HVAC preventive maintenance program field guide

What HVAC preventive maintenance is and why it pays

HVAC preventive maintenance is the scheduled work you do on running equipment to keep it efficient, reliable, and at its full service life. You change the filter, clean the coil, check the charge, tighten the connections, and verify the controls on a calendar, before any of it fails. The alternative is reactive maintenance, where you wait for the call that the building is hot and then pay for the emergency. PM moves the cost from the breakdown to the schedule, and the schedule is always cheaper.

The money case is not subtle. Reactive repair runs roughly three to five times the cost of the same work done on a planned visit, once you add the after-hours rate, the rush part, the secondary damage, and the lost comfort hours in an occupied building. A compressor that dies from a plugged condenser coil costs many times what the coil cleaning would have. The PM is the insurance you actually collect on.

PM also holds efficiency. Equipment does not fail all at once. It degrades. A dirty filter starves airflow, a fouled coil loses capacity, a slipping belt moves less air, and the building gets the bill in kilowatt-hours long before anything trips. A unit that is never maintained can drift well off its rated efficiency within a few seasons and nobody notices, because it still blows cold air. The PM program is what keeps the equipment running at the performance the owner paid for, and it is what keeps a service contract profitable instead of a string of emergencies.

What is the difference between reactive, preventive, and predictive maintenance?

There are three maintenance strategies, and most buildings run a blend of them whether they planned to or not. Reactive is run-to-failure: you fix it when it breaks. Preventive is scheduled: you service it on a fixed interval regardless of its condition that day. Predictive is condition-based: sensors and trending watch the equipment and you act when the data says a fault is developing, not when the calendar says so.

Reactive has its place. On a cheap, redundant, non-critical piece of equipment, running it to failure can be the right economic call. The trouble is using it as the default for everything, because the failures never arrive at a convenient time and the collateral damage is where the cost hides. A bearing that seizes does not just cost a bearing. It can take the motor and the shaft with it.

Preventive is the workhorse and the subject of this guide. It is predictable, it is schedulable, and it catches the slow degradation that reactive never sees coming. Its one weakness is that fixed intervals do some work that the equipment did not yet need, which is why predictive is the direction the industry keeps moving. Predictive does not replace PM. It tells you where to spend the PM hours and where you can stretch the interval, by trending vibration, motor current, temperatures, and pressures over time. Start with a solid preventive program, then layer condition monitoring on the critical and expensive assets where the sensors pay for themselves.

The PM schedule and how often the work comes due

A PM program is a calendar of tasks at different frequencies, not a single annual visit. Some work comes due monthly, some quarterly, some twice a year, and some once a year. The art is sorting each task into the right bucket so the high-impact items get done often and the deep work gets done on a sane cycle.

Filters are the most frequent line. Many commercial units want a filter check monthly and a change every one to three months depending on the filter and the dirt load, though the manufacturer's data and the building's conditions move that number. Belts and drives, drain cleaning, and a quick operational check fit a monthly or quarterly rhythm. Coil cleaning, full electrical checks, control verification, and the seasonal startups land on a semiannual or annual cycle.

The two anchors of the year are the seasonal startups. Cooling startup in the spring gets the equipment ready before the first hot week, and heating startup in the fall does the same before the first cold snap. Those are the two visits that catch the failure before the building feels it, because the worst time to find a dead compressor is the day the load arrives. ANSI/ASHRAE/ACCA Standard 180 frames maintenance around inspection tasks and intervals by equipment type, but it sets minimums; the manufacturer's instructions and the equipment's duty and environment set the real frequency.

Why is the filter change the most important PM task?

The filter change is the highest-impact, most-skipped task in the program, and it is first because everything downstream of it depends on airflow. A plugged filter starves the system of air. The blower works harder for less flow, the evaporator coil gets too little air across it and can drop below freezing and ice up, and on the heating side the heat exchanger overheats and short-cycles on its limit. One neglected filter quietly degrades capacity, efficiency, and equipment life all at once.

Match the filter to the system, not to the box that fits. MERV rating sets how fine the media is, and a higher MERV catches smaller particles but adds pressure drop, so a filter that is too restrictive for the equipment starves it as surely as a dirty one. The design and the unit's available static pressure decide the MERV ceiling. Use the right rating, seat it square so air does not bypass around a loose filter, and write the size and rating in the record so the next tech does not guess.

The schedule is where filters get lost. Monthly is a reasonable check interval for most commercial units, with the change driven by actual loading, every one to three months in a clean office and far more often in a kitchen, a renovation, or a dusty plant. The honest tell that a program has gone soft is filters that are gray and bowed in the rack. If you find that, the coil behind it is already getting dirty, because the filter is the coil's protection and a failed filter passes the load straight to the coil. The air handler guide covers the filter section in the context of the whole airflow path.

Cleaning the evaporator and condenser coils

Dirty coils cost capacity and efficiency, and they do it gradually enough that nobody connects the high energy bill to the fins. The condenser coil rejects heat outdoors. Foul it with dust, cottonwood, and grease and it cannot reject heat, so head pressure climbs, the compressor runs hotter and draws more current, and on a bad day it trips on high-pressure or fails outright. The evaporator coil absorbs heat indoors. Coat it in dirt that gets past a weak filter and it loses capacity, runs colder, and can ice over and dump water.

Clean the condenser at least seasonally in most climates, more in a dirty environment, with a coil cleaner suited to the metal and a gentle wash that does not bend the fins. High-pressure water flattens fins and a flattened fin blocks air as well as dirt does, so the pressure and the angle matter. Comb out fins that are already bent. On the evaporator side, inspection and cleaning are the only way to catch fouling early, because it lives inside the cabinet where nobody looks until the capacity is gone.

The number that exposes a dirty condenser is the head pressure and the corresponding condensing temperature, run against the outdoor ambient. A condensing temperature far above ambient with a coil that looks clean from the front often means it is packed solid in the second and third rows where you cannot see. The capacity that walks away from a fouled evaporator is covered in the air handler guide, which treats the coil as part of the conditioning path.

Belts, bearings, motors, and the drive

Belt-driven fans need the belt and the drive checked, and this is straightforward mechanical PM that gets skipped because the unit still runs with a bad belt, just worse. A loose belt slips, moves less air, glazes, and throws black dust. A tight belt loads the motor and fan bearings and shortens both. You want the tension the manufacturer specifies, checked with deflection or a tension gauge, not by feel, and the sheaves aligned so the belt runs straight instead of cutting into the flange.

Bearings are where a small miss becomes a big repair. A bearing that is run dry or over-greased fails, and when a fan bearing seizes it can take the shaft and the motor with it. Lubricate to the manufacturer's schedule and quantity, because over-greasing blows out seals and is as harmful as running dry. Listen and feel for the early tells: a bearing going bad gets loud and hot before it lets go, and a vibration that was not there last visit is the equipment telling you something is loosening or wearing.

Check the motor itself. Read the running amps against the nameplate full-load amps, because a motor pulling over its rated current is working too hard and will not last. Look at the motor and fan mounts for cracked welds and loose bolts that let vibration build. On direct-drive and EC-motor units there is no belt, which removes a maintenance item and is one reason the industry has moved toward them, but the bearings and the current draw still get checked.

The condensate pan, drain, and trap

A clogged condensate drain floods, and the flood is rarely confined to the unit. The cooling coil pulls moisture out of the air and it has to go somewhere. When the drain line plugs with the algae and slime that grow in a warm wet pipe, the pan overflows, and on an indoor air handler that water finds the ceiling below, the electrical, or the floor of the tenant space. The drain is a small line that causes large claims.

Clear and flush the drain on a regular interval, treat the pan with tablets or a treatment that slows biological growth, and confirm the line actually runs by pouring water through it, not by looking at it. Check the trap, because a draw-through unit runs the pan under negative pressure and without a proper trap the unit either holds water in the pan or pulls air backward through the drain, and both leave standing water. The RTU startup guide covers trapping the condensate at install, and the same trap is what you verify on every PM.

Look at the pan itself. A pan with standing water that never drains is growing the mat that will plug the line next month, and a rusted or cracked pan needs replacement before it leaks on its own. A float switch or overflow switch is cheap insurance that shuts the unit down before the pan overflows, and confirming it works is part of the visit, because a switch that has corroded open protects nothing.

How do you check the refrigerant charge during PM, and why you never just top off?

You verify the charge by superheat or subcooling against the manufacturer's target, not by guessing from the gauges and adding gas. On a fixed-orifice metering device you check superheat. On a system with a thermostatic or electronic expansion valve you check subcooling, because that is what the charge actually controls. The numbers compared to the manufacturer's charging chart tell you whether the system is correct, low, or overcharged, with the metering type deciding which one to trust.

Here is the rule that separates a tech from a parts-changer: a system that is low on refrigerant has a leak, and topping it off treats the symptom while the gas keeps going to atmosphere. Refrigerant does not get used up. If it is low, it left through a leak, and adding more without finding that leak means another low-charge call next season, a bigger environmental loss, and a customer paying twice. Find the leak, fix it, evacuate, and weigh in the correct charge from the nameplate. Charge by weight, not by pressure, especially with the zeotropic blends now common, where adding vapor changes the mixture.

There are rules behind this, not just good practice. Under EPA Section 608, intentional venting is prohibited and larger systems carry leak-repair obligations once they cross the charge thresholds, so a leak you keep feeding is also a compliance problem. The charge check on a PM is a check, not a top-off. If the numbers are off, the deliverable is a diagnosis and a leak search, recorded and quoted, not a quiet pound of gas added to make the call go away.

Electrical: connections, contactors, capacitors, and the amp draw

The electrical PM starts with the connection nobody can see going bad. A loose lug or terminal builds resistance, resistance builds heat, and heat builds slowly behind a cover for months before it discolors the insulation, melts a connector, or starts a fire. You tighten connections to the manufacturer's torque value, not by feel, and a tug test on each wire confirms it is seated. Over-torque is its own failure, so the spec matters in both directions.

An infrared scan finds the hot connection without taking the panel apart and without touching anything live. Run the unit under load and the thermal camera shows the loose terminal, the failing contactor, the unbalanced phase, and the overloaded conductor as a hot spot against its neighbors. It is the fastest way to find the problem that a visual inspection misses, because the fault hides behind covers and insulation exactly where the eye cannot go.

Then the components that wear. Contactor contacts pit and burn from arcing every cycle, and a contactor that is welded or badly pitted gets replaced before it sticks closed or fails to pull in. Capacitors lose capacitance as they age, and a weak run capacitor makes the motor draw high and run hot and can make the contactor chatter, so you read the microfarads against the rating, not just look for a bulged top. Across all of it, read the amp draw against the nameplate full-load amps. A motor or compressor pulling over its rating is the single most useful electrical number on the visit, because it points at the problem before the breaker does.

Heating-side and combustion safety PM

On furnaces and boilers the PM crosses from efficiency into life safety, because the equipment makes carbon monoxide by design and the only thing keeping it out of the building is a sound heat exchanger and a clear flue. You clean and inspect the burners, check the flame for the right color and pattern, confirm the flue and venting are clear and pitched right, and run a combustion analysis. This is not optional work to defer when the visit runs long. CO is the hazard that kills people in their sleep.

The combustion analyzer is the tool that turns this from a guess into a measurement. It reads oxygen, carbon monoxide, and efficiency in the flue, and the numbers tell you whether combustion is clean and complete. A cracked heat exchanger shows up as a change in the flue readings when the circulating blower starts, because the blower pressure pushes air through the crack and shifts the oxygen and CO. A documented method watches for an oxygen change over about half a percent or a CO change over roughly 25 ppm when the blower kicks on. A cracked heat exchanger is a replace, not a patch, on most equipment, and it is a no-run condition until it is fixed.

On hydronic and steam systems the heating PM covers the burner and combustion the same way, plus the water side: the safety controls, the relief valve, the low-water cutoff, the expansion tank, and the boiler's operating and limit controls. The safeties are the items you test on every heating PM, because a low-water cutoff or a relief valve that has corroded into place is a serious event waiting on the day the control it backs up fails.

Controls, sensors, and calibration

Controls drift, and a system controlling to a wrong sensor wastes energy while looking like it is working. The PM checks that the thermostat or the building automation system is reading true, that sensors are calibrated against a known reference, and that the sequence of operation actually does what the design intended. A space sensor reading three degrees off makes the equipment overcool or overheat the zone all day, and nobody sees it on the front end because the room number on the screen looks fine.

Verify the setpoints and the schedules, not just the hardware. Buildings collect overrides: a setback that someone disabled during a complaint and never restored, a schedule that runs the equipment all weekend for a tenant who left a year ago, a setpoint pushed to an extreme to win one argument. Each one runs equipment and burns energy with no comfort benefit. Walking the actual setpoints and schedules against what the building needs is some of the cheapest energy savings in the program.

On a building automation system, check that the points are reading and that the alarms are alive. A dead sensor that the BAS still trusts feeds the control loop a lie, and a flood of nuisance alarms that everyone has learned to ignore is the same as no alarms at all. The control PM is where you confirm the equipment is being told the truth and is doing what it is told. The air handler guide covers the reset and economizer logic that ride on these controls.

Why is the economizer the most-neglected check in the program?

The economizer is the single biggest source of wasted energy on commercial units, and it is the item that almost never gets checked after the unit is installed. When outdoor air is cool enough, the economizer opens the outside-air damper and cools the building for free instead of running the compressor. When it works. The failure is silent, because a broken economizer stuck at minimum position still cools the building, just on mechanical cooling that costs money during weather that should have been free.

The parts that fail are mechanical and electrical. Damper linkages wear and seize, actuators fail or lose calibration, and the outdoor air sensor drifts so the logic locks out free cooling at the wrong temperature. Any one of those leaves the damper parked, and because the building still gets cool air, the fault can run for months or years and waste thousands of dollars a year on a single unit before anyone connects it to the energy bill.

The check is not hard, which makes the neglect worse. Drive the dampers through their full travel and watch them move, confirm the actuator and linkage are tight and respond to the control, verify the outdoor air sensor against a reference, and run the economizer through its changeover so you see it actually take over from mechanical cooling. Do this on the annual PM and you recover energy nobody knew they were losing. Skip it and the economizer is the most expensive idle part on the roof.

Airflow, static pressure, and balance drift

Airflow is the quiet variable that everything else rides on, and it drifts over the life of a building. A system that was balanced at commissioning does not stay balanced. Filters load, coils foul, belts slip, dampers get bumped, and tenants close registers, so the airflow that was set to design slowly wanders off it. The PM is where you catch the drift before it becomes a comfort complaint or a frozen coil.

Read the external static pressure across the unit and compare it to commissioning. A static that has climbed says something is restricting the airflow, usually a loaded filter or a fouled coil, and it is an early warning before capacity or comfort suffers. A static that has dropped can mean a slipping belt, a duct that came apart, or a damper that opened. Either direction off the baseline is information, which is why the baseline has to exist in the record to compare against.

When the drift is real, it is a test-and-balance question, not a guess. Major changes, a renovation, or a retrofit can move the balance enough to need a fresh TAB, and the testing and balancing trade has the procedures and the standards for that work. On routine PM you are checking for drift and correcting the obvious, not rebalancing the building. The air handler guide covers the fan and airflow path that these readings track.

PM by equipment type

The tasks are the same physics, but the equipment changes which ones matter and how you get to them. A standardized program has a task list per equipment type, because a PM checklist that fits a rooftop unit does not fit a chiller plant. Build the list once per type and the field work becomes repeatable instead of improvised.

Packaged rooftop units carry filters, coils, belts, the economizer, the charge, and on gas units the combustion side, all in one cabinet on a roof, which adds the curb, the roof access, and the fall exposure to the visit. The RTU startup guide covers the install and commissioning that set the baseline these PMs maintain. Air handlers move the same filter, coil, condensate, and fan work indoors, fed by a central plant for heating and cooling, and the air handler guide walks that whole airflow path. Splits and mini-splits and VRF add line sets, multiple indoor heads, and refrigerant distribution, where the charge and the indoor coil and filter work spread across many small units.

The plant equipment is its own discipline. Chillers want the tube bundle, the approach temperatures, the oil and refrigerant analysis, and the water treatment watched and trended. Boilers carry the combustion and water-side safeties covered above. Cooling towers add the fill, the basin, the drift eliminators, and the water treatment that controls scale and biological growth, including the Legionella risk that makes tower water treatment a health item, not just an efficiency one. Match the task list to the equipment and the program scales without losing rigor.

Seasonal startup and shutdown

The two seasonal visits are the spine of the calendar, because they put the equipment under inspection right before it goes under load. Cooling startup in the spring gets the cooling side ready before the first hot stretch: clean the condenser, verify the charge, check the contactors and the compressor, confirm the economizer and the controls, and run the unit in cooling to see it perform before the building needs it. The worst day to discover a dead compressor is the first 95-degree afternoon with a full building.

Heating startup in the fall does the same for the heating side: clean and inspect the burners, run the combustion analysis, check the heat exchanger and the flue, test the safeties, and prove the unit makes heat before the first cold snap. On a heat pump the changeover between heating and cooling is its own check, because the reversing valve, the defrost cycle, and the auxiliary heat all have to work in the season they are about to enter.

The changeover is also when you correct the season's accumulated drift. Filters get changed, balance gets a look, and the overrides that crept in over the last season get cleaned up. Two well-run seasonal visits catch most of what a building will throw at you, because they hit the equipment at the moment its weakness is about to be exposed and there is still time to fix it on a planned visit instead of an emergency.

The PM task list and standardized procedure

A PM program is a written task list per equipment type, performed the same way every time, by whoever shows up. Without the written list, the quality of the visit depends on which tech caught the call and how their day is going, and the program becomes a set of individual habits instead of a standard. The list is what makes the work repeatable and auditable.

Each task on the list carries a frequency and an as-found measurement, not just a checkbox. A line that reads only done tells you nothing six months later. A line that records the static pressure, the amp draw, the superheat, and the head pressure as you found them builds a history you can trend, and the trend is what turns a PM program from a cleaning service into a predictive one. The reading is the deliverable, not the checkmark.

ANSI/ASHRAE/ACCA Standard 180 is the framework for this. It lays out inspection and maintenance tasks by equipment type with minimum intervals, and the residential counterpart, ACCA Standard 4 (4 QM), does the same for residential systems, both ANSI-recognized and manufacturer-endorsed. Use them to build the task list, then add what the manufacturer's instructions and the specific equipment require, because the standard sets a floor and the equipment sets the rest.

Documentation and the service record

The PM that is not documented did not happen, as far as anyone who was not on the roof can tell. The service record is what proves the work, justifies the contract, and turns a string of visits into equipment history you can act on. Log every visit with the date, the equipment, the tasks performed, and the readings, and the program becomes defensible and useful instead of a recurring charge nobody can account for.

Record as-found and as-left. As-found is the condition you arrived to, the dirty filter, the high static, the low charge, the stuck economizer. As-left is the condition you handed back. The two together prove the value of the visit and surface the trend: a static pressure or an amp draw or a subcooling number that creeps the same direction across visits is the equipment telling you what will fail next, and you only see it if the numbers are written down in a form you can compare. Capture deferred repairs too, the items you found but did not fix, with a quote, so nothing falls through and the customer owns the decision to defer.

This is where a recurring schedule and a service log earn their keep, and it is what FieldOS is built to handle: the recurring PM schedule that puts the right tasks on the calendar by equipment and frequency, and the service log that captures the readings, the as-found and as-left, the photos, and the deferred work against each piece of equipment over time. A program with a real record trends the equipment and proves the work. A program without one is just visits.

What belongs in an HVAC PM service agreement?

The PM agreement is the business side of the program, and the scope is where it lives or dies. A full-coverage agreement bundles the scheduled maintenance with repairs and parts, so the customer pays a predictable amount and the contractor carries the risk of what breaks. An inspection-only or labor-only agreement covers the PM visits and the labor to find problems, but bills repairs and parts separately. Both are legitimate. What kills the relationship is a scope that is vague about which one it is, so the customer expects full coverage and the invoice says otherwise.

For the contractor, the recurring agreement is the steadiest revenue there is, and the steadiest because it is earned. A serviced base of equipment generates fewer emergencies, the visits surface the repairs before they become failures, and the relationship means the customer calls you first. The PM contract is the difference between a service business and a string of one-off emergency calls that compete only on who answers the phone.

For the customer, the value is the equipment running longer, cheaper, and more reliably, plus a record that supports warranty claims and capital planning. Spell out the scope, the frequency, the equipment covered, the response time, and what is included versus billed, in writing. A recurring schedule and a service history, the kind FieldOS keeps, is what lets you show a customer at renewal exactly what they got for the agreement, which is the easiest renewal conversation there is.

Predictive maintenance, sensors, and the smart building

Predictive maintenance watches the equipment continuously and acts on the data, instead of waiting for the calendar or the breakdown. Sensors track vibration, motor current, temperatures, pressures, and runtime, and analytics trend them to flag a developing fault before it becomes a failure. A bearing trending warmer and rougher over weeks gets caught and scheduled, not run to seizure. This is condition-based maintenance, and on critical or expensive assets it pays for the instrumentation.

Fault detection and diagnostics, layered onto a building automation system, is the commercial version. The analytics watch the points the BAS already collects and surface the faults that hide in plain sight: a simultaneous heating and cooling fight, an economizer stuck shut, a sensor reading nonsense, a unit running off schedule. Studies put roughly a third of commercial HVAC energy use down to faults like these, much of it invisible because the building still holds temperature while it wastes the energy.

Predictive does not replace the wrench. Someone still changes the filter, cleans the coil, and tightens the connection, and the basic PM still gets done. What the data changes is where you aim the hours. It tells you which unit needs attention now, which interval you can safely stretch, and which fault to chase before it becomes a callback. Start with a solid preventive program, instrument the assets where downtime or energy is expensive, and let the trend tell you where the next failure is forming.

The energy case for PM

Equipment loses efficiency without maintenance, steadily and invisibly, and the building pays the difference in energy long before anything fails. A dirty coil makes the compressor work harder for less heat transfer. A clogged filter makes the blower fight for airflow. A slipping belt moves less air per watt. A stuck economizer runs mechanical cooling during weather that should have been free. None of these trips a breaker. All of them show up on the utility bill.

The savings from PM are real and measurable. Clean coils, correct charge, proper airflow, calibrated controls, and a working economizer keep the equipment near its rated efficiency, and the gap between a maintained unit and a neglected one widens every season the neglect continues. The economizer alone, when it has failed shut, can waste thousands of dollars a year on a single unit, which is why the energy case and the economizer check are the same conversation.

Frame the energy savings for the customer the way the invoice frames the repair cost. A PM program that holds efficiency pays part of its own cost back in energy every month, on top of the avoided emergencies and the extended equipment life. The owner who treats PM as a cost to cut is paying for it anyway, in kilowatt-hours and in the replacement that arrives years early.

Technician safety on the PM visit

PM puts a tech around energized equipment, moving parts, the roof edge, refrigerant, and combustion, often alone, and the routine of it is exactly what makes the hazard easy to forget. Lockout/tagout is the first habit. Before you open a panel or reach into a unit, the energy is isolated, locked, and verified dead with a meter you trusted on a known source. The compressor that starts under your hand because a control closed does not care that you were only changing a filter.

The roof is its own set of hazards. Rooftop units sit near edges and skylights, and fall protection is the requirement that techs skip because the visit is quick and they have done it a hundred times. The hundred-and-first is the one. Confined spaces, the mechanical room, the chiller pit, the tower basin, carry their own entry procedures and atmospheric hazards, and they are not a place to improvise.

Electrical work brings arc-flash and shock exposure that the right PPE and the dead-verified state control, and refrigerant brings displacement, frostbite, and the hazard of working with the newer mildly flammable A2L refrigerants now entering equipment. Combustion work brings carbon monoxide, which is the reason the analyzer is a safety tool and not just an efficiency one. None of this is a reason to slow the program down. It is the reason the program has procedures, because the routine visit is where the guard comes down.

PM for data centers and critical facilities

Critical-facility HVAC raises the rigor because the cooling cannot stop. A data center, a hospital surgical suite, a lab, or a telecom site runs a load that does not tolerate the building getting warm, so the maintenance has to happen without taking the cooling down. That changes everything about how PM is planned, even though the tasks on the equipment are the same filters, coils, charge, and controls as any other unit.

Redundancy is what makes the no-downtime maintenance possible. A facility built to N+1 or better has spare cooling capacity, so a unit can be isolated, locked out, and serviced while the redundant equipment carries the load, and a concurrently maintainable design lets every component be taken out of service one at a time without losing cooling. The PM program has to respect that topology: you service one path at a time, you confirm the redundant path is healthy before you isolate anything, and you never have two units down at once in a way that erases the redundancy.

The standards shift toward the load. ASHRAE Technical Committee 9.9 gives the thermal guidelines for data processing environments, the temperature and humidity envelopes the equipment has to hold, and the Uptime Institute tier framework defines the redundancy and concurrent maintainability the facility was built to. The PM here is tighter, better documented, and planned around the redundancy, because the cost of getting it wrong is not a comfort complaint, it is downtime measured in dollars a minute.

Common mistakes

• Skipping or stretching the filter change until the filter is gray and bowed and the coil behind it is loading up.
• Never cleaning the condenser coil, then blaming the compressor that died on high head pressure.
• Topping off a low system with refrigerant instead of finding and fixing the leak.
• Ignoring the economizer, leaving a stuck damper to run mechanical cooling through free-cooling weather for years.
• Recording done instead of the as-found readings, so there is no trend and no proof of value.
• Running reactive-only, paying the three-to-five-times emergency premium and the collateral damage every time.
• Skipping the seasonal startup and discovering the dead compressor or furnace on the first day of load.
• Checking connections by feel instead of to a torque spec, and never running an infrared scan.
• Deferring the combustion analysis and the heat-exchanger check, which is a life-safety item, not a comfort one.

Field checklist

Standards and references

The governing maintenance standard for commercial HVAC is ANSI/ASHRAE/ACCA Standard 180, Standard Practice for Inspection and Maintenance of Commercial Building HVAC Systems, developed jointly by ASHRAE and ACCA. It sets minimum inspection and maintenance tasks by equipment type with intervals, and it is the framework a written PM program is built on. The residential counterpart is ACCA Standard 4 (4 QM), Maintenance of Residential HVAC Systems, ANSI-recognized and manufacturer-endorsed. Both set a floor, not a ceiling.

Frequencies are where you have to hedge honestly. The standard gives minimum intervals, but the manufacturer's installation and maintenance instructions and the specific equipment's duty and environment set the real schedule, and they take precedence where they are stricter. A unit in a clean office and the same unit in a dusty plant do not get the same filter interval. Treat the standard's intervals and any rule of thumb in this guide as a starting point, and let the manufacturer and the conditions move them.

Other bodies govern the pieces. ASHRAE 62.1 covers ventilation and the outdoor air the economizer manages, ASHRAE 90.1 the energy requirements, and ASHRAE Technical Committee 9.9 the data-center thermal envelopes. EPA Section 608 governs refrigerant handling, leak repair, and the prohibition on venting. The testing-and-balancing standards from NEBB and AABC cover the airflow work when drift calls for a rebalance. Cite the body that controls the point, and confirm the current edition and any local adoption before you put a number on a submittal.

Units, terms, and what they mean

PM crosses several measurements and a few terms that read differently across a manufacturer sheet, a controls screen, and a service ticket, so the same idea can wear more than one name.

Preventive maintenance is abbreviated PM, scheduled or planned maintenance is the same idea, and the condition-based versions go by predictive maintenance or PdM and fault detection and diagnostics, FDD. Static pressure is read in inches of water column, written in. w.c. or in. wg, with Pa in metric sources. Refrigerant charge is verified by superheat and subcooling, both in degrees, and the charge itself is weighed in pounds or kilograms. Airflow is in cubic feet per minute, CFM, or cubic meters per hour in metric. MERV rates filter efficiency, higher being finer. Amp draw is compared to nameplate full-load amps, FLA.

PM / PdM

Preventive maintenance on a schedule; predictive maintenance driven by condition data and trending

FDD

Fault detection and diagnostics, analytics on BAS data that surface faults like a stuck economizer

MERV

Minimum Efficiency Reporting Value, the filter's particle-capture rating; higher is finer and more restrictive

Superheat / subcooling

Charge-check measurements in degrees; superheat for fixed-orifice, subcooling for expansion-valve systems

Static pressure (ESP)

External static pressure across the unit in inches of water column, the airflow-restriction indicator

FLA

Full-load amps on the nameplate, the current a motor or compressor should not exceed in service

Economizer

Dampers and controls that cool with outside air when conditions allow, the most-neglected energy item

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• What is an air handling unit?
• How do you start up an RTU?
• What is a data center PM program?
• What is a fleet maintenance program?
• What is pavement preservation?