- On-Off Controller Definition: An on-off controller is defined as a control system that fully opens or closes the control element when the process variable crosses a preset level.
- Working Principle: The on-off controller works by fully opening or closing the output value, causing the process variable to change direction and cycle continuously.
- Example of Application: A typical example is the cooling fan control in transformers, which activates based on temperature levels.
- Dead Time: Practical systems experience a delay, known as dead time, between the control signal and the action.
- Ideal vs. Actual Response: The actual response curve of an on-off control system differs from the ideal due to the presence of dead time.
What is an On Off Controller?
An on-off controller switches its output between two defined states instead of adjusting it continuously. For a valve, those states may be fully closed and fully open. For a heater or fan, they may be off and on. The control system that applies this two-state action is called on-off control. The two output values may be 0% and 100%, but those values are not required by the control principle.
The controller compares the process variable with one or more switching thresholds. Whether it energises the output as the process variable rises or falls depends on the required control action and the process. Cooling usually turns on at a high temperature, while heating usually turns on at a low temperature.
When the process variable crosses a switching threshold, the controller commands the other output state. The process variable may continue in the same direction for a time because sensors, actuators and the process itself do not respond instantly.
Many on-off controllers use separate switch-on and switch-off thresholds. The gap between them is hysteresis, also called a switching differential. It reduces rapid switching near the setpoint. The output may alternate between its two states when the process needs repeated correction, or it may remain in one state while conditions stay on one side of the relevant threshold.
A common example of on-off control is transformer cooling. Load losses heat the transformer. When the temperature or calculated cooling demand reaches a configured start threshold, the control system can energise one or more fan banks on the electrical power transformer. The number of stages and their thresholds depend on the transformer and its cooling design.
Forced airflow increases heat transfer from the transformer. When the measured temperature or cooling demand falls below the assigned stop threshold, the controller de-energises the relevant fan bank. Separate start and stop thresholds, and sometimes time delays, help prevent short on-off cycles.
With the fans off, load losses may cause the temperature to rise again. If it reaches the start threshold, the controller energises the relevant cooling stage. The temperature will fall only when heat removal exceeds heat generation under the current load and ambient conditions.
An ideal diagram may treat switching as instantaneous and omit sensor, actuator and process delays. This simplification shows the two-state logic without modelling the equipment dynamics.
In real equipment, the sensor, controller, relay or contactor, actuator and process can each add delay or lag.
Dead time has a specific meaning: it is the interval after an input changes but before the observed output begins to respond. A gradual response after that interval is process lag and can be described by one or more time constants. Dead time and lag can both affect an actual response, but they are not the same quantity.
The following diagram illustrates one possible practical response. Its exact shape depends on the equipment and operating conditions.
Suppose the transformer temperature starts rising at time T O . The measuring instrument may not indicate the change immediately. Sensor placement, thermal inertia, signal filtering and controller scan time can all contribute to delay or lag. At time T1, the indicated temperature begins to rise.
At point A, the controller reaches the switch-on threshold and commands the cooling fans. After the switching and process delays represented by T2, forced airflow begins. The transformer temperature may continue rising during this interval. It begins to fall only if the cooling system removes heat faster than the transformer generates it. The curve is not necessarily exponential.
At point B, the controller reaches the switch-off threshold and commands the fans to stop. After the delay represented by T3, forced airflow stops. The subsequent temperature path depends on load, ambient temperature, residual cooling and the transformer’s thermal dynamics. It does not have to follow the same curve as the cooling part of the cycle.
N.B.: This illustration assumes that the loading condition of the electrical power transformer, ambient temperature, and other surrounding conditions remain constant. Real operating conditions can change during a control cycle.
