Deadband in Control Systems: Formula & Examples
Quick answer
Deadband is a range in which a change of input does not cause a corresponding output or state change. In an alarm, it usually separates the trip threshold from the return-to-normal threshold; in on/off control, it separates the turn-on and turn-off setpoints.
Key Takeaways
- Deadband is a range in which a change of input does not cause a corresponding output or state change. In an alarm, it us...
- Intermediate-level topic in Industrial Control Concepts
- Commonly used in: Analog alarm chatter prevention, Pump, heater and compressor on/off control
Detailed Definition
Deadband prevents chatter when a noisy or slowly varying process sits near a decision threshold. For a high alarm with a 100 °C trip point and 2 °C deadband, the alarm activates at the configured high threshold and does not clear until the value drops below the reset threshold defined by the implementation—commonly 98 °C. A low alarm normally clears after the value rises above its threshold plus deadband.
Deadband is not automatically the same as mechanical hysteresis, measurement resolution, a PID neutral zone or historian compression tolerance. Those concepts can all create a region of no visible response but have different causes and equations. State the exact algorithm, units, inclusive comparison and whether deadband affects activation, clearing or both.
Evidence and scope
Alarm-reset behavior was checked against current Rockwell Logix Designer ALMA documentation. “Deadband” is used differently in alarms, on/off control, motion and valve specifications, so every value must be reported with its algorithm, units and direction.
Technical review:
Critical behavior
- For a high alarm, deadband generally moves the return-to-normal point below the trip threshold; for a low alarm it moves it above.
- Rockwell ALMA documents that level-alarm deadband affects return to normal, not the transition into alarm.
- Deadband values must use the same engineering units and scaling basis as the compared process value.
- A wider band can reduce chatter but also delays clearing or increases process variation.
- Input filtering and time delay solve different problems; they must not be substituted blindly for deadband.
Verification checklist
- 1Record the algorithm, units, trip boundary, reset boundary and inclusive operators.
- 2Test rising and falling ramps through both boundaries.
- 3Test realistic sensor noise and the fastest credible process change.
- 4Verify alarm deadband, time delay and input filtering independently.
- 5Confirm HMI, historian and controller use the same scaled process value.
IEC and vendor terminology
Similar-looking instructions do not always have identical execution, initialization or storage behavior.
| Platform | Common term | What to verify |
|---|---|---|
| Alarm management | Alarm deadband | Difference between activation threshold and return-to-normal boundary; direction depends on high or low alarm. |
| On/off control | Switching differential | Separate on and off thresholds prevent rapid cycling of a pump, heater or compressor. |
| Valve / mechanics | Mechanical deadband or hysteresis | Input reversal may not produce motion until backlash, friction or linkage effects are overcome. |
| Historian / telemetry | Exception deadband | Suppresses reporting or storage of changes smaller than a configured tolerance; it is not a control threshold. |
High alarm with a separate reset boundary
PV rising: -----------------------> trip at 100.0 °C
ALARM: FALSE ----------------------------- TRUE
PV falling: <---------------- clear below 98.0 °C
|<----- 2.0 °C deadband ----->|Deadband calculations by use case
The same word describes several algorithms. Document the equation actually implemented.
| State / phase | Activation | Return / opposite action | Example |
|---|---|---|---|
| High alarm | PV ≥ HighLimit | PV < HighLimit − Deadband | 100 °C trip, 98 °C clear with DB=2 °C |
| Low alarm | PV ≤ LowLimit | PV > LowLimit + Deadband | 20% trip, 23% clear with DB=3 points |
| Fill pump | Level ≤ StartLevel | Level ≥ StopLevel | Start 30%, stop 70% |
| Historian exception | New sample differs enough | No state reset; publishing rule | Store when |PV − last| exceeds tolerance |
Working LD and ST example
Implement a high alarm with a two-degree reset deadband
Latch the alarm at the trip threshold and clear it only after the process moves below the reset boundary.
Ladder Diagram
|----[ Temperature >= 100.0 ]----------------(S) HighTempAlarm----|
|----[ Temperature < 98.0 ]------------------(R) HighTempAlarm----|Structured Text
IF Temperature >= 100.0 THEN
HighTempAlarm := TRUE;
ELSIF Temperature < 98.0 THEN
HighTempAlarm := FALSE;
END_IF;
// Trip = 100.0 °C; reset boundary = 100.0 - 2.0 = 98.0 °C.Expected result: Values between 98.0 °C and 100.0 °C preserve the previous alarm state, preventing repeated toggling near the trip threshold.
Failure modes and diagnostic checks
| Symptom | Likely cause | Check next |
|---|---|---|
| Alarm still chatters | Noise exceeds the band, units are wrong or multiple alarm evaluators use different values. | Trend the exact evaluated PV, trip and clear boundaries at source resolution. |
| Alarm never clears | Reset direction is reversed or deadband is too wide for normal operation. | Calculate the explicit reset boundary and test just above, at and below it. |
| Pump short cycles despite a wide band | Minimum run/off time, sensor turbulence or equipment feedback is the real issue. | Trend commands, feedback, level and timing; add justified time constraints separately. |
| Historian and HMI trends disagree | Historian exception deadband or sample rates hide intermediate movement. | Compare raw controller values with collector and historian configuration. |
Current primary and technical sources
These sources support the behavior summarized on this page. The project's controller, firmware and IDE help remain authoritative for implementation.
- Analog Alarm (ALMA), Logix Designer
Rockwell Automation — Level-alarm deadband definition, valid range and return-to-normal behavior
Continue with the practical guide
Common Questions
What is Deadband?
Deadband is a range in which a change of input does not cause a corresponding output or state change. In an alarm, it usually separates the trip threshold from the return-to-normal threshold; in on/off control, it separates the turn-on and turn-off setpoints.
When should I use Deadband?
Deadband is particularly useful in scenarios such as Analog alarm chatter prevention and Pump, heater and compressor on/off control. Consider implementing it when you need reliable, efficient solutions for these types of applications.
What should I verify before using Deadband?
Record the algorithm, units, trip boundary, reset boundary and inclusive operators. Test rising and falling ramps through both boundaries. Test realistic sensor noise and the fastest credible process change. Verify alarm deadband, time delay and input filtering independently. Confirm HMI, historian and controller use the same scaled process value.
What are related concepts I should learn?
To fully understand Deadband, you should also familiarize yourself with PID Control, Cascade Control, and Feedforward Control. These concepts work together in industrial automation systems.
Continue Learning
Ready to deepen your understanding of Deadband? Here are some recommended resources:
Was this helpful?
Let us know if this glossary term helped you understand Deadband better.
Your feedback helps us improve our glossary and create better content for the PLC programming community.
Quick Info
- Category
- Industrial Control Concepts
- Difficulty
- Intermediate
- Tier
- Important
About Industrial Control Concepts
Process control, PID tuning, and automation strategies