Air Balancing Air Systems

The rest of this chapter walked the air from the handler out to the boxes and taught what every station is supposed to do. This page is about proving it does. One question, start to finish: how do you balance and commission the air side so every zone actually receives the design flow it was drawn for — and how do you know when it doesn't?

What the Box Thinks It's Moving

Start where every air balance starts: the number the box reports. A pressure-independent VAV box measures its own airflow with a flow ring — a ring or cross of averaging ports across the inlet, one set of holes facing into the airstream and one facing away. The difference between them is the velocity pressure, and velocity pressure rises with the square of air speed, so airflow follows a square root of it:

CFM = K × √VP  (also written Q = K√ΔP)

K is the whole personality of the reading. It's the box manufacturer's coefficient for that particular inlet — literally the CFM the ring reads at exactly 1.0 in. w.c. of velocity pressure, since √1.0 = 1. A box with K = 1000 seeing 0.36 in. w.c. believes it's passing 600 CFM (1000 × √0.36 = 1000 × 0.6). That K comes off a catalog for the box size, and — like the K-factor itself — the whole velocity-pressure frame is US-native; the number printed on a metric box is a different figure on a different label, so this section keeps VP and K in their IP field frame the way the airflow tool does. What matters for balancing is that K is not a law of nature. It's a nameplate estimate, and the installed reality — duct approach, transitions, the specific ring — shifts it. An uncommissioned box obeys its flow loop perfectly, around a number nobody ever checked.

So the first act of air balancing is to distrust that number and go get an independent one. A duct traverse — a grid of Pitot readings averaged across the duct by the log-Tchebycheff points, not a single centerline poke — is the honest measurement the box's ring is only estimating. Put the two side by side. If the traverse (or a calibrated flow hood on the diffusers) says the box is really moving 660 CFM while the controller computes 600 CFM at the same velocity pressure, the box is reading about ten percent low — and every setpoint you program into it will be off by that ten percent until you fix K. The fix is arithmetic: the true K is 660 ÷ √0.36 = 1100, so you trim the coefficient from 1000 up to 1100 (the airflow tool's second mode back-solves exactly this), and only then does a programmed setpoint mean the CFM it says. The independent measurement is not optional; it's what separates a balanced box from a confident one.

Three Setpoints in the Box

With K honest, the box's flow number finally means something, and now you can program the setpoints the balance is built around. Three of them do the work. The maximum — the cooling design flow — is the ceiling the box drives toward when the zone is hottest; it comes off the mechanical schedule as the zone's design cooling load expressed in air. The minimum is where the flow loop stops when the zone satisfies, and it is the setpoint people get wrong, because they read it as a comfort number when it's really a ventilation floor. And on a box that can heat, a separate reheat (heating) minimum — sometimes higher than the cooling minimum — sets the flow the box holds while its reheat coil warms the air.

The minimum deserves the emphasis. Occupied zones need their code-required share of fresh air, and that share rides in on the supply air; close a box too far and the temperature can be fine while the ventilation quietly isn't. The floor is not a round number somebody liked — it's the outcome of the ASHRAE 62.1 ventilation math (the per-person and per-area rates, divided by the zone air-distribution effectiveness) turned into a minimum CFM at the box. Set it below that and you've balanced a zone that meets its cooling load and starves its occupants of air; set it needlessly high and you pay to over-cool the zone and then reheat it back, which is the plant heating air it just paid to cool. Getting the minimum right, rather than generous, is most of the craft — and it's why the balance report records the minimum as a ventilation figure, not just a comfort one.

One trap belongs here because it hides inside the box's own honesty. The box holds flow, but the fresh-air fraction that flow carries is set at the unit, and on a VAV system the supply fan slows as the boxes throttle — so a fixed outside-air damper passes less actual outside air at low fan speed. A box minimum that delivers its ventilation share on the design day can quietly under-ventilate on a mild one if the unit isn't measuring or controlling its outside air. The box's minimum is a necessary floor, not a sufficient one; the unit has to hold up its end too.

Pressure-Independent, Pressure-Dependent, Bypass

Not every terminal defends its own flow, and which kind you're balancing changes what the balance even means. A pressure-independent box is the one the last two sections assumed: it reads its flow ring and drives the damper to hold a CFM setpoint, so when trunk static wiggles — a neighbor throttling, the fan speeding up — the flow loop absorbs it and the zone never feels it. Balance that box once and the setpoints hold, because the box itself keeps holding them. It's the air-side cousin of the PICV on the water side, and for the same reason: it meters flow, not position.

A pressure-dependent box has no flow ring and no flow loop. Its damper is commanded to a position, and the air that flows through that position depends entirely on the inlet static behind it — so the moment another zone opens or closes and shifts trunk pressure, this zone's flow moves with it, to a load that never changed. There is nothing to program a CFM setpoint into. Balancing a pressure-dependent system means setting damper positions (and often manual balancing dampers in the branch) at one operating condition and accepting that the balance is only true near that condition — the same fragility a calibrated balancing valve has on the water side. It's why pressure-independent boxes became the default: the balance survives the building changing its mind.

Bypass systems solve the mismatch differently, at the unit instead of the zone. A bypass box (or a bypass damper from supply to return) dumps excess supply air straight back to the return when the zones close down, so the fan sees a nearly constant volume while the zones see variable. The bypass damper is controlled off duct static pressure: as the zone dampers close and static rises, the bypass opens and spills the surplus, holding the trunk at a setpoint. It keeps a constant-volume unit alive on a variable-flow building — cheap, common on small packaged systems — but it saves no fan energy (the fan moves design air all day, some of it in a circle) and it can short-circuit cold supply air back over the return sensor. Balancing a bypass system means setting that static setpoint so the bypass carries exactly the surplus and no more, then proving the zones still make their design flows with the bypass doing its job. It is the opposite philosophy from a pressure-independent box, which holds flow regardless of static rather than spilling to defend static.

Setting the Building Slightly Positive

Zone-by-zone flow is only half of a balance; the building as a whole has to keep its own books. Every cubic foot the unit brings in as outside air has to leave somewhere, and the residual — after the relief path, the dedicated exhausts, and the leakage have taken their share — is the building pressure. The target isn't neutral: designers hold a building slightly positive, a setpoint around +0.02 to +0.05 in. w.c. relative to outdoors, so conditioned air seeps outward through the cracks instead of unfiltered, humid air seeping in. Get it wrong and the doors tell on you — too positive and exterior doors drift open against their closers; too negative and they go heavy and whistle at the weatherstripping.

For the balancer this lands as a tracking relationship: supply, return, exhaust, and relief have to add up to the small positive surplus you want. On a built-up unit that usually means setting the return or relief fan to track the supply fan at a measured-flow offset — supply minus return equals the standing withdrawal the building's dedicated exhausts steal from the unit's territory, plus the surplus that holds it positive. The offset is not a fudge factor; it's an air-ledger figure, and it has to be commissioned against the exhausts that were actually installed, not the ones on the drawing. Getting that offset right is where zone balancing and building balancing meet — the whole envelope story is its own page of this chapter, and the balance isn't finished until the pressure sits where it belongs across the operating range, not just at one damper position.

Proportional Balancing: Evening Out the Zones

Now the field method itself, and it's the same one the water side uses because it solves the same problem: the terminals interact. Throttle the near ones and the pressure they were hogging shifts to the far ones, which then read differently — so you can't just walk the trunk once, set each terminal to its design flow, and call it done. The first terminal you touch is already wrong by the time you reach the last.

The answer is proportional balancing. Pick a reference: the index terminal — the worst-case run, usually the farthest or most-restricted zone, the one that will lose flow first when pressure drops. Measure every terminal as a fraction of its own design flow, and instead of chasing absolute CFM, bring each one to the same fraction as the index. When every terminal is delivering, say, ninety percent of its design, the system is in proportion even if it's not yet at full flow — and now one move fixes all of it: set the fan (or open the index wide) to lift the whole proportioned set up to one hundred percent together. Because the terminals are already in the right ratio, raising the common supply raises them all in step. Then go back and check, because the first pass moved the reference; two or three iterations is normal, more on a big system. It is real work — half a day for a small system, a week for a large one — and the payoff is a building that's even, not one that's merely full.

Two air-side wrinkles the water side doesn't have. First, on pressure-independent boxes much of this is done in software — you're proportioning by trimming K and the min/max setpoints so each box's held flow matches its design share, then letting the boxes hold it — but the index-zone logic is identical: prove the worst-case zone can make its number, and the rest follow. Second, the total has to reconcile: the sum of the balanced box flows should agree with an independent traverse of the supply trunk, and with the design outside-air fraction the unit is set to bring in. A building whose zones each read right but whose trunk traverse doesn't add up is telling you a box is lying, a duct is leaking, or the schedule and the installation disagree.

Verifying Against Design

A balance isn't a set of dial positions; it's a record. The deliverable is the air-balance report, and it's the thing the next tech — and the commissioning authority, and the owner — actually relies on. For each terminal it states the design flow, the measured flow, and the percentage of design; for each unit it states the total supply, return, outside, and exhaust air, so the ledger visibly closes. The report is where the cross-checks live: the sum of the box flows against a supply-trunk traverse, the delivered air against the unit's nameplate with a CFM-per-ton and CFM-per-person sanity check, and the outside air against the 62.1 minimum. Any two of those disagreeing is a finding, not a rounding error.

And the record is what makes the balance survivable. The reason a building drifts out of balance months after the balancer leaves — a K nudged during a filter change, a box re-tuned and never re-checked, a unit whose fan curve moved when a coil was swapped — is that nobody could tell what "right" used to look like. A balance report with real numbers, and boxes whose setpoints mean the CFM they say, is the baseline that turns "the far zones feel starved" from a mystery into a measurement. That's the whole point of doing it honestly the first time: not that the air side is balanced today, but that anyone can prove whether it still is tomorrow.