Episode Description:
When a PID loop starts hunting, a valve keeps swinging, or a space refuses to settle at setpoint, changing tuning values can feel like the obvious next move.
But what if the PID loop is not actually the problem?
In Episode 564 of the Smart Buildings Academy Podcast, you’ll get a practical look at what PID loops are really doing inside building automation systems. Instead of treating PID settings like mysterious numbers, this episode helps you understand how the controller responds to what it sees and what the behavior of a loop can tell you about the system around it.
You’ll also explore why problems that appear to be tuning issues can originate somewhere else entirely. Understanding that distinction can save time, reduce unnecessary adjustments, and make troubleshooting more systematic.
Topics Covered
• What a PID loop is actually doing behind the scenes
• How proportional, integral, and derivative behavior affects a control loop
• Why hunting, overshoot, offset, and windup occur
• How trends can help you interpret loop behavior
• Why sensors, sequencing, outputs, and mechanical issues should be investigated before tuning
The goal is not to memorize tuning values. It’s to understand the loop well enough that when something goes wrong, you know what questions to ask next.
Click here to download or listen to this episode now.Podcast Video
PID loops show up throughout building automation systems. They control valves, dampers, pumps, pressure, temperature, and other processes that need to maintain a target condition.
Yet PID control is often treated as a set of numbers that should be left alone until something goes wrong.
Then a loop starts hunting. A valve swings open and closed. A discharge air temperature overshoots its setpoint. A space never reaches the desired temperature. Someone opens the controller, changes a tuning value, and waits to see what happens.
That approach can create more problems because not every control problem is a tuning problem.
Understanding what a PID loop is actually doing makes it easier to diagnose unstable systems, interpret trends, and determine whether the controller, sensor, sequence, output device, or mechanical system needs attention.
At its core, a PID loop compares two values:
The difference between those values is the error.
The controller uses that error to determine where its output should be. Depending on the application, that output might command a valve, damper, variable frequency drive, or another controlled device.
The controller does not understand the physical building. It does not know that a conference room is warm, a coil is undersized, or a valve is stuck.
It sees numbers.
PID tuning determines how the controller responds to those numbers.
That response is built around three components: proportional, integral, and derivative.
The proportional component responds to the current error.
The larger the difference between setpoint and actual condition, the stronger the controller responds.
If the measured condition gets closer to setpoint, the response becomes smaller.
This sounds straightforward, but proportional control creates an important tuning consideration. If the proportional response is too aggressive, the loop can repeatedly overshoot its target and correct in the opposite direction.
That behavior can appear as hunting on a trend.
If the proportional response is too weak, the system may respond slowly or struggle to maintain the desired condition.
Proportional control can also leave an offset between the actual condition and setpoint. That is where integral action becomes useful.
Integral control considers how long an error has existed.
Instead of responding only to the current difference between setpoint and measured value, it accumulates error over time.
If the system remains slightly away from setpoint, integral action continues adjusting the output until that persistent error is eliminated.
This makes integral control valuable, but excessive integral action can create its own problems.
The controller may continue building a correction while the physical system is still responding. By the time the measured value changes, the controller has already pushed the output farther than necessary. The result can be slow oscillations around the setpoint.
Integral action can also create windup.
Consider equipment that has been turned off while its control loop remains active. The controller continues seeing an error and keeps accumulating integral action even though the equipment cannot respond.
When the equipment starts again, the accumulated correction can drive the output aggressively.
Control loops should be coordinated with equipment status so the controller is not trying to control a process that cannot physically respond.
Derivative control looks at the rate of change.
Instead of focusing only on the current error or accumulated past error, derivative action considers how quickly the process is moving.
That can help dampen overshoot in certain applications.
The challenge is that derivative action is sensitive to noise. Sensor fluctuations and small changes in measured values can affect the controller's response.
For many building automation applications, derivative is not used and may be set to zero.
That means many BAS technicians spend most of their tuning efforts working with proportional and integral behavior.
PID terminology is not always implemented consistently across controllers.
One manufacturer may express tuning values differently from another. Controllers may use different units, ranges, or forms of PID control.
That creates a common troubleshooting risk.
A technician can understand PID theory and still make the wrong adjustment because the specific controller interprets the entered value differently than expected.
Before changing tuning parameters, understand how that particular product implements PID control.
Know the units. Know the PID form. Know how increasing or decreasing a parameter affects the controller.
Then verify the actual behavior.
A trend can turn an abstract tuning problem into something you can evaluate.
At minimum, look at values such as:
The shape of those values over time can reveal how the system is behaving.
Is the output constantly moving between extremes?
Does the process repeatedly overshoot setpoint?
Does the measured value move slowly while the controller output changes rapidly?
Does the output remain at 100 percent while the process stays far from setpoint?
Each pattern provides information.
Instead of looking at the PID values and guessing, look at the relationship between the controller's output and the physical system's response.
When making tuning changes, change one parameter at a time. Then observe the result.
Changing several values simultaneously makes it difficult to determine which adjustment caused the new behavior.
Not every system responds at the same speed.
A pressure control loop can react quickly. A space temperature loop may take much longer because the room, walls, furniture, occupants, airflow, and thermal load all influence the result.
That difference matters.
A controller that works well on a fast process may be far too aggressive for a slow process.
Tuning should reflect the physical process being controlled.
The goal is not to find universal PID values. The goal is to configure the controller so its response makes sense for the equipment and process connected to it.
A PID loop can only respond to the information it receives.
If the sensor is wrong, poorly located, or measuring something different from what you think it is measuring, the controller can appear to be poorly tuned even when it is responding correctly to its input.
Consider a space temperature sensor installed on an exterior wall, near heat-producing equipment, or in direct sunlight.
The controller may receive a valid electrical signal, but that value may not represent the actual space condition.
The same issue can occur with a discharge air sensor located where the air is not properly mixed.
Pressure sensors introduce another possibility. A damaged, pinched, disconnected, or leaking tube can produce misleading pressure readings.
Before changing PID values, confirm that the input is accurate and represents the condition you intend to control.
The controller may calculate the correct output while the controlled device fails to respond.
A valve actuator might receive a command without moving correctly. A damper could be mechanically stuck. An output could be scaled incorrectly. A command might indicate 50 percent while the physical device is doing something very different.
The PID loop cannot identify all of these conditions by itself.
Verify that changes in the controller output produce the expected physical response.
That connection between software and equipment is critical when troubleshooting control problems.
Two individually functioning loops can create a poorly controlled system when their sequences conflict.
Heating and cooling are a common example.
If heating and cooling setpoints overlap or are too close together, both systems may try to control the space at the same time.
One loop adds heat while another removes it.
From the controller's perspective, each loop may be responding correctly to its own inputs and setpoints. From the building's perspective, the system is wasting energy and creating unstable conditions.
Proper sequencing can prevent this.
Deadbands can create a range in which neither heating nor cooling is active. Another approach is to use a single control output that is split between heating and cooling so both cannot operate simultaneously.
When multiple loops interact, evaluate the sequence before assuming the individual PID parameters are wrong.
One of the most important lessons in PID troubleshooting is that a loop at 100 percent output does not automatically indicate a tuning failure.
It may indicate that the controller is asking for everything the system can provide.
A failed pump can prevent flow.
A closed isolation valve can stop water from reaching a coil.
A clogged strainer can restrict the system.
Chilled water may be warmer than the design condition.
A valve or damper may not physically reach its commanded position.
If the equipment cannot deliver what the controller is requesting, the PID loop may drive its output to 100 percent and remain there.
Changing tuning values will not fix that problem.
In this situation, the trend becomes a diagnostic tool. It shows that the controller is demanding additional capacity but the measured condition is not responding.
That points the investigation toward the mechanical system rather than the PID parameters.
Before changing PID values, answer four questions:
If those areas check out, tuning becomes the next logical step.
Then use trends to understand what the loop is doing. Observe the shape of the response. Make one adjustment at a time. Give the physical system enough time to react before drawing conclusions.
This creates a repeatable troubleshooting process instead of relying on trial and error.
A PID loop is not a mysterious black box.
It compares what you want with what you have, then determines how to move the output based on present error, accumulated error, and, when used, rate of change.
The tuning values teach the controller how to respond to the type of system it is controlling.
The trend shows you whether that response is working.
Once you understand that relationship, troubleshooting changes. Instead of immediately asking which tuning value should be changed, you can ask a more useful question:
What is causing the loop to behave this way?
Sometimes the answer will be proportional or integral tuning. Other times it will be a sensor, actuator, sequence, equipment condition, or mechanical limitation.
Understanding the difference is what turns PID tuning from guesswork into a practical building automation troubleshooting skill.
For a deeper discussion and insights from the field, listen to this episode on the Smart Buildings Academy podcast.