Savi LoopIQ official brand logo

Interactive engineering visualization

See process gain, lag and dead time.

Change one parameter at a time. Watch the predicted process-variable deviation respond to a synthetic controller-output step. This is an open-loop FOPDT demonstration, not a closed-loop PID tuner or an equipment interface.

Simulated PV deviation from baselineFOPDT · seconds
Synthetic first-order-plus-dead-time step responseThe process variable remains at its baseline until the input step and dead time have elapsed, then approaches the gain times input-step steady-state change with an exponential lag. Numerical milestones are listed below the chart.-1.41.053.55.958.40s9s18s27s36sΔPV (engineering units)Time (s)
Process responseOutput stepResponse onset63.2% point
Final ΔPV7
Response begins6 s
63.2% reached12 s

Adjust the model

1. Process gain determines direction and magnitude

For a steady step, ΔPV = Kp × ΔCO. If the gain or step is negative, the response moves in the opposite direction; if either is zero, there is no modeled response.

2. Dead time delays the first response

Until time stepAt + θ, this idealized PV deviation stays zero. A sensor filter or transport delay can appear as effective delay in a fitted model.

3. The time constant sets the approach rate

One τ after the response starts, an ideal first-order process has completed about 63.2% of its total change. After five τ it is about 99.3% complete.

Make the model earn your confidence

  1. Increase only θ. Explain why the onset moves but the eventual ΔPV does not.
  2. Double τ. Compare the time of the 63.2% point; explain why gain remains unchanged.
  3. Choose the negative-gain preset. Predict the response sign before looking at the graph. Explain why a negative steady-state gain is not an inverse-response (initial wrong-way) transient.
  4. State what this ideal model does not represent: noise, saturation, stiction, interacting dynamics, unstable or integrating behavior.

Engineering safeguard: Never infer a safe field test, controller action or valve fail position from this model alone. All results here are synthetic and ungraded.