Slow Processes and Long Dead Time

The Core Idea

A gradual response and a delayed response are different process behaviors. Lag spreads a change over time; dead time adds an interval before the measurement begins to respond.

What you will learn

• Distinguish a process that moves slowly from one that responds after a delay.

• Recheck a tune when the process changes.

• Allow enough observation time before accepting or rejecting a response.

Why it matters

A controller can make more changes while the effect of an earlier one is still unseen. Understanding that wait helps you judge when another adjustment is warranted.

Try the same tune on three processes

Start with P = 3% output per °C, I = 0.3% output per (°C·s), and D = 0. Apply that same tune to each supplied preset. Before running, predict which process will make an immediate gain increase hardest to judge.

All outputs span 0–100%, but these are different processes with different starting temperatures and targets. Compare the response relative to each target and step span, rather than treating the height of two temperature traces as directly equivalent.

Run the original tune on each process, then change one gain at a time and restart. Finish 300 simulated seconds per comparison. Record your settings alongside the result; selecting another process resets its controls to the original tune.

Separate waiting from a missing response

During dead time, a changed output has not yet affected the measured process. Raising the command repeatedly during that interval can create a large response that arrives later. After the delay, the lag describes how gradually the process approaches its new condition.

The sluggish preset has a long lag and also includes a stated load event. A dip near 90 s is therefore not evidence that your gains suddenly changed. Observe its recovery after the load ends at 120 s as well as its first rise.

Use an observation window that fits the process

For a clean ending-error judgement, allow at least dead time + four time constants of an unchanged condition: 65 s, 125 s, and 34 s for these presets. These are observation guides, not promises that every tune will settle in that time.

The sluggish process needs additional time after its later load change; 300 s covers that recovery too. Simulator speed changes how quickly the display plays, while the controller still samples every 100 ms.

The response limits here are overshoot no greater than 10% of the start-to-target span, ending average absolute error no greater than 2% of that span, and settling within 110 s for the temperature and long-delay presets or 230 s for the sluggish preset. Read each condition separately.

Adapt the approach

For the long-delay process, try gentler P and I and wait for the effect to arrive. For the sluggish process, look for a sustained trend before deciding the output is doing too little. Make deliberate trials rather than repeatedly editing a running loop before its earlier commands become visible.

Hint 1 — Do not judge the first few seconds alone

At 5 s, the long-delay preset has still not shown the effect of its command. Inspect controller output (CV) as well as process variable (PV), keep the target fixed, and finish the run before comparing gain settings.

Hint 2 — A patient starting trial

For the long-delay process, try P = 0.4, I = 0.05, D = 0. For the training temperature process, try P = 1.2, I = 0.08, D = 0. Keep the units from the opening brief.

Worked comparison after your trials

On the long-delay preset, the original P = 3, I = 0.3 tune produces sustained large excursions. P = 0.4, I = 0.05 meets the stated limits and settles in about 37 s. A second passing option, P = 0.5, I = 0.04, settles more slowly, around 94 s.

The training temperature process meets the limits with P = 1.2, I = 0.08; the sluggish preset meets them with P = 2, I = 0.1. These different responses support different settings. A tune that works on one process is a starting hypothesis on another.

Check your understanding

Record one completed trial on each preset, including the gain settings and relevant delay or load event. Explain why you waited, what evidence prompted a gain change, and why your final choice fits that process. Then mark the lesson complete.

Enable JavaScript to run the embedded simulator and complete the interactive exercises.