What a Setpoint Is and What It Is Not

Why this matters

A setpoint is the most-quoted number on any job and the least-documented. Customers repeat it, dispatchers write it on tickets, and techs adjust it, and almost nobody records the three facts that make it mean anything: where it is measured, under what condition it was set, and how far the result is allowed to sit from it. Without those, a setpoint is a number that two people can both read correctly and completely disagree about, which is how a system that is working exactly as designed generates a callback.

Before you move one

Changing a setpoint changes how close the equipment runs to its protective devices, and that is the part people skip.

• Establish the margin before you raise a target toward a limit. Read the limit's trip point, read the current operating value, and know the gap you are eating into. A setpoint raised into a limit produces nuisance trips that get diagnosed as a bad limit weeks later by somebody else.
• Never raise, bypass, jumper or replace a limit or an interlock to make room for a setpoint. If a protective device has been opening, the required first step is establishing why it opened. Replacing a correctly-operating limit reaches the same end state as jumpering it, one step slower and with a part number on the invoice.
• Where you must reach a setting inside an energized control enclosure, 29 CFR 1910.333(a)(1) permits energized work only where de-energizing introduces additional or increased hazards or is infeasible due to equipment design or operational limitations; use a meter and leads rated CAT III at or above the circuit voltage, and work to the boundaries and protective equipment NFPA 70E-2021 assigns the task. Where the panel can be dead, open the disconnecting means, lock and tag it under 29 CFR 1910.333(b)(2) in general industry or 29 CFR 1926.417 in construction, and prove dead per NFPA 70E-2021, 120.5.
• On combustion equipment, wear a personal carbon monoxide monitor before the appliance fires, and keep it on you for the whole visit rather than clipping it to the toolbag. A setpoint change that alters firing rate or cycle length changes combustion behaviour, and that is exactly when you want the monitor on your person.

What a setpoint actually is

A setpoint is a target value for one specific measured quantity, at one specific measurement point, under a stated set of conditions. All three qualifiers are load-bearing.

It is not a promise about any other point in the system. A discharge air setpoint promises something about the air at the sensor, not at the diffuser sixty feet away. A supply water setpoint promises something at the sensor well, not at the last terminal on the branch. Every foot of distance between the controlled point and the point the customer experiences is a place where the number changes for reasons the loop has no ability to see or correct.

It is also not a promise the equipment can keep. A setpoint is a request. Whether the request is achievable depends on capacity against load, and a loop asked for something it cannot deliver runs its output pinned and reports nothing wrong, because from the loop's point of view nothing is wrong: it did everything it could.

The four things that get called setpoints and are not

The record a setpoint needs to be usable

This is the artifact. Nine fields, and every one of them exists because leaving it out has produced a callback.

• Controlled variable - what is being held (temperature, pressure, level, flow, concentration).
• Measurement point - the physical location, described well enough that a different tech puts an instrument in the same spot. "Supply duct, 8 ft downstream of the coil, center of duct" is a location. "Discharge air" is not.
• Target value, or the schedule that produces it - if the target is reset off another variable, write the two endpoints of the schedule, not the current output.
• The condition it was set under - outdoor condition, load, occupancy, product, whatever was true at commissioning. A target set at design load and one set on a mild day are different numbers and they will not both work.
• Acceptable deviation - the band inside which the system is performing correctly. Without this field, any deviation is a complaint.
• Differential or deadband, and minimum on and off times - what the controller is allowed to do around the target.
• Related limits and their trip points - so the next person knows the margin they are working inside.
• Interlocks that override this setpoint - the permissives that can hold the loop off entirely and make it look dead.
• Change authority and step size - who may change it and by how much at a time. A 2 F step with a review after one full load cycle is a reasonable default for a comfort temperature loop; tighten it on a process where product is at stake, and state the step in the record rather than leaving it to judgment.

The filled-in record: one air handling loop

Controlled variable: supply air dry-bulb temperature. Measurement point: supply duct, 8 ft downstream of the cooling coil, probe at duct center, downstream of the fan. Target: reset on outdoor air. 55 F supply at 75 F outdoor, rising to 65 F supply at 50 F outdoor, straight line between them. Condition set under: commissioned at 78 F outdoor, building at design occupancy, all zone boxes proven modulating. Acceptable deviation: plus or minus 1.5 F from the scheduled target, in steady state, at least 10 minutes after any load step. Differential: modulating output, no deadband; minimum valve position 0%, minimum stage-off time 5 minutes on the compressor stage. Related limits: low supply temperature cutout, and a freeze protection device on the coil. Read and record both trip points at every service visit. Interlocks: fan proof of flow, filter pressure alarm, smoke detector shutdown. Any of these open and the loop is not the story. Change authority: the mechanical contractor of record; 2 F per step, reviewed after one full load cycle.

Now run the customer's complaint through it. They say the setpoint is 55 F and the unit is blowing 61 F, and they want to know why the equipment is not holding its setpoint.

Read the schedule. The slope is 10 F of supply change across 25 F of outdoor change, which is 0.4 F of supply per degree of outdoor. At 60 F outdoor, the scheduled target is 65 minus 0.4 times the 10 degrees above 50, so 65 minus 4, or 61 F. Check the far end to make sure the arithmetic is not backwards: at 75 F outdoor, 65 minus 0.4 times 25 is 65 minus 10, or 55 F, which matches the schedule's own endpoint. The unit is holding its target exactly. The customer is reading a number that was true on a design day against a target that moves every hour.

Then go one step further, because the interesting half of this call is not the schedule. The customer does not stand in the supply duct. They stand under a diffuser at the far end of a run through an unconditioned ceiling space. Air gains temperature along that run, and how much depends on the duct's insulation, the temperature of the space it passes through, and the velocity, so it has to be measured on that job rather than assumed. Measure it: put a probe at the diffuser and read the difference against the duct sensor at the same moment, at steady state. If that measured gain turns out to be 4 F, the customer is feeling 65 F air while a perfectly-functioning loop holds 61 F at the point it was told to hold.

That is the whole lesson in one call. The setpoint was honored. The record made it possible to say so in one visit instead of three, and it also identified the real work: the duct run, not the controller.

What would flip the conclusion. If the measured gain along that run were near zero and the diffuser still read high, the air is being mixed with something warmer before it arrives - an open bypass, a leaking damper on a return path, or a terminal unit reheating when it should not - and now the loop's target genuinely is not reaching the space.

The failure mode. Lowering the schedule's endpoints to make the diffuser read what the customer wants. It works for a week. It also drags the coil colder for the same load, pushes the low-temperature cutout's margin down, and typically shows up in mid-winter as a freeze protection trip that nobody connects to a summer setpoint change.

Confirming a setpoint is actually in force

Two checks that catch the two ways a documented setpoint stops being the real one.

Read the live target, not the configured one. On any controller with a reset schedule, an override, or a scheduled event, the number in the configuration screen and the number the loop is currently comparing against are different fields. Find the one the loop is using and record both. A great many "the setpoint drifted" complaints are an override left active from a previous visit.

Compare the record against the controller at every visit, and date the comparison. An undocumented change is invisible until somebody has the old value to compare with. If the record and the controller disagree and nobody can say who changed it, that is the finding, and it belongs on the ticket rather than being quietly corrected - a setpoint that moves without a trail will move again.

References

• 29 CFR 1910.333(a)(1) and (b)(2), OSHA general industry work practices for energized and de-energized electrical work
• 29 CFR 1926.417, OSHA construction lockout and tagging of circuits
• NFPA 70E-2021, 120.5, verifying an electrically safe work condition
• See related: What a Control Loop Actually Is; Why Actual Never Equals Setpoint