Bad HVAC Sensors and Forgotten Overrides That Waste Energy

AI-generated cover image for illustration.

The dashboard shows a warm room, so the system sends more cooling. The room is already comfortable, but its sensor sits in direct sun. Elsewhere, a pump continues to run because a temporary override outlived the repair that required it. In both cases the equipment may be following its instructions correctly while consuming unnecessary energy.

A plausible reading or green status icon can look reassuring. Check what the building needs, what the controls believe and what the equipment physically does. Start with a few high-impact points before changing setpoints.

Original Nexora schematic. Conceptual only; not a construction or control detail.

Choose the points that can affect the most equipment

Begin with sensors and overrides affecting central equipment, then recurring problem zones. Outdoor temperature may determine economizer or heating enable; supply temperature can influence many zones; pressure may drive fan or pump speed. Rank their influence rather than alarm counts.

Rank candidates by consequence, duration, evidence and ease of verification. A central sensor driving equipment to an extreme may outrank a small room-sensor disagreement. Safety, critical conditions and equipment protection come first.

PNNL's research on self-correcting HVAC controls addresses temperature and humidity sensor faults, damper faults and persistent manual overrides. It supports treating these as identifiable operating problems, not simply accepting them as the inevitable behavior of an older building.

Log the equipment, point, suspected issue, supporting trend, current setting, owner and next check. Mark candidates unverified until testing establishes the cause; an analytics alarm is a lead rather than a diagnosis.

Separate calibration from a poor location

Calibration asks whether an instrument measures accurately under the conditions being checked. Location asks whether those conditions represent the space or air stream the controller is supposed to manage. An accurate sensor in direct sun can still be a poor source of room-temperature feedback.

Compare a suspect room sensor with a suitable calibrated reference instrument at the same location. Allow both to stabilize, follow their measurement procedures and consider their stated accuracy. Then compare conditions at representative occupied locations. Avoid holding a reference in your hand beside the sensor and treating the immediate difference as a reliable offset.

Look for sunlight, nearby heat sources, supply-air jets, restricted circulation, and air leaking through a wall penetration behind the device. Trane's zone sensor installation instructions illustrate why representative placement matters. Use the instructions for the installed model when deciding on a correction.

Avoid software offsets that conceal changing local influences. A sensor heated by morning sun may read normally later. Have the technician select and verify appropriate relocation, shielding, sealing or another physical remedy.

Check the reading all the way to the graphic

A field instrument can be accurate while the displayed value is wrong. Trace the physical device through the controller input, engineering units, scaling, software point and graphic label. Confirm which equipment and location the point actually represents. Similar names are common in repeated floors and copied graphics.

Check for unit conversion errors, an incorrect input range, a stale value, an averaging function or a substituted value after a communication fault. Record timestamps and quality flags where available. A line that stays perfectly flat through changing conditions deserves investigation, but a stable process can also produce a genuinely steady reading.

Use related readings as a plausibility check. Interpret mixed-air temperature alongside outdoor and return conditions, sensor locations and mixing quality. Stratification can mislead. A weather website provides context for an outdoor sensor, not a calibration reference at its location.

Humidity, pressure and airflow points need appropriate reference methods and competent personnel. Check temperature effects, tubing condition, probe orientation and operating range as relevant. Each measurement has its own verification requirements.

Distinguish an instruction from physical feedback

An on command shows that a controller requested operation. It does not prove that a fan is turning, a pump is circulating water or a valve reached its intended position. Equally, an off command does not prove the equipment stopped. Check what each status point is derived from before relying on its label.

Useful evidence can include an independently sensed run status, drive feedback, measured power, airflow, differential pressure or a temperature response. Each has limits. Motor current can show electrical activity without proving useful flow, and actuator position does not prove that a valve seals. Choose evidence that answers the particular question.

Berkeley Lab's monitoring-based commissioning plan template includes comparing controller outputs with actual conditions to identify failed devices. Apply that principle to a specific discrepancy. If a valve reports closed while the air warms across its coil, verify the thermal behavior and valve condition rather than editing the graphic until it looks consistent.

Qualified staff should follow isolation, access and testing procedures without defeating interlocks. Record the confirming observation so later operators can distinguish measured status from an echoed command.

Give every override an owner and an expiry

An override may have been a reasonable response to a failed sensor, a comfort complaint, a special event or commissioning work. Start by recovering its purpose. Ask whether the initiating condition still exists, what must be repaired, and who is responsible for returning the system to automatic control.

Build an override list from the controls system's supported reports and the team's records. Include equipment, point, forced value, reason, owner, start date, planned expiry and the automatic behavior expected after release. A local hand setting or controller-level override may require an authorized field check if it is not visible in the supervisory report.

Review unexplained and long-running overrides first, but avoid a bulk release. A forced fan or pump might be compensating for an unresolved fault. Removing it could interrupt a real service. Confirm that automatic operation is ready, arrange an appropriate test window and release only the approved item.

Watch the transition and subsequent sequence. Verify that normal demand drives the output and physical response without new alarms or comfort problems. If the test fails, restore the documented safe state and reopen the underlying fault.

Measure improvement without turning every alarm into savings

Preserve baseline trends, hours, weather and occupancy. After correction, compare equivalent conditions and check the wider response. Less reheat, fewer unnecessary hours or lower measured power can support the result, provided environmental requirements remain satisfied.

A sensor offset alone does not establish an energy saving. The consequence depends on which sequence uses that sensor, how often the error changes a decision, and how the equipment responds. Estimate avoided energy from measured power and time where possible, or use a transparent engineering calculation with its assumptions stated.

Avoid double counting related faults. A drifting zone sensor might cause both a cooling request and a manual override. If one repair resolves both, their energy estimates may overlap. Maintain one explanation of the operating change and one defensible total, rather than adding every alarm's estimated saving together.

Close a fault only when the evidence shows the sensor or control path is reliable and automatic operation behaves as intended. Save the before-and-after readings, instrument used, original and revised settings, approver and follow-up date. Keep edits limited to the authorized operating scope; network access and security settings are separate work.

The DOE's re-tuning guidance treats operating data as a basis for controls improvements. A good routine makes that data trustworthy first. For wider energy-use context, Nexora's Cool Savings guide is optional further reading. The essential maintenance habit is simple: verify the reading, understand the override, and confirm what the equipment actually does.