Direct answer
A control loop is a set of devices that measures a process variable, compares it with a setpoint and acts on the process. The main types are open loop (no feedback), closed loop (feedback), cascade (two nested loops) and split range (one controller output driving two final elements).
Institutional content: Authomathika
Published:
Every industrial process has variables that must be held near a target: temperature, pressure, level, flow, density. A control loop is the arrangement of instruments and logic that does this job. Choosing the right loop type is a design decision that affects stability, safety and how easy the plant is to commission and maintain.
What is a control loop?
A control loop is a group of devices that measure a process variable, decide what to do about it and act on the process to keep the variable near a desired value. The desired value is the setpoint. The measured value is the process variable (PV). The correction sent to the actuator is the controller output (OUT).
In a plant, the decision is made by a controller that runs inside a PLC or a DCS. The same concepts apply to a standalone controller on a skid. What changes between loop types is how many measurements are used and how the controller output is applied.
What is the difference between open loop and closed loop?
An open loop sends a command to the process without checking the result. A closed loop measures the result, compares it with the setpoint and corrects the command continuously. That comparison is called feedback.
Open loop is simple and cheap, but it cannot compensate for disturbances. Closed loop needs a sensor and a controller, and it can become unstable if it is poorly designed or tuned. For a simpler walkthrough, see open loop vs closed loop control with simple examples.
| Aspect | Open loop | Closed loop |
|---|---|---|
| Feedback from the process | No | Yes |
| Reaction to disturbances | None | Automatic correction |
| Typical use | Timers, fixed speed pumps, sequences, manual valve positioning | Temperature, pressure, level, flow and density control |
| Main risk | Drift away from the target without anyone noticing | Oscillation or instability if poorly specified or tuned |
| Instrumentation needed | Actuator only | Sensor, controller and final element |
When should you use cascade control?
Use cascade control when a slow main variable is disturbed by a faster, measurable variable that sits between the controller and the process. Cascade control is an arrangement of two controllers in series: the primary (master) controller sets the setpoint of the secondary (slave) controller.
A common case is a heated vessel. The primary controller handles product temperature and sends its output as the setpoint of a secondary loop on heating medium flow or jacket temperature. A disturbance in the medium supply is corrected by the fast inner loop before it reaches the product.
Criteria for choosing cascade:
- The inner variable responds clearly faster than the outer variable.
- The inner variable can be measured reliably.
- The disturbance enters through the inner loop.
- The inner loop is stable on its own before the outer loop is switched to automatic.
- The control system supports the cascade mode and a clean bumpless transfer between modes.
Tune the inner loop first, then the outer loop. The details of tuning are covered in PID control: how the loop works and how to tune it.
What is split range control for?
Split range control uses one controller output to drive two or more final elements, each acting over a part of the output span. It is used when a single variable must be managed by two different actions, such as heating and cooling or pressure relief and pressure makeup.
In a reactor temperature loop, for example, the lower part of the output range may open the cooling valve in a decreasing way, while the upper part opens the heating valve. Around the middle of the range, both are closed. That middle region is a dead zone that avoids fighting between heating and cooling.
Points to define in the design:
- The output range assigned to each final element and the overlap or dead zone between them.
- The failure position of each valve, which must be safe for the process.
- The behavior at the transition point, where gain or valve characteristics can change.
- How the operator sees and overrides the two outputs on the HMI.
For valve behavior and characteristics, see control valves: what they are, how they work and types.
Which elements make up a control loop?
A closed loop has a sensor, a transmitter, a controller and a final element, plus the process itself. Each part adds delay, error or noise that the loop must tolerate.
| Element | Function | What to check |
|---|---|---|
| Sensor and transmitter | Measure the variable and send a signal | Range, accuracy, response time, installation point |
| Controller (PLC or DCS function block) | Compares PV with setpoint and calculates the output | Algorithm, scan time, modes, limits, alarms |
| Final element (control valve, VFD, actuator) | Applies the correction to the process | Sizing, characteristic, speed, failure position |
| Process | Responds to the action | Dead time, lag, gain, interactions with other loops |
| Signal path | Carries signals between field and controller | Wiring, shielding, grounding, network health |
Before start-up, every loop should be verified from field to screen. A field checklist for loop check and commissioning helps confirm range, direction of action and alarms. Practical examples for two very common variables are in temperature and pressure control loops.
At Authomathika, loops like these are part of the automation work, which is evaluated by scope and budget.
What mistakes are common when specifying control loops?
Most specification mistakes come from picking the loop type before understanding the process, or from treating the final element as an afterthought. These are the ones that show up most often:
- Using a closed loop with a slow or noisy sensor, which forces a weak controller action.
- Oversized control valves that work almost closed, leaving poor control resolution.
- Cascade loops where the inner loop is not really faster than the outer one.
- Split range with no dead zone, causing heating and cooling to fight each other.
- Wrong direction of action (direct versus reverse), which makes the loop run away.
- No defined failure position for the final element.
- Missing limits, anti-windup or manual and auto transfer rules in the control logic.
- Ignoring interactions between loops that share the same process.
A good specification lists, for each loop, the variable, range, setpoint source, mode behavior, final element, failure position, alarms and interlocks.
