Glossary term
Dead Time
A delay between a change at a system input and the first observable effect at the output.
Definition
quantityA delay between a change at a system input and the first observable effect at the output.
Dead time is a pure or approximate transport delay that adds phase lag without initially changing magnitude. It is critical in process control, motion systems, networks, thermal systems, and sampled control because it reduces achievable bandwidth and can destabilize feedback loops.
Dead time is the interval after an input changes before the output begins to respond. In process control, it may come from transport delay through a pipe, sensor placement, mixing delay, analyser sampling time, actuator travel, computation, communication, or thermal inertia before a measurable effect reaches the sensor.
Engineering role
Dead time limits feedback performance. A controller acts on information from the past; if the delay is large, aggressive correction can arrive too late and drive the system into oscillation. This is why processes with large dead time are harder to control than processes with the same time constant but little delay.
Representation
In transfer-function models, ideal dead time is often represented by:
where L is the delay. First-order-plus-dead-time models are common in process identification:
where K is process gain and \tau is time constant. Since pure delay is awkward for some analyses, approximations such as Pade approximations may be used, but they add modelling assumptions.
Control impact
Dead time adds phase lag proportional to frequency. This reduces phase margin and limits the closed-loop bandwidth that can be achieved safely. PID tuning rules often become more conservative as the ratio of dead time to time constant increases. Feedforward, Smith predictors, model predictive control, or better sensor placement may be used when delay dominates.
Measurement
Dead time can be estimated from step tests, pulse tests, cross-correlation, timestamps, or process data. The value should be measured at the operating condition of interest because flow rate, line length, sample handling, network loading, and actuator dynamics can change it.
Common mistakes
Common mistakes include hiding dead time inside an apparent time constant, ignoring sensor and communication delays, and tuning a loop from simulation without including delay. Another error is assuming that faster computation fixes dead time when the dominant delay is physical transport or measurement location.