Two Complaints From One Bay
The tenant on level four logs the same desk twice a day. At half seven it is cold, and the person sitting there has a fleece over the back of the chair from October to April. At four in the afternoon the same desk is warm enough that they have opened the manual override on the diffuser above them, which the controls contractor has now reset three times. Nothing about the terminal serving that bay is broken. The box modulates, the valve strokes, the sensor reads within half a degree of the one beside it, and the trend logs show it doing exactly what the sequence tells it to.
What is wrong is upstream of all of it. The coil in that terminal was selected against a heating load, the load run took a U-factor off the window schedule, and the number on the window schedule described the glass rather than the wall. Between the centre of a lite and the finished bay there is a frame, a mullion every 1.5 metres, a transom, a slab-edge cover and a spandrel, every one of them a worse thermal path than the glass whose number got used. The coil is undersized because the wall is worse than the arithmetic said — and the afternoon complaint is a separate failure entirely, the same glass working in the other direction, dumping solar gain into a zone whose terminal is already at its heating minimum.
So the sizing problem is not one calculation. It is a short chain that has to be walked in order, because each link changes the meaning of the next: what the wall really loses, what that does to the inner face the occupant is sitting beside, what the sun does to the same wall on the same day, how little air the terminal is permitted to hold at minimum, and only then what the coil has to put back into that air. Skip a link and you will size a perfectly correct coil for a building that does not exist.
The Bay, From the Plenum Down
It helps to draw the thing before calculating anything about it, because a perimeter zone is one of the few parts of a building where the mechanical work and the envelope work are stacked physically on top of each other. Everything that sets the heating load is below the ceiling line. Everything that answers it is above the ceiling line, discharging through a slot a hand's width from the glass. In a core zone those two conversations happen a long way apart; here they are separated by about 300 millimetres of ceiling void, and the drawings that show them belong to two different consultants.
Read the stack from the top and the sizing sequence falls out of it. The branch off the main sets what the terminal can ever be given. The terminal's own minimum sets what it will actually be holding when the coil is called. The coil sets the discharge temperature. The diffuser decides whether that discharge reaches the floor or slides back into the return. And the wall below decides how much of it is needed in the first place.
One perimeter bay, taken apart from the plenum down
- Primary air branch and terminal casing — sized on the bay's cooling peak, which is the one condition of the year the box is never reheating in HVAC Duct Size Calculator
- Reheat coil on the terminal discharge — does its work at the terminal's heating airflow, not at the airflow the casing was selected for VAV Terminal Reheat Coil Sensible Heat Capacity Calculator
- Ceiling plane and perimeter slot diffuser — aimed along the glass to break the falling cold film, which is a different job from distributing air to the room Duct Air Velocity Calculator
- Vision glass, transom and mullion — three thermal paths in parallel with three different U-factors, only one of which is printed on the glass order Window Wall Area-Weighted U-Factor Calculator
- Sill, slab edge and occupied floor — where the cold air sheeting off the glass arrives, at ankle height and away from every wall sensor
The Schedule Describes the Glass; the Load Needs the Wall
A U-factor is only meaningful once you know what area it was measured across. Centre-of-glass is the best number the assembly will ever produce, because it excludes the sealed edge, the spacer, the sightline and the frame. A whole-product rating determined under ANSI/NFRC 100, Procedure for Determining Fenestration Product U-Factors, already blends the frame in — but it blends in the frame of the tested product size, and a curtain wall bay in a real building is not that size. Widen the lite and the assembly improves; narrow it and add an intermediate mullion and it gets worse. In Europe the same distinction runs through ISO 10077-1 for the window as a whole and ISO 10077-2 for the frame profile by numerical method, with BS EN ISO 12631 covering curtain walling specifically.
For a heating load you want the number that applies to the actual area of wall the zone is standing behind, which means weighting each component by how much of that area it occupies. That is a two-line piece of arithmetic and it is routinely skipped, usually because the glass number was already to hand and the frame number was not. It is worth doing even when the answers seem close, because the frame fraction is where the surprises live: a bay with a single large lite might be twelve per cent frame, and the same elevation cut into narrower modules for shipping or for a shading fin can arrive at twenty-five per cent without anyone re-issuing a thermal figure. A frame at three or four times the U-factor of the glazing does not need much area to move the composite.
Two things the weighting does not catch, and both belong in the same conversation. The first is the spandrel and the slab-edge zone: a shadow box or an insulated back-pan is a different assembly again, it sits directly above the vision glass in every bay, and on a floor-to-floor curtain wall it can be a third of the elevation. Weight it in as a third component or run the vision assembly and the spandrel assembly separately and add the losses. The second is linear heat flow at the junctions, which no area-weighted average can represent at all: the mullion-to-slab connection, the head and sill and the perimeter of every spandrel each carry a linear thermal transmittance, calculated in detail under ISO 10211 or taken from the tabulated defaults in ISO 14683. On a well-glazed elevation those linear terms are not a rounding error.
Finally, check what was actually bought. The gap between the specified assembly and the delivered one opens during value engineering, not during design: warm-edge spacer swapped for aluminium, a thermal break narrowed, argon dropped from the shop drawings, a coating moved from surface two to surface three for a visual reason. All small on their own, all landing on the same coil. Ask for the fabricator's assembly figure and the test report behind it before the load is frozen.
| Number quoted | Area it applies to | What it leaves out |
|---|---|---|
| Centre of glass | The middle of a lite, away from every edge | Spacer, edge seal, sightline, frame — everything that makes a bay worse than a sample |
| Whole product, NFRC 100 | One tested product at one tested size | The frame proportion of your bay, which is set by the module width, not by the test |
| Area-weighted assembly | The vision area of the bay as built, glass plus frame | Spandrel and slab-edge zones, and all linear junction losses |
| Elevation average | Vision, spandrel and opaque back-pan together | Junction psi-values, and any local difference between a corner bay and a typical bay |
| Linear transmittance, ISO 10211 or 14683 | A junction, per metre run rather than per square metre | Nothing the area terms cover — it is the piece they cannot represent |
Take the glass area and the opaque frame area straight off the shop drawing for one typical bay, not off the window schedule, and see what the composite does when the frame fraction climbs from a single-lite module to a mullion-heavy one.
The center-of-glass (or whole-glass) U-factor for the glazing component.
The total visible glass area in the window wall assembly.
The U-factor for the frame/mullion component, typically higher (less insulating) than glass.
The total opaque frame/mullion area in the window wall assembly.
Composite area-weighted U-factor
0.3769 BTU/(hr·ft²·°F)
They open the calculator with your figures already in it
Window Wall Area-Weighted U-Factor Calculator: 0.3769 BTU/(hr·ft²·°F) — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Add the equipment this sizes
This result is a specification — 0.3769 BTU/(hr·ft²·°F) — not a quantity. Put the thing it sizes into your project: how many, what you call it, and your supplier’s price.
What this calculation does not cover
- The assembly is split into two areas here, glass and frame, while the NFRC method the source names uses three: centre-of-glass, a 63.5 mm (2.5 in) edge-of-glass band where the spacer and sealant conduct around the perimeter of each lite, and frame. Entering a centre-of-glass U-factor across the whole glass area therefore averages that colder band in as though it were centre glass, and the composite lands lower — better — than the rated whole-window U-factor for the same product.
- Nothing checks that the glass area and the frame area actually tile the opening. Taking the frame area from the outside frame footprint while the glass area comes from the daylight opening double-counts the overlap and pushes the composite up; missing a mullion, a transom or a stack joint leaves that area out and pulls it down. Neither mistake raises an error, because the two figures are only ever used as weights against their own sum.
- Area weighting does not capture thermal bridging. The anchors, brackets and the slab-edge and head-and-sill details that tie a window wall back to the structure conduct around both of the areas entered here, and they are quantified by a linear psi-value or a two-dimensional thermal model rather than by an average of two U-factors.
- The output is a conduction rating, not a heat loss and not an energy performance figure. It is computed in W/m²K and restated as BTU/(hr·ft²·°F) when the page is set to US units, while the two U-factor boxes above stay in W/m²K whichever way that switch is set. Multiply it by the assembly area and the design temperature difference to get watts, and handle solar separately, since gain is governed by SHGC and shading — a low composite U-factor says nothing about overheating on a west elevation. Air leakage through the perimeter seal and the joint to the slab is outside the rating as well.
- An area-weighted average is a mean, and condensation is a local event. The sight line, the spacer and the frame itself all run colder than the composite figure implies, so an assembly that meets a U-factor target can still run wet at the glass edge; that question needs the manufacturer's condensation resistance figure or a thermal model of the actual profile.
A Number the Occupant Can Feel
Conduction through the wall is the part that reaches the coil, but it is not the part that generates the complaint. What the person at that desk responds to is the temperature of the surface they are sitting a metre from and the sheet of cold air falling off it. Both come from the same assembly figure, and both can fail while the zone air temperature sits obediently on setpoint — which is exactly why the trend log looked clean and the sensor reading was defensible.
The inner surface runs colder than room air by an amount proportional to the assembly U-factor and to the temperature difference across it, moderated by the interior air film. Push it far enough below room air and two separate things happen. Radiant asymmetry becomes measurable, which ASHRAE Standard 55, Thermal Environmental Conditions for Human Occupancy, and ISO 7730 both treat as local thermal discomfort in its own right rather than as something an air temperature compensates for. And the film of air touching the glass becomes dense enough to run downward, gathering speed over the storey height and arriving at ankle level as a draught. That second effect scales with the height of the glass as well as with the surface temperature, which is why a 3-metre floor-to-ceiling bay with a respectable U produces draughts a 1.4-metre punched window with a worse U never would. Raising the supply temperature does not fix it; putting air movement along the glass does.
Take the same surface temperature far enough and it stops being a comfort question and becomes a water question. Condensation risk is assessed against the inner surface temperature under winter design conditions and the dew point of the room air, and the test methods behind the published ratings are AAMA 1503, Voluntary Test Method for Thermal Transmittance and Condensation Resistance of Windows, Doors and Glazed Wall Sections, and NFRC 500, Procedure for Determining Fenestration Product Condensation Resistance Values. Neither produces a number you can use directly against a specific room; both give you a comparative rating and a method. Run the surface temperature for the assembly you actually have, against the interior humidity the building will actually be held at, and do it before anyone specifies a humidifier — a raised winter setpoint on humidity is the single most common way a facade that was fine becomes a facade that streams.
Put the composite assembly U-factor from the previous step in here rather than the glass figure, because the frame and the sightline are where the surface runs coldest and where the water appears first.
The interior air design temperature used for the winter condensation check.
The exterior winter design temperature for the project location, typically a 97.5% or 99% design condition.
The overall U-factor of the spandrel/vision glass assembly being checked.
The interior air film coefficient, commonly taken as 8 W/m²K per standard convention.
The dew point of the interior air, based on its temperature and relative humidity.
Interior glass surface temperature
57.2 °F
The calculated interior glass surface temperature stays above the interior air's dew point shown below, so condensation is not predicted under these steady-state design conditions. This checks room-side vision/spandrel glass surface condensation risk only, per NFRC 500/AAMA 1503 steady-state methodology. Shadowbox CAVITY condensation — a distinct, common failure mode driven by trapped cavity moisture and solar vapor drive behind the spandrel panel — is NOT covered here and needs separate ventilation/vapor analysis per GANA/NGA guidance. No risk predicted under these conditions is not the same as none. The conditions are the ones you entered, and one surface is not the assembly.
- Interior air dew point
- 50 °F
They open the calculator with your figures already in it
Curtain Wall Spandrel Condensation Risk Checker: 57.25 °F — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
What this calculation does not cover
- Models conduction only. Warm indoor air pushed out through an imperfect air seal — a splice joint, an anchor pocket, a gasket that has taken a set — carries far more moisture into the assembly than diffusion does, and it condenses wherever it cools rather than on the face being checked. A wall that passes on paper can still run with water, or ice, inside the mullion, and the first sign of it is usually a stain at the head of the panel below.
- Uses the exterior air temperature, and the outer surface does not sit at it. On a clear, calm night the glass radiates to the sky and settles several degrees below the surrounding air, which drags the interior surface down with it — and those are exactly the nights condensation forms. A result that clears the dew point by a degree or two at the design air temperature has no margin left for the sky.
The Same Bay, Six Hours Later
A west-facing perimeter zone in January is a heating zone at eight in the morning and a cooling zone at half past three, on the same day, with no change in the weather. That is not a control fault and it is not something to design out; it is the defining behaviour of a glazed perimeter and the reason the zone has a reheat coil rather than a radiator. What matters is that the two conditions are sized separately and that neither is allowed to be discovered on site.
Solar gain through the assembly is governed by the solar heat gain coefficient — rated at normal incidence under ANSI/NFRC 200 — modified by whatever shading actually intervenes, and the sun that causes the problem is the low one. A horizontal overhang sized against a summer noon altitude does nothing at a winter afternoon altitude of twenty degrees, because the shadow it casts is short and the sun is coming at the glass nearly face-on. The profile-angle method in the fenestration chapter of the ASHRAE Handbook—Fundamentals is what turns a projection depth and a solar altitude into a shaded fraction of the glass, and running it at the low angle rather than the high one is what separates a shading device that works from one that photographs well.
The consequence for the terminal is specific. That afternoon the box is cooling a zone whose neighbours are still calling for their winter condition, and if the perimeter has been zoned generously it is reheating one half of its own area while overcooling the other. Sizing the cooling condition is therefore part of sizing the heating condition: the primary airflow needed at the solar peak sets the casing, and the casing's turndown sets the minimum airflow the coil has to heat.
Test the overhang at the solar altitude that produces the afternoon complaint rather than at a summer noon angle, because a projection that shades the whole lite in June frequently shades none of it in February.
The glass's own solar heat gain coefficient with no shading applied.
The sun's angle above the horizon at the time of day/year being checked.
How far the horizontal overhang or shading device projects out from the wall face.
The vertical distance from the underside of the overhang down to the top of the window.
The full height of the window glass being shaded.
Effective SHGC
0.15 SHGC
This simplified geometric approximation is for a window facing directly toward the sun (zero azimuth difference) and treats shaded glass as receiving negligible direct gain, which overstates shading benefit since diffuse/reflected radiation still reaches shaded glass. For precise energy modeling, use ASHRAE 90.1/LBNL Projection Factor tables or full solar gain software.
- Shaded fraction of window
- 0.63
They open the calculator with your figures already in it
SHGC Effective Shading Fraction Calculator: 0.1495 SHGC — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Add the equipment this sizes
This result is a specification — 0.15 SHGC — not a quantity. Put the thing it sizes into your project: how many, what you call it, and your supplier’s price.
What this calculation does not cover
- One instant, not a season. This is the shading at a single solar altitude — one time of day on one day of the year — and a fixed overhang cannot be tuned to more than a few of them. Whether the overhang is any good comes from the balance across the year: summer shading set against the winter sun it also blocks, and how it behaves through the mid-afternoon hours when a cooling peak usually lands rather than at noon, when it always looks best.
- The overhang is treated as infinitely wide. The shadow here drops straight down the glass with no ends to it, but a real overhang stops somewhere, often at or just past the jambs, and sun reaches in around both sides for much of the day. The narrower the overhang relative to the window, and the further the sun swings off the facade's normal, the more that end effect eats into the shaded fraction this returns.
- A horizontal overhang is close to useless on east and west glass, and nothing here will say so. The sun is low whenever it faces those elevations, so the shadow drop the formula returns is small at exactly the hours the gain is worst. Those windows need vertical fins, exterior blinds or glass with a lower SHGC — size an east or west shade off a noon altitude figure and you build a device that shades nothing when it matters.
Why the Terminal Cannot Simply Close
The instinct when a zone is being overheated by its own supply air is to shut the box further. Three separate things stop you, and all three have to be settled before a coil duty means anything.
The first is ventilation. Under the Ventilation Rate Procedure of ASHRAE Standard 62.1, Ventilation for Acceptable Indoor Air Quality, the zone has an outdoor air requirement built from its floor area and its population, and in a single-duct VAV system that outdoor air arrives diluted in the primary air. Cut the primary airflow and you cut the outdoor air with it. Worse, the standard's zone air distribution effectiveness table penalises warm air supplied and returned at the ceiling once the supply runs a stated margin above room temperature — so reheating hard can raise the outdoor air the zone must be given at the moment you were trying to reduce its airflow. Read that value out of the edition the project is filed under; it has moved between revisions.
The second is the energy standard. ASHRAE Standard 90.1, Energy Standard for Buildings Except Low-Rise Residential Buildings, contains a simultaneous heating and cooling limitation whose zone control provisions cap how much primary air a terminal may still be passing before it is permitted to start reheating. The cap is expressed as the largest of several competing floors — a fraction of the zone peak, the ventilation minimum, and a rate per unit of floor area — and both the fractions and the exceptions have changed edition to edition. The corresponding requirement in the International Energy Conservation Code commercial provisions, and in Approved Document L for non-domestic buildings in England, is aimed at the same waste by different wording.
The third is the diffuser, forgotten because it is somebody else's drawing. A slot has a flow range below which it stops behaving like a slot: the throw collapses, the jet leaves the ceiling plane close to the outlet, and the air the coil just warmed goes back into the return without reaching the person it was heated for. Hold all three together and you arrive at the modern answer, a dual-maximum sequence — a low cooling minimum for the hours the zone is neutral, and a separate, higher heating airflow the box steps up to when the coil is working. ASHRAE Guideline 36, High-Performance Sequences of Operation for HVAC Systems, sets it out in full, and it changes the selection completely, because the coil is now sized at the heating maximum rather than at the cooling minimum.
The Coil Is Sized at a Flow the Box Rarely Sees
With the airflow settled, the coil duty is the least contentious arithmetic on the page. Sensible heat added to an airstream is the mass flow multiplied by the specific heat multiplied by the temperature rise, and the trade collapses that into a constant: about 1.08 in imperial units, where 60 minutes an hour times 0.075 pounds per cubic foot times 0.24 Btu per pound-degree multiplies out, and about 1.2 watts per litre per second per kelvin in metric, from 1.2 kilograms per cubic metre and roughly 1.006 kilojoules per kilogram-kelvin. Multiply, and you have the sensible capacity the coil has to deliver.
The trap is in which airflow goes into it. Use the terminal's design maximum and the answer is a large coil that will never see that flow while heating, because a box at its cooling peak is not reheating. Use the heating airflow the sequence actually commands — the heating maximum in a dual-maximum control, or the single minimum in an older single-maximum box — and the answer is both smaller and correct. Then check the discharge temperature that duty implies at that flow, because a rise big enough to satisfy the load at a small airflow can put the leaving air above what the diffuser will deliver to the floor and above the margin the ventilation effectiveness table punishes.
The constant has conditions attached, and the one that catches people out is altitude. It encodes standard air density, so on a job at 1,500 metres the real factor is meaningfully lower and a coil sized on the sea-level number arrives short exactly when the weather is worst. Entering air temperature matters too: a terminal downstream of an air handler running supply temperature reset heats from a different starting point in shoulder season than in deep winter, and the rise is measured from wherever the primary air actually is.
None of this is a coil selection. A sensible-heat duty tells a manufacturer's selection program what to aim at; the program decides rows, fin spacing, circuiting, air-side pressure drop and — for a hot water coil — the flow and entering water temperature needed to reach that duty at the temperature the plant is really running, which on a condensing-boiler or heat-pump plant can be far below the flow temperature the coil schedule was drafted around. Coil ratings are established under AHRI Standard 410, Forced-Circulation Air-Cooling and Air-Heating Coils, and the terminal as a whole is tested under ANSI/ASHRAE Standard 130, Methods of Testing Air Terminal Units. Take the duty to the selection; do not take it to the order.
| Assumption inside the constant | Where it stops holding | What to do instead |
|---|---|---|
| Standard air density at sea level | Any project at altitude — the factor tracks density downward | Correct the constant for site elevation before sizing, not after |
| The airflow you enter is the airflow while heating | Single-maximum boxes at cooling minimum, dual-maximum boxes at heating maximum | Take the flow from the control sequence, not from the terminal schedule |
| A fixed entering air temperature | Air handlers running supply air temperature reset | Measure the rise from the primary air condition at the hour the load peaks |
| Sensible heat only | Nowhere, for a dry reheat coil — this one is safe | No adjustment; reheat adds no moisture |
| Water hot enough to reach the duty | Low-temperature plant, condensing boilers, heat pump sources | Re-select at the real entering water temperature and flow, not the schedule's |
Enter the heating airflow the sequence commands and the rise measured from the primary air temperature you will really have, then treat the answer as the target you hand to a selection program rather than as a coil size.
The VAV terminal's airflow rate at the reheat condition.
The desired air temperature increase across the reheat coil, from entering to leaving air.
Required reheat coil sensible capacity
25,900 BTU/hr
This is a simplified sensible-heat sizing check using the standard-air 1.08 factor — it is not a substitute for full manufacturer coil selection, which also accounts for entering water/electric element temperature, coil rows, fin spacing, and actual air density (the 1.08 factor itself shifts at high altitude).
They open the calculator with your figures already in it
VAV Terminal Reheat Coil Sensible Heat Capacity Calculator: 25,917 BTU/hr — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Add the equipment this sizes
This result is a specification — 25,900 BTU/hr — not a quantity. Put the thing it sizes into your project: how many, what you call it, and your supplier’s price.
What this calculation does not cover
- The temperature rise is not a free choice, and nothing here stops you entering one the room cannot use. Air leaving a ceiling diffuser much more than about 15 °F (8 °C) above room temperature stops mixing and stratifies at the ceiling — ASHRAE 62.1 penalizes a zone's air distribution effectiveness for exactly that condition. A large rise on a small minimum airflow therefore buys a capacity that heats the plenum while the occupied zone stays cold, and the remedy is more airflow, not more rise.
- Nothing here sizes what feeds the coil. The same duty is either an electric element pulling a breaker and a feeder off a panel, or a hot-water coil demanding a flow rate through a control valve — and across a floor of thirty terminals the summed electrical demand, or the pump flow and pipe size to serve them, is a design quantity in its own right that a per-box capacity never adds up.
Warm Air Does Not Want to Come Down
Every figure so far assumes the heat arrives where the person is. Buoyancy disagrees. Air supplied warm from a ceiling is already lighter than the room it is entering, so the jet loses its downward component quickly, and the greater the temperature difference the sooner it happens. Beyond a certain rise the warm air simply spreads along the ceiling, stratifies, and finds a ceiling return, leaving the occupied zone below it colder than the thermostat believes and the coil running to heat a plenum.
So the discharge temperature is a distribution limit before it is a coil limit, and the answer to a stubborn perimeter is usually more air rather than hotter air. It is also the argument for putting the supply where the problem is: a linear slot along the glass, throwing down the face of the wall, breaks the falling cold film where it forms. Where the load is genuinely large, perimeter radiation under the sill answers the draught directly and lets the terminal go back to ventilation and cooling — worth pricing before assuming overhead reheat carries all of it. Room air distribution has its own test method in ASHRAE Standard 113, Method of Testing for Room Air Diffusion, and throw data is published at stated temperature differences for this reason: the catalogue figure was almost certainly measured isothermal.
One Zone, or Four
How well any of this works is decided by where the zone boundaries were drawn, which happens early, on a floor plate with no furniture on it yet. A perimeter zone is conventionally taken as a band of the order of four to five metres deep from the glass, because beyond that the envelope stops dominating and internal gains take over. That depth is a convention drawn from practice and from the load calculation chapters of the ASHRAE Handbook—Fundamentals, not from a code clause, and on a heavily glazed floor with low gain density the wall's influence reaches further in.
The error that costs the most is a zone that spans two orientations. Wrap a single terminal around a corner and you have joined a bay that peaks at nine in the morning to a bay that peaks at four in the afternoon, and the box will spend most of the day satisfying the average of two conditions that never occur together. The sensor sits in one of them. Corners deserve their own terminal for the same reason they get their own load: they see two elevations, they have more envelope per unit of floor area than anywhere else on the plate, and they are where the complaint concentrates.
A zone that is correct on the base-build plan also goes wrong when the fit-out puts a full-height meeting room across two modules of glass, or when a tenant demises the floor down a line the ductwork does not follow. None of that is predictable at design stage, which is an argument for leaving flow setpoints adjustable — not for oversizing the coil, which makes the afternoon condition worse rather than better.
Proving It With the Building Standing Up
The chain ends where every chain in this trade ends, at measurements taken on the finished floor. Airflow first: the terminal has to be verified at its heating flow as well as at its maximum, and a box balanced only at maximum is a box whose reheat condition nobody has ever seen. Procedural standards for that work are published by NEBB and AABC, ASHRAE Standard 111, Measurement, Testing, Adjusting, and Balancing of Building HVAC Systems, sets out the measurement methods, and the commissioning process around all of it is described in ASHRAE Guideline 0 and ASHRAE Standard 202, Commissioning Process for Buildings and Systems.
Then check the two claims this guide has been making. Put a thermometer in the discharge with the coil at full call and confirm the rise the arithmetic promised — if it is short, the water temperature or the flow is the first suspect, not the coil. And on a genuinely cold morning, before the sun is on the elevation, take an infrared reading of the inner glass surface and the sightline at the mullion and compare it with the surface temperature the assembly figure predicted. That single reading tells you whether the U-factor you loaded the whole calculation with was real, and it takes about a minute.
Finally, trend rather than sample. A perimeter terminal that looks correct on a commissioning day in April will show you its actual behaviour over a week of logs in January: how many hours it spends at heating maximum, whether it ever leaves it, whether the reheat valve and the cooling call overlap in the afternoon, and whether the zone temperature and the discharge temperature ever agree. Those logs are also what settles who owns the complaint, a conversation a glazed perimeter will produce whether or not anyone prepared for it.
- Verify the terminal at both flows — cooling maximum and the heating airflow the sequence commands — and record which sequence the box is actually running.
- Measure the discharge temperature with the coil at full call and compare the rise against the duty the sizing assumed.
- Check the entering water temperature and coil flow at the same moment, since a short rise is usually a water problem wearing a coil's clothes.
- Take infrared surface readings on the vision glass and at the mullion sightline on a cold morning before sun-up, and compare against the predicted surface temperature.
- Confirm the outdoor air the zone receives at its heating airflow, not only at its cooling maximum.
- Trend zone temperature, discharge temperature, primary airflow and reheat valve position through a full winter week before signing the zone off.
Compare what the box measured against what the sequence asked for at each flow setpoint, and treat the heating flow reading as the important one, because that is the condition the coil was sized in.
The zone's specified design airflow from the mechanical drawings.
The zone's actual field-measured airflow.
Airflow balance deviation
-8 %
Within the commonly-specified ±10% TAB tolerance.
They open the calculator with your figures already in it
HVAC Zone Airflow Balancing Percent Calculator: -8 % — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
What this calculation does not cover
- The deviation is only as trustworthy as the reading behind it. A balancing hood is accurate to a few percent at best, and considerably worse when it is not matched to the diffuser, is held off square, or back-pressures a low-flow outlet enough to depress the very flow it is measuring. So a zone showing a small deviation may be a hood problem rather than a damper problem — confirm a marginal reading by a second method before anything gets adjusted.
- Zones are not independent, so any one zone's percentage is a moving target. Closing a damper on one branch pushes that air into every other outlet on the same trunk, which means a zone set exactly to design early in the pass has moved by the time the last zone is set. Balancing is proportional and iterative across the whole system, and a per-zone figure only means something on a final pass with everything else settled in position.
- Sitting inside the tolerance is not the same as delivering what the zone needs. Where the design airflow was set by ventilation rather than by heating or cooling load, a reading 9 % under design passes this arithmetic while the space receives 9 % less outdoor air than it is required to. And a system whose zones all sit on the low side of tolerance sums to a supply total well short of the fan's design flow with every individual number still reading acceptable.
The six numbers a perimeter terminal is sized on
Settle these in order before a coil duty is issued — each one changes what the next one means, and the coil duty is the only one anybody usually writes down.
- Area-weighted assembly U-factor for one typical bay — Glass and opaque frame area off the shop drawing, with spandrel and slab-edge zones weighted in separately and junction psi-values noted alongside.
- Inner surface temperature at winter design — Run it against the composite figure and against the room's actual humidity setpoint, since comfort and condensation both fail here before the air temperature does.
- Effective SHGC at the low afternoon sun — Test the shading at the altitude that causes the complaint; a projection sized on summer noon usually shades nothing in winter.
- Heating airflow the control sequence commands — The heating maximum under a dual-maximum sequence, or the single minimum otherwise — bounded by the ventilation requirement and the diffuser's usable range.
- Coil sensible duty and the discharge temperature it implies — Correct the sensible-heat constant for site altitude, measure the rise from the real primary air condition, and check the leaving temperature against what the diffuser can push to the floor.
- Entering water temperature the plant will really deliver — A low-temperature or heat-pump plant re-sizes the coil rather than adjusting it; confirm before selection, not at commissioning.
Opens the calculators above on one screen with the dimensions from this article already filled in. Quantities only — this site publishes no price list, because local prices vary too much to publish honestly.