How do I calculate Kp, Ki and Kd for a PID controller?

Enter the setpoint, process variable and sample time in the PID Solver 360 calculator to evaluate u = Kp·e + Ki·∫e + Kd·de/dt. For plant-based design, identify a first-order-plus-dead-time (FOPDT) model and use the auto tuner (Ziegler–Nichols, Cohen–Coon, IMC/Lambda, SIMC, AMIGO or Tyreus–Luyben) to compute Kp, Ki and Kd, then simulate the closed loop.

How do I tune a PID controller online for free?

Open the solver, load an example or enter your process gain K, time constant τ and dead time L, then run the auto tuner. Compare the classical rules, inspect overshoot and settling time, and optionally let the optimiser refine the gains against IAE, ISE or ITAE. No install or sign-up is required.

What is the best PID tuning method — Ziegler–Nichols, Cohen–Coon or IMC?

Ziegler–Nichols is fast but aggressive (often 20–50 % overshoot). Cohen–Coon is better on dead-time dominant FOPDT plants. IMC/Lambda lets you pick the closed-loop time constant λ to trade speed against robustness. For most process loops, start with IMC, SIMC or Tyreus–Luyben, then verify margins in the simulator.

How do I tune a PID temperature controller?

Temperature loops are usually lag-dominant with some dead time. Load the electric-oven example, or enter your heater’s FOPDT parameters, apply IMC or SIMC with a conservative λ, enable anti-windup and a derivative filter, then check overshoot, settling time and phase margin before you deploy.

How do I reduce PID overshoot and settling time?

Lower Kp or increase Ti to cut overshoot; add filtered derivative (Kd) to damp the response and shorten settling time. Enable anti-windup if the actuator saturates. Use the optimiser with an overshoot or ITAE objective, and keep phase margin around 30–60° and gain margin above 6 dB.

What is a FOPDT model and why is it used for PID tuning?

A first-order-plus-dead-time model G(s) = K·e^(−Ls)/(τs+1) captures the gain, lag and delay of most self-regulating processes with three numbers. Almost every classical PID tuning rule — Ziegler–Nichols reaction curve, Cohen–Coon, IMC, SIMC and AMIGO — is derived from that model.

Should I use a PI or a PID controller?

Start with PI for flow, pressure and most process loops: it removes offset and stays quiet on noisy measurements. Use full PID when the plant has significant lag or dead time (temperature, many motion loops). Always filter the derivative and prefer derivative-on-measurement to avoid kick.

Does PID Solver 360 include Bode and Nyquist plots?

Yes. After a simulation the solver computes open-loop frequency response, gain and phase margins, maximum sensitivity Ms, and Bode/Nyquist charts so you can check robustness as well as the time-domain step response.

Can I identify a plant model from bump-test data?

Yes. Paste a CSV of time, input and output from an open-loop step. The Identify tab fits a FOPDT model (Sundaresan–Krishnaswamy) and you can apply K, τ and L to the solver, then auto-tune.

What is relay autotune for a PID controller?

Relay autotune drives the loop into a controlled oscillation to measure the ultimate gain Ku and period Pu, then applies Ziegler–Nichols or Tyreus–Luyben. Use it when you do not have a reliable FOPDT model.

PID Solver 360 Powered by Solver360

Launch solver

Solver360 · Control systems module

PID Controller Calculator, Simulator & Tuning Tool

A free online PID controller calculator, closed-loop simulator and tuning tool. Calculate Kp, Ki and Kd, run true closed-loop simulations, auto-tune with Ziegler–Nichols, Cohen–Coon or IMC, and export a full report — no install, no sign-up.

Launch the solver Identify from CSV

✓ RK4 numerical engine✓ Anti-windup & derivative filtering✓ 21 passing unit tests✓ Unit-tested tuning formulas

The control loop

Everything a PID workflow needs

Not a one-line “output = Kp × error” form. PID Solver 360 combines a calculator, a simulator, an auto tuner, an optimiser, visualisation, education and reporting into a single engineering-grade tool.

PID Calculator

Step-by-step Kp·e + Ki·∫e + Kd·de/dt with setpoint weighting, saturation and anti-windup — every term shown symbolically and numerically.

Closed-loop Simulator

RK4 integration of first-order, FOPDT and second-order plants, with transport dead time, derivative filtering and digital sample time.

Auto Tuner

Ziegler–Nichols (reaction curve & ultimate cycle), Cohen–Coon, IMC/Lambda and Tyreus–Luyben, each with its assumptions and equations.

Identify from CSV

Upload or paste bump-test data (t, u, y) to fit K, τ and L. Download a sample CSV template — processed entirely in your browser, no server upload.

Optimiser

A Nelder–Mead search finds Kp, Ki, Kd that minimise overshoot, settling time, rise time or IAE/ISE/ITAE, and compares before vs after.

Performance Metrics

Rise time, peak time, overshoot, settling time, steady-state error, peak error, IAE/ISE/ITAE, oscillations and windup — in a KPI dashboard.

Export & Report

Interactive graphs with zoom, pan and hover, plus CSV data, PNG charts, JSON configuration and a printable PDF engineering report.

Classical tuning, done for you

Enter a plant model and the auto tuner returns gains from every major method, each with the exact formulas and the assumptions behind them — so you know why, not just what.

Open the auto tuner

Ziegler–Nichols

Classic reaction-curve and ultimate-cycle rules — fast, aggressive quarter-amplitude decay.

Cohen–Coon

Tuned for dead-time dominant FOPDT plants; faster disturbance rejection than Ziegler–Nichols.

IMC / Lambda

Choose the closed-loop time constant λ directly to trade speed against robustness.

Tyreus–Luyben

A damped ultimate-cycle variant giving well-behaved, industrial responses.

SIMC / Skogestad

The process-industry default: pick τc ≈ L for a fast but robust PI.

AMIGO

Åström–Hägglund robust FOPDT PID — a little slower, much more forgiving.

Real-world engineering examples

Start from a realistic, pre-populated model. Each example opens the solver with matching plant parameters, gains, units and actuator limits — ready to simulate.

DC motor speed control

Armature-controlled DC motor, speed regulated by drive voltage.

first-orderSP 1500 rpmK 120

Temperature control (electric oven)

Resistive heater with thermal lag and transport delay.

fopdtSP 180 degCK 2.5

Tank level control

Gravity-drained tank with a control valve on the inlet.

fopdtSP 2.5 mK 0.8

Tank level (integrating)

Non-self-regulating tank — the inlet valve sets the rate of change of level, not the level itself.

integratingSP 3 mK 0.02

Flow control

Fast flow loop — small time constant, noisy, derivative usually off.

fopdtSP 45 L/minK 1.4

Pressure control (gas receiver)

Compressor feeding a receiver vessel with a bleed valve.

first-orderSP 6 barK 0.09

HVAC zone temperature

Air-handling unit serving a room with long thermal lag.

fopdtSP 22 degCK 0.45

Hydronic heating system

Boiler + radiator circuit, slow with a large dead time.

fopdtSP 21 degCK 1.1

Robot joint position control

Second-order servo axis — inertia plus viscous damping.

second-orderSP 0.5 mK 1

Vehicle cruise control

Longitudinal vehicle dynamics with drivetrain lag.

fopdtSP 27.8 rad/sK 0.6

Understand PID control

A PID controller continuously computes the error between a setpoint and the measured process variable and drives an actuator so the error goes to zero. The three terms answer three different questions: how big is the error now (P), how long has it persisted (I), and how fast is it changing (D).

u(t) = Kp·e(t) + Ki·∫e(t)dt + Kd·de(t)/dt

equivalently u(t) = Kp [ e(t) + (1/Ti)·∫e(t)dt + Td·de(t)/dt ] with Ti = Kp/Ki and Td = Kd/Kp

Proportional — Kp

Reacts to the present error. Raising Kp gives a faster rise time and a smaller offset, but too much gain makes the loop oscillate and eventually unstable. Pure P control always leaves a steady-state offset on a self-regulating process.

Integral — Ki (Ti)

Accumulates past error, so it keeps pushing until the error is exactly zero — this is what removes steady-state offset. The cost is added phase lag: too much integral action creates overshoot, slow oscillation and windup during saturation.

Derivative — Kd (Td)

Anticipates where the error is going. It adds damping, cuts overshoot and shortens settling time, but it amplifies measurement noise, so it is normally filtered (Tf = Td/N) and taken on the measurement to avoid derivative kick.

What happens if I increase Kp, Ki or Kd?

Change	Rise time	Overshoot	Settling time	Steady-state error	Stability margin
Increase Kp	Decreases	Increases	Small change	Decreases	Degrades
Increase Ki	Decreases	Increases	Increases	Eliminated	Degrades
Increase Kd	Small change	Decreases	Decreases	No effect	Improves (noise-sensitive)
Increase Ti	Increases	Decreases	Decreases	Slower removal	Improves
Increase Td	Small change	Decreases	Decreases	No effect	Improves then degrades

Trends assume the other two gains are held constant on a typical self-regulating process; for some plants (for example an integrating process, or derivative feedback on a first-order plant) the effect can reverse. Use the simulator to check your own model.

Free PID controller calculator and simulator engineers actually use

PID Solver 360 is a free PID tuning calculator and PID controller simulator for students and practising engineers. Compute the proportional, integral and derivative terms from the standard PID equation, switch between the Kp/Ki/Kd and Kp/Ti/Td parameterisations, and run a numerically integrated closed-loop simulation of your process. Auto-tune with the Ziegler–Nichols calculator, Cohen–Coon, IMC/Lambda or Tyreus–Luyben rules, then let the optimiser refine the gains for overshoot, settling time or IAE/ISE/ITAE.

Use it as a PID temperature controller tuning calculator, a FOPDT PID tuner, or a general closed-loop PID simulator with anti-windup, derivative filtering, Bode and Nyquist plots, and gain/phase margin. Every run reports rise time, overshoot, settling time and steady-state error — with a clear split between computed metrics and engineering suggestions.

Ready to tune your loop?

Open the solver, load an example or enter your own model.

Launch PID Solver 360

PID controller calculator FAQ

Long-tail answers for the searches engineers actually type — how to calculate PID gains, which tuning rule to use, and how to check overshoot, settling time and stability margins.