What Is a VAV System? Variable Air Volume Explained

Look into the ceiling void of a typical air conditioned office tower and you will find a large duct running down the corridor with a series of insulated metal boxes tapped off it, one per zone, each carrying a small controller. Those boxes are VAV terminal units, and between them they decide how much conditioned air every part of that floor receives. VAV is not a complicated idea, but the interaction between the boxes and the fan that feeds them is genuinely the thing newcomers find hardest, and it is also where most of the wasted energy in commercial buildings hides.

The message up front: a VAV system varies the quantity of air delivered to each zone while holding the supply air temperature roughly constant. Constant volume does the opposite: fixed airflow, varying temperature. VAV won the commercial office market because varying flow saves fan energy and because one central plant can serve many zones with independent temperature control. The catch is that VAV only performs as designed while its minimum flows, flow sensors, damper linkages and static pressure reset are all still correct, and in most buildings at least one of those has quietly stopped being correct.

1. What variable air volume actually means

Every air conditioning system has to remove a certain quantity of heat from a space. The physics gives you two levers to do it with: how much air you push into the room, and how cold that air is. Multiply the two together, roughly speaking, and you get the cooling delivered. Any air based system is therefore a choice about which of those two levers you modulate and which one you hold still.

A variable air volume system modulates the quantity. The central plant, normally an air handling unit, conditions air to a reasonably steady supply temperature and pushes it down a duct. Each zone has a terminal box that opens or closes a damper to take more or less of that air depending on how warm the zone currently is. A conference room filling with people takes more air. An empty corner office takes less. The supply temperature stays where it is; the volume is what changes.

That is the whole core idea, and a lot of confusion downstream comes from not having it straight. When somebody says a zone is "starving", they mean the box cannot get enough air volume, not that the air is not cold enough. When somebody says a floor is "overcooled at low load", they usually mean boxes are sitting at a minimum flow higher than the zone needs, so cold air keeps arriving whether the zone wants it or not. Both are volume problems.

VAV sits inside the wider family of air based systems covered in the HVAC systems and components guide. The unit that generates the conditioned air and the duct pressure the boxes draw from is the air handling unit, explained in detail in what is an AHU. You cannot really understand VAV without understanding the AHU, because the two are a single control problem wearing two names.

2. VAV versus CAV, and why VAV took over the office market

The predecessor to VAV is constant air volume, usually written CAV. A CAV system delivers a fixed airflow to the space continuously and varies the supply temperature to suit the load. In its simplest single zone form, CAV is elegant: one thermostat, one coil, one fan, nothing to go out of calibration. It is still the right answer for a single large space with a uniform load, which is why you find it in theatres, lecture halls, plant rooms, data halls and many industrial applications.

CAV falls apart when you ask one system to serve many zones with different loads simultaneously. The classic workaround was terminal reheat: cool all the air to satisfy the hardest zone, then reheat it at every other zone to stop that zone freezing. It worked, and it wasted an enormous amount of energy doing it, because you paid to cool air and paid again to warm the same air back up.

The energy argument for VAV, stated qualitatively because that is the only honest way to state it without a specific building in front of you, rests on two things. First, a building at part load needs less air, and if you actually reduce the air you move, the fan power required drops much faster than the airflow does. That relationship is the largest structural saving VAV offers, and it is why the industry accepted the extra controls complexity. Second, sizing: because zone peaks do not all occur at the same hour, a VAV system can be sized for the coincident building peak rather than the sum of every zone peak, which means smaller ducts, fans and coils. Running cost and capital cost together are why VAV became the default for commercial offices.

3. Inside a VAV box: damper, sensor, actuator, controller, coil

A VAV terminal unit is a simple piece of equipment. Strip the insulation off and you have five things worth naming.

• The casing and inlet: a sheet metal box with a round inlet spigot connected to the main duct and a rectangular outlet feeding the zone's ductwork and diffusers. Internally lined for acoustic and thermal reasons.
• The damper: a single blade, normally a round disc on a shaft, sitting in the inlet. Rotating it changes the free area and therefore the airflow. This is the only moving part in the air stream on a basic box.
• The flow sensor: usually a multi point averaging pitot ring or a cross shaped flow probe in the inlet, connected by two small plastic tubes to a differential pressure transducer in the controller. It measures velocity pressure, from which the controller infers volumetric flow. On many boxes it is one of the most failure prone components, because two thin tubes are easy to kink, crush, disconnect or block.
• The actuator: a small electric motor, typically 24 volt, that drives the damper shaft. Coupled to the shaft by a clamp or a linkage. Commonly fitted with a manual release so a commissioning engineer can move the damper by hand, which is also how it ends up in the wrong position permanently.
• The controller: the brain. Reads the zone temperature sensor and the flow signal, compares them against setpoints and flow limits, and drives the actuator. On a networked system it also reports to the building management system and accepts setpoint and schedule commands from it.
• The reheat coil, where fitted: an electric element or a hot water coil in the discharge, used to warm the air when the zone needs heating or when the box is at minimum flow and the zone would otherwise be overcooled. Hot water coils bring a control valve with them; electric coils bring a contactor and a safety thermostat.

That is it. The box is not clever. The cleverness, and therefore the fragility, is in the controller's logic and in the calibration of the flow sensor. Everything else is metal.

4. Pressure independent versus pressure dependent, and why it matters

This distinction sounds like jargon and turns out to be one of the most practically important things to know about any given box.

A pressure dependent box has no flow sensor. The controller reads zone temperature and positions the damper to some angle it believes corresponds to the required airflow. The problem is obvious once stated: airflow through a fixed damper opening depends on the pressure in the duct upstream, and that pressure changes constantly as other boxes on the same duct open and close. So the same damper position delivers different airflow at different times of day. A pressure dependent box is cheap and it is adequate for a non critical zone on a system with stable duct pressure, and it is a source of chronic complaints anywhere else.

A pressure independent box has the flow sensor and runs two control loops in cascade. The outer loop looks at zone temperature and decides what airflow the zone needs, expressed as a flow setpoint. The inner loop looks at measured flow and drives the damper until measured flow matches that setpoint. Because the inner loop is closing on measured flow rather than on damper position, upstream pressure changes are rejected automatically: if the duct pressure rises, measured flow rises, and the damper closes to compensate. The box delivers the flow it was asked for regardless of what the neighbours are doing.

Nearly all modern commercial VAV boxes are pressure independent, and you should assume that is what you have unless you find otherwise. The practical consequences are worth spelling out, because they explain a lot of behaviour:

• A pressure independent box lets you set meaningful minimum and maximum flow limits in engineering units, because the controller knows what flow it is actually passing. On a pressure dependent box, "minimum" can only be a damper position, which is not a flow.
• A pressure independent box whose flow sensor has failed does not fail obviously. It fails into confidently controlling to a wrong number. A blocked pitot tube reads low flow, so the controller drives the damper wide open and leaves it there, and the zone is now over supplied while the graphic shows it starving.
• Duct static pressure reset, discussed below, depends on damper position feedback from the boxes. It works with pressure independent boxes because they are already modulating to a flow target rather than fighting the duct pressure.

5. The main box types and where each is used

The generic VAV box has a handful of variants, and knowing which one you are standing under changes what you check and what you can expect.

A single building routinely contains several of these: cooling only boxes serving the core, single duct with reheat around the perimeter, and fan powered boxes on a few special zones added during a fit out. When diagnosing a comfort complaint, the first question is always which type of box serves that zone, because the answer changes the whole diagnosis.

6. The bit newcomers find hardest: the box and fan interaction

Here is the mental model to get right. Each VAV box is a local controller pursuing its own zone's comfort, and it does so by opening or closing a damper. The boxes do not talk to each other, and in most designs they do not talk to the fan either. They simply act, all of them, on the same shared duct.

Now think about the duct when building load falls. Zones cool, boxes close their dampers. Closing dampers restrict the duct, and a fan pushing against a more restricted duct produces less flow at a higher pressure. So as boxes close, duct static pressure rises. Left alone, that rising pressure works against every box on the system: they fight more pressure to deliver the same flow, the dampers throttle harder, the boxes get noisier, and the fan burns energy pressurising a duct nobody asked to be pressurised.

The fix is to give the supply fan its own control loop. A static pressure sensor is fitted well along the main duct run, and the fan speed modulates to hold the measured duct static pressure at a setpoint. Boxes close, pressure starts to rise, the fan slows down until pressure returns to setpoint. Boxes open, pressure falls, the fan speeds up. That loop is what converts the boxes' independent damper movements into an actual reduction in fan speed, and therefore into the energy saving that justifies VAV in the first place.

There is one more layer, and it is the one worth understanding properly: static pressure reset. Holding a fixed duct pressure setpoint is safe but wasteful, because the setpoint has to be high enough to satisfy the worst case, and most of the time you are not in the worst case. Reset logic instead watches the damper positions reported by the boxes and asks a better question: is any box struggling? If the most open damper on the system is still comfortably short of fully open, every box has margin, so the pressure setpoint can be lowered. If a box goes fully open and still cannot make its flow setpoint, the pressure setpoint is raised. The system settles at the lowest duct pressure that keeps the hardest working zone satisfied, which is exactly the right place to be.

That is as deep as an explainer should go. The written sequences, the loop tuning, the reset schedules and the failure symptom tables belong in the controls article: see BMS in HVAC: control points and sequences for the sequence of operation level detail, and the complete guide to building management systems for how the boxes appear as points on a supervisory system in the first place.

7. Minimum flow settings and the ventilation consequence

Every pressure independent box has a minimum flow and a maximum flow programmed into its controller. Maximum is a capacity and a noise limit: it is the most air that box and its diffusers can handle. Minimum is the interesting one, and it exists for two separate reasons that are often conflated.

The first reason is air distribution. Below a certain flow, the diffusers stop throwing air properly, the supply air dumps straight down instead of mixing, and the zone gets cold drafts in one spot and stagnation in another. So there is a mechanical floor below which the box should not go.

The second reason is ventilation. The fresh air a zone receives is a fraction of the supply air it receives, because the supply air is a mixture of outside air and recirculated return air. If a box throttles down to a low flow, the zone's fresh air supply throttles down with it. On a floor at low thermal load but full occupancy, which is exactly what a mild shoulder season morning looks like, every box can be sitting near minimum while the people in those zones need the same fresh air they always needed. That is the structural ventilation weakness of VAV, and it is real.

This is where I have to be careful about numbers, and you should be too. How much fresh air a given space is required to receive, and how that requirement is calculated for a variable flow system, is set by the ventilation code in force where the building stands. It is jurisdictional. In practice the widely used technical references are the ASHRAE standards, which are US in origin and adopted or referenced far beyond the US, and CIBSE guidance in the UK. Neither one is automatically the law where you are. Get the current edition of whichever document your authority having jurisdiction actually enforces, and calculate against that rather than against a figure somebody quoted you in a meeting.

The design responses, in rough order of how often you meet them: set box minimums so the zone still receives adequate outside air at minimum flow; increase the outside air fraction at the AHU when total system flow is low; or measure actual demand with carbon dioxide sensing and modulate outside air accordingly. Each is legitimate and each has costs. What is not legitimate is treating the box minimum as a comfort setting only and never asking what it does to fresh air.

8. Why VAV systems drift out of tune

This is the honest core of the article. A VAV system is commissioned once, at a point when the building is empty, the partitions are where the drawings say they are, and a commissioning engineer has just balanced every box. From that day forward it degrades, and almost none of the degradation announces itself. Here is where it comes from, in the order I most often find it.

• Minimum flows set high at commissioning and never revisited. Under time pressure, minimums get set generously, because a generous minimum guarantees no cold complaints and no ventilation argument on handover day. Few ever go back. Years later the building is running with every perimeter box holding a minimum flow far above what its zone needs, and the reheat coils are working steadily to undo it. This is commonly among the largest recoverable wastes in a VAV building, and it costs nothing but engineering time to correct.
• Reheat fighting cooling. The direct consequence of the above, but it also arises on its own: a leaking hot water reheat valve, a contactor welded closed on an electric coil, or a sequence that permits the cooling damper and the heating coil to be active simultaneously. The zone stays comfortable, which is why nobody reports it, while the building pays for cooling and heating the same air. Look for boxes at minimum flow with reheat active during mild weather.
• Failed or drifting flow sensors. The two plastic tubes from the pitot ring to the transducer are the weak link: kinked during a ceiling tile replacement, crushed by a cable tray installer, disconnected and never reconnected, or blocked with dust and construction debris. A box reading low flow opens wide and stays there. A box reading high flow closes down and starves its zone. Both read plausibly on the graphic. This is why a flow reading that never changes, or that sits pinned at zero or at maximum, is worth investigating even when nobody has complained.
• Disconnected or slipping actuator linkages. The clamp on the damper shaft works loose, or somebody used the manual release to prove a point during a call out and left the damper where it was. The actuator now strokes happily through its full range while the damper does not move, and the controller has no way to know: it commands a position, gets no flow response, and integrates further and further in the same direction. A box whose commanded damper position sits permanently at either end of its range is a strong candidate.
• Fit out changes nobody told the controls about. This is the big structural one. A floor is re-partitioned. Two zones become five, or five become two. Meeting rooms appear in what used to be open plan. Server rooms and comms cupboards get created with no cooling of their own. The diffuser layout is altered. The airflow the zone needs, and the location of the thermostat relative to the load, have both changed, and the controls documentation has not. The box is still holding setpoints calculated for a floor plan that no longer exists.
• Static pressure reset disabled or never enabled. Reset gets turned off during a complaint investigation to rule it out, and is not turned back on. Or it was specified, never commissioned, and the fan has held a fixed high setpoint since handover. Either way the building loses the largest fan saving available to it, silently, because the zones are still comfortable.

Notice the common thread. Every one of these faults leaves the occupants reasonably comfortable, which is the only outcome most organisations actively monitor. Comfort is a terrible proxy for correctness on a VAV system, because the system has two energy inputs, cooling and reheat, and it can burn both to reach a comfortable result. A VAV floor can be simultaneously comfortable and badly broken, and that is the sentence I would want a new operator to remember.

9. Where to take this next

Two directions, depending on what you are trying to do.

If you suspect the system is not doing what the design intended, that is a commissioning or retro commissioning question. The graphics, trend logs and operator dashboards you need to see the drift are covered in BMS commissioning, graphics and operator dashboards. Trending damper position, flow setpoint, measured flow, zone temperature and reheat output together for a couple of weeks tells you more than any number of site visits, and it costs nothing if the points already exist.

If the boxes and the AHU are not maintained on any sensible cycle, that is a planned maintenance question. Filters, coils, actuators, linkages, sensors and flow tubing all have a natural inspection rhythm, and preventive maintenance for HVAC systems covers building that into a schedule rather than reacting to complaints. The VAV specific addition I would argue for is periodic flow verification on a rotating sample of boxes, because that is the only way a failed flow sensor gets found before it has wasted a year of energy.

And if you are comparing air distribution strategies rather than operating one, the useful comparisons are the decentralised alternatives: what is an FCU, AHU versus FCU, and what is a heat exchanger for the component doing the actual heat transfer in the coils. For primary technical references rather than a summary, ASHRAE and CIBSE are the bodies whose published guidance the industry works from.

The idea to walk away with

VAV is a simple idea implemented as a large number of small independent control loops sharing one duct. Vary the quantity of air, hold the temperature steady, give every zone its own box, and let the fan slow down as the boxes close. That is the design, and when it works it is a genuinely good design, which is why it dominates the commercial office market.

But a system made of many small loops has many small ways to go wrong, and on a VAV system almost all of them are invisible from the occupied space. The minimum flow nobody recalculated, the flow sensor with a kinked tube, the actuator turning a shaft it is no longer clamped to, the reheat valve seeping, the reset schedule somebody disabled in 2019: none of those produce a complaint, and all of them produce a bill. If you take one operating habit from this article, make it this: judge a VAV system by what its dampers, flows and reheat outputs are doing, not by whether anybody has complained.

Final thoughts

The reason VAV is worth understanding properly, even if you never touch a controller, is that it is where the building's air, energy and comfort all meet in one small metal box. A facilities manager who can look at a floor's VAV trend data and say "those four perimeter boxes are sitting at minimum with reheat on in October, why" is doing more for that building's running cost than any number of plant upgrades will.

One honest qualification to close on. Ventilation adequacy and the code compliance of any flow setting are jurisdictional and, in many places, matters for a competent designer to sign off rather than for an operator to adjust. Understand the system as thoroughly as you can, trend it, question it, and build the evidence. Then bring a competent designer to the decisions that carry compliance consequences, because the cost of getting a ventilation calculation wrong on an occupied floor is not an energy bill.

Related reading: What is HVAC: systems and components, What is an AHU, What is an FCU, AHU vs FCU, BMS in HVAC: points and sequences, BMS commissioning and dashboards, Preventive maintenance for HVAC systems.