131-3 and IEC 61508.
Last reviewed: November 2026. Values come from the oven and servo models; check the real process and the plant’s standards before applying them.