Air balancing and TAB report field guide for HVAC

Test, adjust, and balance, and what the report delivers

Testing, adjusting, and balancing is the methodical work of measuring an air system, adjusting it, and setting it so each space gets the airflow the design called for. Testing is the measurement. Adjusting is moving the dampers and the fan. Balancing is the proportioning that gets every outlet to its share. The deliverable is not a balanced system you can feel. It is a report.

That report is the product. It lists design airflow against measured airflow for every diffuser, grille, and fan, the as-left damper positions, the fan settings, and the deficiencies the tech could not fix. The engineer of record reviews it and accepts it, or kicks it back. The owner gets a documented baseline of how the building was set the day it was handed over.

The thing that separates a real balance from a clipboard exercise is that the numbers have to reconcile. The sum of the outlet readings on a branch should track the branch traverse. The branch readings should track the fan total. When those three do not agree, you have duct leakage, a measurement error, or both, and a TAB tech who signs the report owns finding out which.

Where do you start a TAB? The sequence

You balance from the fan out, not from the registers in. Set total air at the air handler first, then proportion the branches against each other, then set the terminal outlets. The order is not preference. It is physics. Every time you close a damper somewhere, you change the pressure and the airflow everywhere else on that fan, so if you set the outlets before the trunk, the first trunk adjustment undoes all of it.

Before any of that, the system has to be ready, and a tech who skips the prerequisites balances a system that changes under him. The unit must be mechanically complete and running. Filters in and clean, because a filter that loads later moves every number you just set. Access doors closed and sealed. All volume dampers and splitter dampers open to start. Controls functional and the system in the mode you are balancing, usually full cooling airflow. Belts tight, sheaves set, rotation correct.

Then the sequence runs from the source outward. Read and set the fan total with a pitot traverse in the main. Traverse the main branches and proportion them to each other. Then proportion the outlets on each branch. Then come back and trim the fan to land the total, because the proportioning changes the system curve. NEBB and AABC procedural standards lay this out the same way: mains, then branches, then terminals, with the fan set to carry it. Confirm the exact steps against the standard the spec names.

What is proportional balancing?

Proportional balancing sets every outlet and branch as a ratio to each other first, then sets the fan to scale the whole system up to design at once. You do not chase each diffuser to its design number one at a time, because adjusting one outlet changes all the others and you would chase your tail until the belt wears out. You make them proportional, then you turn up the fan.

The method runs on the index outlet, the one reading the lowest percentage of its design airflow, usually the outlet farthest from the fan. You leave that damper wide open and never touch it. Then you throttle the closer, over-supplied outlets down until each one reads the same percentage of its design as the index. As you close the near dampers, air shifts to the far ones, so the index reading climbs while you work. When every outlet on the branch reads the same fraction of design, the branch is proportional.

Now the payoff. Increase the fan until the index outlet hits 100 percent of its design airflow. Because every outlet is proportional to the index, they all rise together and land at design at the same time. Set it once and the ratios hold: change the fan later and every outlet moves by the same fraction, so the balance does not come apart. That is the whole reason the trade balances this way instead of one register at a time.

Setting total airflow at the fan

Total air comes first because it is the budget every branch and outlet spends from. You establish it with a pitot traverse in the main supply duct, read it against the fan's rated performance, and adjust the fan to hit design total CFM before you proportion anything downstream.

Reading the fan means three numbers together: the traverse CFM, the fan speed in RPM, and the static the fan is working against. Plot them on the manufacturer's fan curve and you know whether the fan is where it should be or off in a corner of its curve where a small static change swings the airflow hard. A belt-drive fan gets dialed in with a sheave change or an adjustable-pitch sheave; a fan on a VFD gets dialed with the speed setting. Either way you are moving the fan along its curve to the design point.

The trap is setting total air against a static the duct will not actually run at. If the system static is high because a coil is dirty or a fire damper is half closed, you set the fan for that wrong static, and when the real restriction gets cleared the airflow walks off. Static pressure is the first thing to verify, which is why the external static reading frames the whole balance. See the duct static guide for measuring and reading TESP before you trust the fan setting.

Total air

The full airflow the fan delivers, measured by a pitot traverse in the main and set before branch balancing

Fan curve

The manufacturer plot of airflow against static at a given speed, used to confirm the fan is at its design point

How do you measure airflow in a duct?

Run a pitot-tube traverse: take velocity pressure readings at a grid of points across the duct, average them, convert to velocity, and multiply by the duct's free area to get CFM. A single reading in the center lies, because the velocity profile is not flat. It is fast in the middle and slow at the walls, so you sample the whole cross section and average.

The grid is not eyeballed. The log-Tchebycheff distribution, which traces back to ISO 3966 and is the method ASHRAE Standard 111 describes for traversing, places the points so the simple average of the readings is the true average velocity, weighting the points toward the walls where the boundary layer drags the air down. For a rectangular duct that is a grid of equal areas with the points set by the log-Tchebycheff spacing, commonly a minimum of 25 points. For a round duct you read along two or three diameters at the log-Tchebycheff radii, with the point count rising with diameter. Pick a straight run, ideally several diameters downstream of the last fitting, because elbows and takeoffs swirl the air and wreck the profile.

The equal-area method is the older cousin: divide the duct into equal areas and read the center of each. It works, but it slightly overstates velocity near the walls, so log-Tchebycheff is the more accurate choice and the one the standards point to. Either way, the traverse is the reference measurement the whole balance reconciles against.

The 4005 constant assumes standard air at about 0.075 pounds per cubic foot. At altitude or in hot or cold air the density shifts, so correct the velocity and the resulting CFM per the balancing standard before you record the reading, or the whole report drifts off by the same error.

Pitot traverse

A grid of velocity-pressure readings across a duct, averaged to get true average velocity and then airflow

Log-Tchebycheff

The traverse point spacing (from ISO 3966, used in ASHRAE 111) that makes the simple average of readings the true average velocity

Velocity pressure

The pressure from air motion that a pitot tube reads and converts to velocity, distinct from static pressure

Flow hood, anemometer, and the right tool for each

Three instruments cover most air-side measurement, and using the wrong one for the spot is how a balance reads clean and is wrong. The pitot traverse with a manometer is the reference for duct airflow: fan totals, branch totals, and any time you need the real number. The flow hood, or balometer, reads the airflow at a diffuser or grille directly by capturing all of it into a known throat with a velocity grid. The rotating-vane or hot-wire anemometer reads a velocity at a point or across a grille face, which you turn into airflow with the free area.

The flow hood is the workhorse at the outlets because it is fast and reads CFM straight off, but it has its own lies. It adds a little back pressure that can pull the reading low on a high-flow diffuser, so the better hoods carry a back-pressure compensation, and a slotted linear diffuser or a sidewall grille may need a specific hood top or a correction factor from the maker. The anemometer earns its keep on a return or exhaust grille and on supply registers a hood will not seal to, where you traverse the face and apply the effective free area, not the gross opening.

Match the tool to the location and the accuracy you need. Fan and branch totals get the pitot traverse. Outlets get the flow hood. Awkward grilles get the anemometer and a free-area calculation. Read the hood's manual for the diffuser type, because the correction is not optional on the ones that need it.

What air balancing tolerance is acceptable?

Outlets are commonly balanced within plus or minus 10 percent of design airflow, and the system total to within roughly plus 10 percent at the fan, but the project specification and the named TAB standard set the actual numbers and they are sometimes tighter. The plus or minus 10 percent at outlets is the figure you will see most often in a balancing spec. It is a common target, not a universal law, so read the contract.

Tolerances vary by what the space does. A general office runs the standard outlet tolerance. A pressure-critical space, an operating room, an isolation room, a cleanroom, or a lab fume hood makeup, gets a tighter band and a directional pressure requirement on top of the airflow number. Some specs hold the total to a one-sided tolerance, zero to plus 10 percent, so the system never delivers less than design but is allowed a margin over it to cover leakage. Confirm which framing the spec uses before you call a reading a pass.

Treat a reading that lands right at the edge of tolerance the way you treat a calculation that barely passes. Verify the instrument, the location, and whether the system was in the mode the spec balances to. A diffuser reading 91 percent of design is inside a plus or minus 10 percent band, but if it is the third one in a row reading low, the branch or the fan has a problem the outlet damper will not fix.

Outlet balancing at the damper, not the diffuser

Set the air with the volume damper at the branch takeoff, not the blades or the core of the diffuser. The diffuser face is the last place you want to throttle, because choking it there spikes the velocity through the remaining free area and turns a quiet outlet into one that whistles. The takeoff damper is upstream and out of the occupied space, so the noise it makes stays in the ceiling.

Work the branch proportionally. Read every outlet on the branch with the flow hood, find the one reading the lowest percentage of design, and leave its damper open. Throttle the higher ones down until they match that percentage, re-reading as you go because each adjustment feeds the others. When the branch reads even as a percentage, it is proportional and ready for the fan to scale it to design.

The diffuser pattern is a separate setting from the volume. Many diffusers have a pattern adjustment, blades or a core, that aims the throw without changing the CFM much. Set the volume at the damper, then set the throw at the diffuser so the air covers the space without dumping on someone's neck. The complaint that comes back is rarely the wrong airflow. It is the right airflow aimed wrong.

The fan, the sheave, and the VFD

Once the branches are proportional, you set the fan to carry the total, and the fan laws tell you exactly how the airflow, the pressure, and the power move when you change fan speed. Airflow rises in direct proportion to RPM. Static pressure rises with the square of RPM. The brake horsepower, and the motor amps that go with it, rise with the cube of RPM. That cube is the one that bites: a 10 percent speed increase to find a little more air costs about 33 percent more horsepower.

On a belt-drive fan you change speed by changing the sheave. An adjustable-pitch motor sheave opens or closes the pulley faces to trim speed; for a bigger change you swap to a fixed sheave of a different diameter and reset the belt. Calculate the new RPM you need from the airflow ratio, pick the sheave, then verify the result with a tach, because the calculation gets you close and the duct decides the rest. On a VFD you just command the speed, which is why variable-speed fans are quicker to balance, but the fan-law power penalty is identical.

Check the motor amps every time you speed a fan up. The cube law means a fan set faster than its design point can pull the motor past its full-load amps and its service factor, and a motor running over its rating does not last. Clamp the motor leg, compare it to the nameplate full-load amps, and if you are over, you have a fan asked to do more than its motor was sized for, which is a finding for the report, not something you bury by accepting the high amps.

Return, exhaust, and building pressurization

Supply is only half the balance. The return and exhaust have to be set too, because the difference between what you push in and what you pull out is what pressurizes or depressurizes the building. Balance the returns and exhausts proportionally the same way you balance supplies, then check the result at the building level, not just the grille.

Most commercial buildings are designed to sit slightly positive, with supply outrunning return plus exhaust by a modest margin so conditioned air leaks out at the doors instead of dirty, humid, or freezing air leaking in. Go negative and the building pulls infiltration through every crack, doors swing hard or stand open, and in a humid climate the wall cavities load up with moisture they were never meant to see. The makeup air unit exists to replace the air the exhaust fans throw out, and if it is short, the building makes up the difference by sucking it through the envelope.

Verify pressure with a manometer reading the building relative to outside, at a door or a dedicated port, with the system in its normal operating mode and the exterior doors closed. A target is often a small positive value the spec calls out. Kitchen hoods, lab exhaust, and large toilet exhaust are the usual culprits behind a building that will not hold pressure, because the exhaust got installed and the makeup air did not keep up. Set supply, return, and exhaust as a system, then prove the pressure the design intended.

Outside air and minimum ventilation

Setting and verifying minimum outside air is part of the balance, because the ventilation the occupants are owed is an airflow number like any other. ASHRAE Standard 62.1 is the basis most commercial specs use for the minimum outside air rate, and the TAB tech confirms the system actually delivers it at the mixing box or the outside air intake.

Measuring outside air is harder than measuring supply, and that is exactly why it gets faked. A wide-open outside air louver with a low face velocity reads poorly on a hood or an anemometer, and the intake is rarely a clean straight duct you can traverse. The common methods are a traverse of the outside air duct where one exists, a velocity grid at the louver with the free area applied, or a temperature-mixing calculation that backs the outside air fraction out of the mixed, return, and outside air temperatures when the temperature spread is wide enough to be meaningful. Each has error, so name the method you used in the report.

Set the minimum outside air damper position with the system at minimum, then confirm the economizer drives the damper open on a call for free cooling and returns to minimum when it should. An economizer stuck at minimum throws away free cooling; one stuck open overventilates and freezes coils in winter. The balance documents the minimum position and confirms the damper strokes, and the rest of the economizer logic is verified in commissioning.

Static pressure and the system effect

A balance that cannot be hit usually starts as a static pressure problem, so the external static reading is where a smart TAB tech looks first. If the fan is fighting more static than the design assumed, it is already low on airflow before you touch a single damper, and no amount of outlet adjustment makes air that the fan is not moving.

System effect is the static penalty you pay for bad fan inlet and outlet conditions that the catalog fan curve never saw. An elbow slapped right on the fan outlet, an inlet duct that turns hard into the fan, or a missing straight run spins or distorts the air entering or leaving the wheel, and the installed fan delivers less than the curve promises at the same speed. You measure it as a fan that needs more RPM than the calculation said to make its airflow. It is a design and installation problem, not a balance problem, and the honest move is to flag it rather than overspeed the fan to paper over it.

Read total external static, split it supply from return to localize the restriction, and tie it back to the airflow the fan is actually moving off its table. The duct static guide walks through measuring TESP, reading it against the blower table, and finding the side that carries the restriction, which is the diagnostic that frames the whole balance.

Field example: balancing a branch proportionally

A supply branch off a rooftop unit feeds five ceiling diffusers, each designed for 250 CFM, so the branch design is 1250 CFM. The fan traverse reads 1240 CFM total, basically on design, which is the trap. Total air is fine and every room is still wrong.

Read the outlets with the flow hood and the proportion is a mess. The first diffuser, closest to the takeoff, reads 310 CFM, 124 percent of design. They fall off down the branch to the last diffuser at 180 CFM, 72 percent. That far diffuser is the index. Leave its damper wide open. Throttle the near diffusers down, re-reading as you go, until all five read the same fraction of their design, which lands around 88 percent once the air has redistributed to the far end.

Now scale the branch with the fan. The outlets are proportional at about 88 percent, so the fan needs roughly a 14 percent bump in airflow to bring the index to 100 percent, and the others ride up with it. After the fan is trimmed, every diffuser lands between 244 and 255 CFM, inside the plus or minus 10 percent band. The total barely moved. What changed is that the air is where the design put it, which is the entire point of a balance and the thing a single fan-total reading will never tell you.

Instruments and calibration

A balance is only as honest as the instruments behind it, and uncalibrated instruments produce a report that is precise and wrong. Every measuring device the report relies on, the manometer, the flow hood, the anemometer, the tachometer, the temperature probes, carries a calibration date and a traceable reference, and the dates belong in the report.

The procedural standards require calibration on a defined interval against a known field standard, commonly annually, with the certificates kept on file. A manometer that drifted reads a static and a velocity pressure that are off by its drift on every point of every traverse, and a flow hood out of calibration biases every outlet the same direction. Before the day starts, zero the manometer to open air, and zero it again when you change locations or the readings stop making sense.

Carry the calibration paper to the job, because the commissioning agent or the engineer can ask for it, and a balance report from a certified firm is expected to show that the instruments were in calibration on the date of the work. An out-of-date sticker is a finding against the report, not a detail. The whole value of the document is that the numbers can be trusted, and that trust starts at the instrument.

What goes in the TAB report?

The TAB report lists design versus measured airflow for every outlet and every fan, plus the equipment data, the as-left settings, the instrument calibration record, and the deficiencies. For each air handler and fan it shows the nameplate data, the design and measured CFM, the fan RPM, the motor amps and voltage, and the external static. For each outlet it shows the design CFM, the measured CFM, the percentage of design, and the damper or device setting. The deficiency list captures what could not be balanced and why.

Format follows the standard the spec names. NEBB and AABC both publish report forms and procedural standards, and a project specified to one of them expects that body's forms filled out completely, with the firm's certification and the supervising professional's stamp where required. ASHRAE Standard 111 covers the practices for the measurements behind the numbers. The report is signed and certified by the balancing firm, which is how the building department and the engineer know a qualified TAB agency stands behind it.

Then the engineer of record reviews it. A report that shows everything inside tolerance with no deficiencies on a real building gets a hard second look, because a building with zero findings usually means a balance that was written, not performed. The credible report names the diffusers that would not make design, the fan that needed more speed than it should, and the spaces that could not be brought into pressure, because those are the items the engineer needs to rule on before the system is accepted.

Deficiencies and the back-check

A TAB tech finds the problems the installation hid, and the job is to report them, not to quietly absorb them. The recurring finds are duct leakage that bleeds air before it reaches the outlets, dampers left closed or fire and smoke dampers stuck part shut, a fan with the wrong sheave or a slipping belt, an undersized return or trunk that throttles the whole branch, and design airflows the duct simply cannot pass at the static the fan can make. Some you can fix on the spot. Some are above your scope.

Know the line between balancing and fixing. Opening a damper, trimming a sheave, and setting outlet volumes is balancing. Sealing a leaking duct, replacing a wrong sheave, cutting in a bigger return, or correcting a fan installed with a system-effect penalty is a repair that belongs to the installing contractor, and you flag it. A balance that hides a design or installation defect by overspeeding a fan or starving the far rooms is worse than no balance, because it certifies a problem as solved.

After any fix, you re-balance. Seal the leaking trunk and every downstream reading changes, so the proportioning you did before the fix is no longer valid. Replace a sheave and the fan total moves, which moves every branch. The back-check is not optional paperwork. It is the second balance the first one's findings made necessary, and the final report reflects the system as it was actually left, not as it read before the repair.

Where TAB meets commissioning

TAB and commissioning are different jobs that hand off to each other, and the balance feeds the functional tests. The TAB tech sets and documents the airflows; the commissioning agent verifies the system performs to the design intent across its sequences. A completed, accepted balance report is usually a prerequisite for the functional performance tests, because you cannot meaningfully test a VAV box sequence or an economizer changeover on a system whose airflows were never set.

The functional tests lean on the balanced state. A VAV box minimum and maximum airflow setpoint verification starts from the box airflows the balance established. An economizer test confirms the dampers stroke from the minimum outside air position the balance set. A building pressure control sequence is checked against the supply, return, and exhaust balance. When a functional test fails, the first question is often whether the balance underneath it was right, which is why the two scopes share so much data.

Treat the TAB report as a commissioning input, not a separate filing. The commissioning agent reads the deficiency list, the measured-versus-design tables, and the fan data, and uses them to target the functional tests and to separate a controls problem from an airflow problem. See the commissioning material for how functional testing and the issues log close out a system. A balance done in isolation gets redone; a balance done as the front end of commissioning closes the job.

What to document

The balance is the report, so the record is the work, not an afterthought to the work. For every system, fan, and outlet, capture enough that the engineer can reconcile the numbers and the next tech can pick up the baseline years later.

Capture the system or outlet identifier, the design airflow, the measured airflow, the percentage of design, the damper or device setting as left, and the pass or fail against the spec tolerance. For each fan add the RPM, the motor amps and voltage, the external static, and the sheave or VFD setting. Add the instrument calibration dates, the measurement method used for outside air, and the deficiency list. The table below is the core outlet record; the fan and system data ride alongside it.

Common mistakes

• Balancing outlets before total air is set, so the first trunk adjustment undoes the outlet work.
• Throttling at the diffuser face instead of the takeoff damper, which makes the outlet whistle.
• Reading a single center-point velocity instead of a full pitot traverse and trusting the number.
• Ignoring return and exhaust, so the building never holds the pressure the design intended.
• Setting the fan against a false static from a dirty coil or a closed damper, so airflow walks off later.
• Overspeeding a fan past its motor amps to mask a system-effect or undersized-duct problem.
• Using instruments with no calibration date and forgetting to zero the manometer between locations.
• Reporting everything in tolerance with no deficiencies on a building that plainly has them.

Field checklist

Standards and references

The TAB procedural standards and the certifications behind them come from NEBB, the National Environmental Balancing Bureau, and AABC, the Associated Air Balance Council, with TABB also certifying technicians and firms. Each publishes a procedural standard and report forms, and a project specifies one of them; the named standard governs the procedure, the forms, and the certification on the report. Confirm which one the spec calls out before you start, and use that body's current edition.

ASHRAE Standard 111, Practices for Measurement, Testing, Adjusting, and Balancing of Building HVAC Systems, covers the measurement methods, including the duct traverse. The log-Tchebycheff traverse point spacing traces to ISO 3966. ASHRAE Standard 62.1 sets the minimum ventilation rates the outside air balance has to deliver, and 90.1 sets energy limits the system has to live within. SMACNA owns duct construction and the leakage classes behind the air you lose before it reaches a register, and also publishes TAB guidance. AMCA rates the fan performance the fan curve traces back to.

The design the balance is checked against runs on the ACCA Manuals for residential and light commercial work, Manual J for the load, Manual S for equipment selection, and Manual D for the duct. Above all of it sits the project specification and the engineer of record, who set the tolerance and accept the report. Cite the body that owns the point, and verify the edition and section, because these documents revise on their own cycles.

Units, terms, and conversions

Air balancing carries its own vocabulary and a few unit systems, so the same quantity reads differently across a balance report, a manufacturer sheet, and a metric drawing.

Airflow is CFM, cubic feet per minute, in the field, and liters per second or cubic meters per hour in metric sources, where 1 CFM is about 0.472 liters per second. Velocity is feet per minute, often written FPM, or meters per second in metric. Static and velocity pressure are inches of water column, in. wg or in. w.c., where 1 in. wg is about 249 pascals. Tolerance is a percentage of design, commonly plus or minus 10 percent at outlets. The work itself goes by TAB, test and balance, or air balancing, and the firms are certified by NEBB or AABC.

TAB

Testing, adjusting, and balancing: measuring an air system, adjusting it, and setting it to design, then reporting it

Proportional balancing

Setting outlets and branches as a ratio to the index (lowest-percent) outlet, then scaling the whole system with the fan

Traverse

A grid of velocity readings across a duct, averaged to get true airflow; log-Tchebycheff spacing is the accurate method

CFM

Cubic feet per minute, the airflow unit a balance sets at each outlet and fan

Tolerance (+/- 10%)

The acceptable band around design airflow, commonly plus or minus 10 percent at outlets, set by the spec

NEBB / AABC

The bodies that publish TAB procedural standards, report forms, and the firm and technician certifications

Index outlet

The outlet reading the lowest percentage of design, left wide open as the reference for proportional balancing

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• How do I measure total external static pressure?
• Do I charge by superheat or subcooling?
• What is a hydronic balancing valve?
• How do you balance a hydronic system?
• What is an air curtain?