All learning experiments

PLANT DYNAMICS

First-order systems and time constants

Connect gain and time constant to a process step response and closed-loop tuning decisions.

G(s)=Kτs+1G(s)=\frac{K}{\tau s+1}

A first-order model describes a process that moves exponentially toward a new steady value. Its gain determines how far it moves and its time constant determines how quickly.

01

The 63.2 percent point

After one time constant, an ideal first-order response has completed about 63.2 percent of its total change. After roughly four time constants it is close to steady state.

TIME VIEW 01

One time constant means 63.2 percent complete

Changing τ stretches or compresses the same exponential shape without changing the final process gain.

One time constant means 63.2 percent completeChanging τ stretches or compresses the same exponential shape without changing the final process gain. Series shown: Fast τ, Medium τ, Slow τ. Key insight: τ controls speed; it does not control final value.0255075100-0.10.30.71.11.41τ = 63.2%Normalized time →Process output
READ THE PLOTτ controls speed; it does not control final value
02

Process gain

The steady change in output divided by the applied input change estimates K. Its sign also determines whether increasing the actuator raises or lowers the measured process.

GAIN VIEW 02

Process gain changes how far the output moves

The response shape stays first-order while K scales the steady change produced by the same input step.

Process gain changes how far the output movesThe response shape stays first-order while K scales the steady change produced by the same input step. Series shown: K = 0.5, K = 1.0, K = 1.5. Key insight: Estimate K from steady output change ÷ input change.0255075100-0.10.40.81.21.6Normalized time →Process output
READ THE PLOTEstimate K from steady output change ÷ input change
03

Why delay changes everything

Adding dead time does not change the final gain but postpones feedback. The controller acts without seeing the result, reducing the gain and bandwidth that remain safe.

DELAY VIEW 03

Dead time adds a silent interval before the rise

These plants have the same gain and time constant. Only the start of the response moves, but that waiting period strongly limits feedback control.

Dead time adds a silent interval before the riseThese plants have the same gain and time constant. Only the start of the response moves, but that waiting period strongly limits feedback control. Series shown: No delay, Medium θ, Large θ. Key insight: θ is not part of τ—measure both.0255075100-0.10.30.71.11.4dead time θNormalized time →Process output
READ THE PLOTθ is not part of τ—measure both