Can a control loop have both deadband and hysteresis at the same time?

Yes. A controller can be explicitly configured with a deadband around its setpoint while the sensor or output relay it uses also exhibits its own physical hysteresis. The two effects stack — the loop won't react within the deadband, and even once outside it, the relay or sensor may still apply its own direction-dependent threshold before actually switching.

Is hysteresis always a bad thing to be eliminated?

No. Some hysteresis is deliberately designed in (as with the two-threshold thermostat) precisely because it prevents rapid on/off cycling that would otherwise wear out contacts, valves, or motors. Problems arise when hysteresis is unexpectedly large, inconsistent, or undocumented — for example, a relay that has degraded and now has far more contact bounce/hysteresis than it did when new, causing erratic switching that wasn't part of the original design.

Does deadband only apply to on/off (bang-bang) control, or also to PID loops?

Deadband applies to both. In on/off control it means the actuator doesn't switch at all inside the band. In a PID loop, a deadband is sometimes applied around the error signal itself — the controller's output stops changing when error falls inside the deadband, even though the underlying PID math is still running, specifically to stop the output from chasing noise once the process is "close enough" to setpoint.

How would I tell, in the field, whether chattering is caused by a missing deadband or by sensor hysteresis?

Check the controller's configuration first — if no deadband is configured (or it's set very narrow) and the process variable is noisy right at setpoint, that's a deadband fix (widen the band). If a deadband is already configured and reasonably sized but the actuator is still chattering, look at the sensor or relay itself: a sensor with a large or drifting response, or an aging relay/contactor, is a hysteresis-related hardware issue that a deadband setting can't correct.

Is the "differential" setting on a thermostat the same thing as hysteresis?

Yes, in common usage. Thermostat "differential" or "swing" settings that specify separate turn-on and turn-off temperatures are describing configured hysteresis — the same concept, just using HVAC industry terminology rather than the general controls-engineering term.

← SCADA & Industrial Controls Studio

Concept Explainer · SCADA

Deadband vs. Hysteresis

Two different reasons a control loop "ignores" small changes — and why mixing them up leads to the wrong fix.

Both deadband and hysteresis stop a control loop from reacting to every tiny fluctuation, and both exist to prevent excessive actuator cycling — a valve, relay, or heater chattering on and off far more often than it needs to. That's where the similarity ends. Deadband is a single, symmetric "do nothing" zone the engineer deliberately configures around one setpoint. Hysteresis is a property — often physical and inherent to a sensor or actuator — where the switching threshold itself depends on which direction the input is currently moving. One is a band around a point. The other is a loop with two different points.