All learning experiments

CLASSIC TUNING

Ziegler–Nichols PID tuning explained

Use reaction-curve and ultimate-cycle tuning responsibly, with assumptions and limitations made explicit.

Kp=0.6Ku,Ti=0.5Pu,Td=0.125PuK_p=0.6K_u,\quad T_i=0.5P_u,\quad T_d=0.125P_u

Ziegler–Nichols rules turn a few measured plant features into controller settings. They are historically important and useful as starting points, but often produce aggressive quarter-amplitude damping rather than a finished production tuning.

01

Two different methods

The reaction-curve method uses an open-loop FOPDT approximation. The ultimate-cycle method uses the gain and period of sustained closed-loop oscillation. Their inputs and risks are not interchangeable.

IDENTIFICATION VIEW 01

Read gain, delay and slope from a reaction curve

An open-loop step reveals the process gain and the delayed S-shaped rise used by the reaction-curve method.

Read gain, delay and slope from a reaction curveAn open-loop step reveals the process gain and the delayed S-shaped rise used by the reaction-curve method. Series shown: Measured response, Tangent estimate. Key insight: Identification quality sets the ceiling for tuning quality.0255075100-0.10.30.71.11.4delay LslopeNormalized time →Process output
READ THE PLOTIdentification quality sets the ceiling for tuning quality
02

Controller form matters

Published coefficients may assume an ideal or standard PID form. PID Loop Lab converts results into its documented parallel PIDF form instead of silently copying gains between incompatible equations.

ULTIMATE VIEW 02

Ku is the edge of sustained oscillation

At the ultimate gain, proportional-only control produces an approximately constant-amplitude cycle with period Pu.

Ku is the edge of sustained oscillationAt the ultimate gain, proportional-only control produces an approximately constant-amplitude cycle with period Pu. Series shown: At Ku, Below Ku. Key insight: This test deliberately approaches instability.02550751000.50.71.01.31.6period PuNormalized time →Process value
READ THE PLOTThis test deliberately approaches instability
03

Treat the result as a starting point

After applying a rule, test output limits, delay uncertainty, noise and disturbances. A robust operating point usually needs less aggression than the textbook rule.

TUNING VIEW 03

Classic Z–N is intentionally aggressive

Quarter-amplitude damping settles through repeated oscillations. Reducing the gains usually gives a calmer production starting point.

Classic Z–N is intentionally aggressiveQuarter-amplitude damping settles through repeated oscillations. Reducing the gains usually gives a calmer production starting point. Series shown: Classic Z–N, Softened gains, Conservative. Key insight: A rule returns a starting point, not a finished design.0255075100-0.10.30.71.11.4Normalized time →Process value
READ THE PLOTA rule returns a starting point, not a finished design