Six-Forty on the Coldest Morning of the Year
The unit came out of night setback at half five, the space sensors are still reading 16 °C, and the outdoor-air damper has just opened to its minimum position because the schedule says the building is occupied. Ninety seconds later the low-limit trips, the supply fan stops, and the manual reset button is behind a filter rack that takes two people to open. Nobody has done anything wrong. The building simply asked a mixing box to hand a coil air it was never going to be able to warm from that starting point.
The awkward part of winter commissioning is that almost every number involved is defensible on its own. The minimum outdoor-air fraction is defensible: it comes out of ASHRAE Standard 62.1, Ventilation for Acceptable Indoor Air Quality, or out of whatever the local ventilation code adopts in its place. The low-limit setting is defensible: it protects a coil that will split if it freezes. The morning warm-up sequence is defensible. It is the arithmetic between them that nobody checks, and that arithmetic is what decides whether the building starts.
So the work on a cold morning is not adjustment, it is prediction. Before touching a setpoint, know what dry-bulb the mixing plenum will actually deliver at the outdoor design condition, know how far that sits above the low-limit setting, and know how much moisture the humidifier is expected to put back into an airstream that arrived from outdoors with almost none. Those two figures — mixed-air temperature and moisture addition rate — settle most of what a winter start-up argues about.
What the Air Passes Through, and in What Order
A central-station unit in a cold climate is a short series of components in a sheet-metal box, and the winter behaviour of the whole thing is decided by the first one and the last one. The first is the mixing box, where outdoor air at design temperature meets return air at room temperature and produces a single stream that everything downstream inherits. The last, in a great many cold-climate units, is a steam dispersion tube putting back the moisture the outdoor air never brought with it.
Between them sit the parts that get all the attention on a hot day and very little on a cold one. Worth noting on the drawing: some units carry a preheat or frost coil ahead of the filter bank, or out in the outdoor-air duct before the mixing box entirely, which changes the whole freeze story — the coil sees undiluted outdoor air, and it must then be arranged so it cannot freeze at part load. Face-and-bypass, integral face-and-bypass, a glycol run-around loop or a two-position steam valve with a steam-distributing tube are the usual answers; a plain modulating valve on a steam coil facing outdoor air is the one arrangement that reliably fails.
The winter air path through a central-station unit
- Discharge plenum and supply duct — the velocity here sets how far downstream the steam takes to disappear, and it is where the high-limit humidistat belongs Duct Air Velocity Calculator
- Steam dispersion tube — sized on the airflow past it and the moisture the space setpoint demands, never on the outdoor-air flow alone Steam Humidifier Moisture Capacity Calculator
- Supply fan — adds sensible heat of its own, all of it downstream of the low limit, so none of it ever helps the coil
- Heating coil bank — duty is the sensible rise from mixed air to supply air, the same relation a terminal reheat coil is sized on VAV Terminal Reheat Coil Sensible Heat Capacity Calculator
- Filter bank — loads unevenly in winter and drags the fan curve, which quietly changes both airflows feeding the mixing box
- Mixing box and dampers — two air streams, one leaving temperature, and the only place on the unit where a damper position becomes a number Mixed Air (Economizer) Temperature Calculator
The Damper Position Is a Number and So Is Its Consequence
Two air streams entering one plenum leave it at a flow-weighted average of their temperatures. That is all the mixing box does, and it is the reason a minimum outdoor-air fraction chosen for indoor air quality lands squarely in the middle of a freeze-protection problem. Take a unit at 8,000 CFM (3,776 L/s) with a 25% minimum outdoor-air position. On a morning at -12 °C (10 °F) against a 21 °C (70 °F) return, the plenum delivers about 12.8 °C, or 55 °F. That is comfortably clear of any sensible low limit, and it is also the number people quote to prove there is no problem.
The trouble starts when the outdoor-air fraction moves. It moves for defensible reasons — an economizer sequence releasing the damper on a mild afternoon and not resetting cleanly, a return fan out of track, a CO2-based demand-control loop overshooting, or somebody who left the damper in hand years ago. Every ten points of outdoor-air fraction on this example costs roughly 3.3 K of mixed-air temperature, and the last few points before a trip go by fast.
Weight by flow, not by damper blade angle. Damper travel is not linear in flow and never has been: an opposed-blade damper delivers most of its flow change over the middle of its stroke, and the relationship between commanded position and actual outdoor-air CFM depends on where the damper sits in the pressure system. If the unit has a measuring station on the outdoor-air duct, use its reading. If it does not, a traverse at the outdoor-air opening is a slower but honest substitute, and it is the only way to find out that a damper commanded to 25% is passing 40%.
One caveat worth carrying, because it works against you rather than for you: the volumetric weighting used here treats the two streams as equally dense, and on a cold morning they are not. Outdoor air at -12 °C is about 12% denser than return air at 21 °C, so a strict mass balance gives a mixed-air temperature roughly 0.8 K (about 1.4 °F) colder than the volumetric figure at that split. That is inside the error of your damper position and well outside the error of a low limit set 2 K away. Treat the calculated figure as the optimistic end of the range.
| Outdoor-air fraction | Mixed air (°C) | Mixed air (°F) | Margin over a 3.3 °C / 38 °F low limit |
|---|---|---|---|
| 10% | 17.7 | 63.9 | +14.4 K |
| 15% | 16.1 | 60.9 | +12.8 K |
| 20% | 14.4 | 57.9 | +11.1 K |
| 25% | 12.8 | 55.0 | +9.5 K |
| 30% | 11.1 | 52.0 | +7.8 K |
| 40% | 7.8 | 46.0 | +4.5 K |
| 50% | 4.5 | 40.1 | +1.2 K |
| 60% | 1.2 | 34.2 | below the limit — trips |
| 70% | -2.1 | 28.2 | below the limit — trips |
Put the two measured flows and the two measured temperatures in and read the plenum before you set the low limit — the answer is what the coil inherits, and every downstream number on the unit is built on it.
The volume of outdoor air being introduced through the economizer/damper.
The dry-bulb temperature of the incoming outdoor air.
The volume of return air being recirculated from the conditioned space.
The dry-bulb temperature of the return air stream.
Resulting mixed air temperature
62.74 °F
- Total combined airflow
- 2,000 CFM
They open the calculator with your figures already in it
Mixed Air (Economizer) Temperature Calculator: 62.74 °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
- Returns the fully-mixed average, and a mixing box rarely mixes. Cold outdoor air crossing the damper does not blend with the return in the short distance before the coil — it arrives as a cold layer across part of the face, ten degrees or more below the average reported here — so a freeze stat trips on the coldest foot of its element and a hydronic coil bursts at its coldest tubes, not at the mean. Blenders, opposed-blade dampers and the geometry of the plenum are what close that gap, and none of them are inputs.
Why the Plenum Reads Colder Than the Arithmetic
A flow-weighted average is a statement about a perfectly mixed stream, and a mixing box is not a mixer. Outdoor air usually enters through one face and return air through another, at right angles, into a plenum a metre deep, and the two streams then travel through the filter bank and onto the coil face largely as they arrived: a cold layer and a warm layer sharing a duct. Traverse the coil face on a genuinely cold morning and a spread of 10 to 15 K across the face is ordinary. The mean of that traverse may match the calculation to a fraction of a degree while one corner sits below freezing.
This matters because nothing downstream responds to the mean. The low-limit thermostat responds to the coldest short segment of its capillary, by design and by construction. The coil freezes wherever the water in one tube stops moving, not wherever the average says it should. A single averaging sensor mounted in the middle of the plenum is therefore the least useful instrument you can install there, and it is the one most units have.
The fixes are physical rather than programmed. Air blenders and static mixers exist for exactly this and are the correct retrofit where the plenum is too short. Parallel-blade dampers arranged so the outdoor and return blades throw their streams into each other mix far better than opposed-blade dampers set to throw them apart, and swapping the blade arrangement on the outdoor-air damper is a cheaper intervention than it sounds. Where the geometry cannot be changed, the answer is to stop asking the mixing box for a low mixed-air temperature at all, and to temper the outdoor air before it arrives.
- Fix the unit at its design supply airflow and the outdoor-air damper at its true minimum position, with the building at steady state rather than in warm-up.
- Divide the coil face into a grid — nine points on a small unit, twenty-five on a large one — and mark them on the access door so the next engineer repeats the same traverse.
- Read dry-bulb at every point with one instrument, not several, and give each reading time to settle; a fast handheld probe in a 12 K gradient reads whatever it touched last.
- Record the minimum, not the mean. The minimum is the number the low limit and the coil both respond to.
- Compare the mean of the traverse against the flow-weighted calculation. A large gap means your flows are wrong; a small gap with a wide spread means your mixing is wrong. They are different faults with different fixes.
- Repeat the traverse after any change to damper arrangement, blender installation or minimum-position setting, and keep both sets of readings.
A Low Limit That Guards the Coil Rather Than the Schedule
Low-limit thermostats respond to the coldest portion of their sensing element — commonly the coldest 300 mm or so of a capillary many metres long — which is precisely the behaviour you want from a freeze protection device and precisely the behaviour that makes it look erratic to anyone expecting an average. Manufacturers' installation literature gives the serpentine coverage rate, conventionally about a foot of element per square foot of coil face, and it gives the mounting face. Follow that literature rather than local habit: an element clipped in a neat loop across the top third of the coil is protecting the top third of the coil.
Setpoint is a specification decision, not a field preference. It comes from the coil manufacturer and the design engineer, it sits above freezing with margin rather than at it, and on a hydronic coil it is chosen with the glycol concentration in mind. Whatever the number, wire the device as a hardwired manual-reset interlock that stops the fan and drives the outdoor-air damper shut — not as a soft point in the controller that a graphic can override. Where the unit is also covered by smoke detection under NFPA 90A, Standard for the Installation of Air-Conditioning and Ventilating Systems, keep the two shutdown paths separate and test them separately, because a technician chasing one will otherwise disable the other.
Percent Is the Wrong Unit for This Job
Relative humidity is a ratio to a moving target, and in winter the target moves further than the number does. Outdoor air at -12 °C and 65% RH sounds damp. Heat that same air to 21 °C without adding a drop and it reads about 6% RH, because heating it to 21 °C multiplied what the air is able to hold by roughly eleven while the moisture in it did not move at all. Nothing was removed. The denominator grew. Any conversation about winter humidification that stays in percent will eventually produce a wrong answer, usually in the direction of undersizing.
Work in humidity ratio instead — grains of moisture per pound of dry air in imperial practice, grams per kilogram in metric — because that is the quantity conserved when you heat air and the quantity a humidifier actually adds. The psychrometric relations behind it are set out in the ASHRAE Handbook—Fundamentals, and the same relations underlie every psychrometric chart in every plant room. The one thing to remember from the chart is that heating moves you horizontally and humidifying moves you vertically, so the two operations never substitute for each other.
The table below is the reason winter humidification loads are almost entirely an outdoor-air problem. Even saturated, air at -12 °C carries about 9 grains per pound; the space you are trying to hold at 30% RH carries about 32. Every cubic foot of outdoor air brought in for ventilation arrives more than twenty grains short even in the wettest case the temperature allows, and that shortfall is the load. It also means that the design outdoor humidity ratio, not the design outdoor dry-bulb, is the input worth arguing about — take it from the mean coincident values in the ASHRAE Handbook—Fundamentals climatic design data rather than assuming saturation.
| Outdoor dry-bulb | Saturation ratio (g/kg) | Saturation ratio (gr/lb) | RH once heated to 21 °C |
|---|---|---|---|
| -20 °C / -4 °F | 0.63 | 4.4 | 4% |
| -15 °C / 5 °F | 1.01 | 7.1 | 7% |
| -12 °C / 10 °F | 1.34 | 9.3 | 9% |
| -10 °C / 14 °F | 1.60 | 11.2 | 10% |
| -5 °C / 23 °F | 2.47 | 17.3 | 16% |
| 0 °C / 32 °F | 3.77 | 26.4 | 25% |
| +5 °C / 41 °F | 5.39 | 37.8 | 35% |
The Steam the Setpoint Actually Costs
Take the same unit: 8,000 CFM total, 2,000 CFM of it outdoor air, outdoor design condition -12 °C carrying 5 grains per pound, space held at 21 °C and 30% RH, which is about 32 grains per pound. The outdoor-air stream arrives 27 grains short. Spread that shortfall across the whole supply airflow and the humidifier has to lift the mixed air by 27 × (2,000 ÷ 8,000), which is 6.75 grains per pound. That is the number the humidifier is sized on, and it is far smaller than the number people expect, because most of the air passing the dispersion tube is return air that never lost its moisture in the first place.
Both routes to the load must agree, and checking that they do is the quickest way to catch an error. Either the humidifier lifts the entire 8,000 CFM by 6.75 grains, or it makes good 27 grains on the 2,000 CFM of outdoor air; 6.75 × 8,000 and 27 × 2,000 are the same 54,000 grain-CFM. If your two routes disagree, one of the flows is wrong — usually the outdoor-air flow, which is the one nobody measures.
Turned into a mass rate, that comes to about 34.7 lb/hr, or 15.7 kg/h, of steam. The imperial form multiplies airflow in CFM by the humidity ratio rise in grains per pound, divides by the 7,000 grains in a pound, and applies a factor of 4.5 — which is nothing more mysterious than sixty minutes per hour times the 0.075 lb/ft³ density of standard air. The metric form does the same job in one line: 3.78 m³/s of air at 1.2 kg/m³ is 4.53 kg/s of dry air, times a 0.96 g/kg lift, times 3,600, giving 15.7 kg/h. That the two agree is a useful check that neither the density assumption nor the unit conversion has gone astray.
Both forms assume standard air, and that assumption is the first thing to fail. Above roughly 600 m of altitude the density factor and the humidity ratios both shift, and the two shifts do not cancel. Non-standard entering conditions, a unit that recirculates part of its own discharge, or an appreciable moisture gain inside the space will all move the answer. Use this figure to select the size of humidifier and to sanity-check a submittal; take the final selection from the manufacturer's own psychrometric sizing software, which is what DriSteem, Condair and Armstrong all publish for exactly this reason.
Then check what feeds it. A steam humidifier drawing from a house boiler adds that mass to the boiler's makeup water demand and to its water treatment load, hour after hour, for the whole heating season — 15.7 kg/h is about 110 kg of water a day at seven hours of run time. A self-contained electrode or resistive humidifier converts the same duty into electrical load — around 10 kW here on latent heat alone, at a little over 2,250 kJ per kilogram of steam, and more again once the feed water has to be brought up from cold. Neither figure belongs to the air handler, and both belong on somebody's schedule before the unit is energised.
Enter the airflow passing the dispersion tube — the supply airflow, not the outdoor-air flow — and the humidity ratio rise the space setpoint demands, and it returns the moisture rate the humidifier has to deliver.
The supply airflow rate passing through the humidifier section.
The target increase in the airstream's moisture content, in grains of moisture per pound of dry air.
Required humidifier moisture capacity
38.6 lb/hr
This uses a standard simplified psychrometric approximation (grains per pound of dry air, which metric practice writes as grams per kilogram, with the 4.5 factor for standard air) common in manufacturer humidifier sizing guides — for a final humidifier selection, use the manufacturer's full psychrometric sizing software, which accounts for actual entering air conditions and altitude.
They open the calculator with your figures already in it
Steam Humidifier Moisture Capacity Calculator: 38.57 lb/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 — 38.6 lb/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
- Nothing here is the absorption distance, and that is what most often goes wrong on site. Steam leaving a dispersion manifold needs a clear run of duct before it meets anything — a coil, a filter, a turning vane, a branch takeoff or the duct wall — and that distance depends on the leaving air's relative humidity, the velocity, the duct dimensions and the manifold type. Too short a run and the steam condenses on whatever it strikes: wet filters, corroding duct, and mold in a system installed to improve the air.
- Pounds per hour is not the service. Steam takes roughly 1,000 Btu to make each pound, so the 38.6 lb/hr (17.5 kg/h) on these defaults is about 38,000 Btu/h (11 kW) of gas input, or the same again in electricity on an electrode or resistive unit — a dedicated feed, and often a three-phase one. Add the water supply, whatever treatment the cylinder needs at your water hardness, and a drain rated for hot water.
- The humidity ratio increase is a figure you supply, and taking it off a mild day undersizes everything downstream. The design case is the coldest outdoor dew point the site sees, because that is when ventilation air arrives driest — and a humidifier that comes up short does not catch up the way a heating coil does, it simply holds the space under setpoint for the whole cold spell.
The Ceiling the Building Puts on the Setpoint
There is no comfort-code floor to argue from. ASHRAE Standard 55, Thermal Environmental Conditions for Human Occupancy, sets no lower humidity limit, so a winter humidity setpoint exists because a process, a collection, a manufacturing tolerance or a persistent complaint put it there. Ask which, because the answer decides how tightly it has to be held and whether it needs to be held at all outside occupied hours.
The upper bound is not a preference either — it is the coldest interior surface in the building. Every setpoint has a dew point, and the moment a window pane, an aluminium frame, a steel lintel or an uninsulated reveal drops below it, the humidification plant is manufacturing condensation and paying for the privilege. Single glazing on a -12 °C morning can hold an internal surface temperature in the low single digits, which puts even 30% RH at risk. Establish the coldest surface first, set the humidity below its temperature with margin, and reset the setpoint downward as the outdoor temperature falls, which is the sequence most control platforms already have and most sites never enable.
| Space RH | Humidity ratio (gr/lb) | Humidity ratio (g/kg) | Dew point |
|---|---|---|---|
| 25% | 27 | 3.8 | 0 °C / 32 °F |
| 30% | 32 | 4.6 | 3 °C / 37 °F |
| 35% | 38 | 5.4 | 5 °C / 41 °F |
| 40% | 43 | 6.2 | 7 °C / 44 °F |
| 50% | 54 | 7.7 | 10 °C / 50 °F |
Where the Tube Can Go
Steam leaves a dispersion tube as visible mist and needs a run of duct to disappear into the air. That run is the absorption distance, and it is a function of the leaving air condition, the duct velocity and the geometry of the tube itself — manufacturers publish it as a table or a selection routine, and it can be anywhere from a few hundred millimetres with a multi-tube dispersion panel to several metres with a single tube in cold, nearly saturated air. Nothing may sit inside that distance: no coil, no filter, no turning vane, no fire damper, no duct-mounted sensor and no lined duct, because everything in that zone gets wet.
The high-limit duct humidistat is the device most often mounted wrong for this reason. Put it beyond the absorption distance and it reads air; put it inside and it reads mist, drives the humidifier off, and produces a complaint about a humidifier that never runs. The same applies to the control humidity sensor, whether it sits in the duct or in the return.
Interlocks are what stop the tube from filling a stopped duct with steam. An airflow proving switch, a high-limit humidistat and the unit's own run status all belong in series with the humidifier's call, and all three want testing individually at commissioning rather than as a set. Water hygiene sits alongside them: a steam humidifier is a far lower risk than an evaporative or cold-water system precisely because it boils, but the feed, the drain and any adjacent wetted plant still fall under the building's water safety regime — ASHRAE Standard 188, Legionellosis: Risk Management for Building Water Systems, in North American practice, and CIBSE TM13 with the HSE's Approved Code of Practice L8 in the UK.
The Order Things Get Switched On
Sequence matters more than settings on a cold start, because half the failures above are failures of order rather than of value. This is the order that lets each stage prove itself before the next one can fault it.
- Prove the dampers mechanically before any control is enabled: full stroke both ways, blades closing tight against their seals, linkage tight, and the actuator's spring-return direction confirmed as closed on power loss.
- Verify the closed position leaks as little as its specification claims — outdoor-air dampers are rated to leakage classes tested under AMCA Standard 500-D, and a closed damper still passes measurable air, which is why an unoccupied unit can freeze.
- Start the fan with the outdoor-air damper shut and set the design supply airflow first, so every flow figure afterwards refers to the same fan operating point.
- Open the outdoor-air damper to its minimum position and measure the outdoor-air flow, by measuring station or by traverse. Do not infer it from a damper command.
- Traverse the coil face for temperature and record the minimum alongside the mean, then compare that mean against the flow-weighted mixed-air calculation.
- Set and test the low limit last on the air side, by hand, at the device, with the fan running — then confirm it dropped the fan and shut the outdoor-air damper.
- Bring the heating coil up and confirm the supply temperature the coil can hold at the measured mixed-air condition, not at the design one.
- Enable the humidifier only once the space is at temperature, and prove the airflow switch, the high limit and the run interlock separately by defeating each in turn.
What You Leave Behind in the Cabinet
The commissioning record is the deliverable, and on a winter start-up it has an unusually short list of numbers that actually matter: design supply airflow and the fan operating point that produced it, measured outdoor-air flow at minimum position, the mean and minimum of the coil-face traverse, the low-limit setpoint and the date it was function-tested, the space humidity setpoint expressed in humidity ratio as well as in percent, and the calculated moisture rate the humidifier was selected against. Six lines, in the unit, laminated. The next engineer to open that door on a cold morning inherits everything the arithmetic already established, and does not have to rebuild it at half five with a torch in their teeth.
The formal framework for that record, where a job has one, is ASHRAE Guideline 1.1, HVAC&R Technical Requirements for The Commissioning Process, with the measurement methods taken from ASHRAE Standard 111, Measurement, Testing, Adjusting, and Balancing of Building HVAC Systems, or from the AABC National Standards for Total System Balance or the NEBB Procedural Standards, depending on whose balancing report the project is buying. In European practice the equivalent equipment requirements sit in BS EN 13053 for air handling unit rating and performance and BS EN 1886 for casing leakage and thermal performance, with CIBSE Guide B covering the design intent behind them. Cite whichever governs the job on the sheet itself; a record that does not say what it was measured against is a page of numbers with no argument attached.
The six figures a winter start-up runs on
Everything above reduces to a handful of measured or specified quantities. Collect these before opening a control graphic, because each one is an input to the next.
- Design supply airflow, and the fan operating point that delivers it — Every other flow figure on the unit is a fraction of this one, so it is fixed first and quoted with the static pressure it was measured at.
- Minimum outdoor-air flow, measured rather than commanded — Set by the adopted ventilation code, but the number that matters is what the duct passes at the minimum damper position, not what the damper was told to do.
- Outdoor design dry-bulb and its coincident humidity ratio — Two separate values from the climatic design data. The dry-bulb drives the freeze problem; the humidity ratio drives the humidification load, and assuming saturation oversizes it.
- Return-air dry-bulb at the mixing box, not at the thermostat — Return air picks up or loses heat on its way back through ceiling plenums and risers, and the mixing box weighs whatever actually arrives.
- Space humidity setpoint, converted to grains per pound or grams per kilogram — Percent is unusable arithmetic here. Convert once, write both, and check the dew point against the coldest interior surface before committing to it.
- Low-limit setpoint, mounting face and capillary coverage — From the coil manufacturer's literature and the specification, with the function test dated — this is the interlock that decides whether the building starts.
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.