Dedicated Outdoor Air (DOAS) Air Systems
A conventional air handler does two jobs in one breath: it brings in the code-required slice of outdoor air for ventilation, and it heats or cools the space — all in a single blended stream, driven by one thermostat call. A dedicated outdoor air system pulls those two jobs apart. It takes 100% outdoor air, conditions it to a controlled supply state, and ducts it to the zones purely as ventilation — while separate terminal units at each zone handle the heating and cooling the space actually needs. This page is about what that split buys you and how the unit is controlled. One question, start to finish: what is a DOAS, and why would a building decouple ventilation from the rest of the load?
This page is the counterpoint to Air Handlers, which walks the conventional mixed-air unit station by station. If the mixing box — where a slug of outdoor air joins the recirculated return and the blend gets conditioned together — isn't familiar yet, read that first. Everything below is about the other way to bring in fresh air: not blended into the return, but on its own unit, with its own coil, its own controls, and its own reason for existing.
One Unit, One Job: Ventilation Air
Start with the definition and hold onto it: a DOAS is a unit dedicated to 100% outdoor air. It draws in nothing but fresh air, conditions that air to a controlled leaving state, and ducts it to the zones as the building's ventilation supply. It never touches return air the way a mixing box does — there is no blend. The heating and cooling the space needs to hold its thermostat setpoint is somebody else's job entirely: the terminal units. Those can be fan-coil units, VAV boxes, chilled beams, or radiant panels — whatever the design picked to move the space's sensible load. The DOAS delivers ventilation; the terminals hold the room.
Contrast that with the conventional mixed-air air handler. There, the outdoor-air damper cracks open to a minimum position, a small fraction of fresh air mixes into the big recirculated return stream, and the blended air crosses one set of coils on its way back to the space. Ventilation is real, but it is entangled with the thermostat: the same fan and the same coils serve both the cooling call and the fresh-air requirement, and the outdoor air rides in as a passenger on whatever the space happens to be calling for. A DOAS un-entangles them. The two streams below are the whole idea — one blended path versus two dedicated ones.
Why Pull Them Apart
Decoupling is not tidiness for its own sake — it solves two real problems that a mixed-air unit is bad at. The first is moisture. A large share of the building's latent load — the moisture the system has to remove — walks in the door with the ventilation air, especially in humid climates. (Some spaces make plenty of their own — a lab, a kitchen, a natatorium — but the outdoor-air stream is the part a mixed-air unit is worst at drying.) Hot, humid outdoor air carries far more water per pound than the conditioned air already in the space, so the fresh-air requirement and the humidity problem are, physically, the same stream. Put that stream on its own unit and you can control humidity independently of temperature: the DOAS is sized and controlled to wring the moisture out of the ventilation air before it ever reaches a zone, instead of hoping the space cooling coil happens to run cold enough, long enough, to dry the air out as a side effect. That side-effect approach is exactly how a mixed-air system ends up with cold, clammy spaces on a mild, humid day — the thermostat is satisfied, so the coil stops, so the drying stops, while the outdoor dew point keeps climbing.
The second problem is guaranteed ventilation. On a mixed-air VAV system the outdoor air rides in as a fraction of the supply, and the supply throttles down as zones satisfy — so a box sitting at its minimum flow can quietly starve its occupants of fresh air even while the temperature is perfect, and the whole system's outdoor-air fraction wanders as the fan speed and damper positions move. A DOAS sidesteps that entirely: it delivers each zone its ASHRAE 62.1 ventilation share as a measured, dedicated flow, regardless of whether the zone is calling for heating, cooling, or nothing at all. Ventilation stops being a fraction you hope is right and becomes a delivered quantity you can meter.
There is a payoff on the equipment, too. Because the terminal units no longer have to bring in and dehumidify outdoor air — the DOAS already did — they only have to move the space's sensible load. That makes them smaller: a fan coil that would have been sized for the ventilation-plus-latent burden on top of the room load can shrink to just the room's sensible swing. The load didn't vanish; it moved to the unit built to handle it.
The Latent / Sensible Split
Here is the load-bearing idea of the whole architecture, stated plainly: the DOAS owns the latent load; the terminals own the sensible load. Latent is the moisture — the energy tied up in water vapor. Sensible is the dry-bulb temperature — the part a thermometer reads. A conventional coil does both at once and can't really separate them; a DOAS design deliberately assigns each to the equipment suited for it. And doing the latent half well drives an unusual-looking move at the DOAS coil.
To pull moisture out of air, a cooling coil has to run its surface colder than the air's dew point, so water condenses onto the fins and drains away (the same physics behind the drain pan on any cooling coil, and the reason a dew point matters). To dry the ventilation air enough, the DOAS drives its coil deep — down to a low apparatus dew point, the effective cold-surface temperature the leaving air is pulled toward. A common target is a leaving-air dew point around 55 °F, and where a space needs tighter humidity the coil is pushed further — toward a leaving dew point in the mid-to-upper 40s °F. That is far colder than any space wants its supply air.
So the deep coil creates a problem it then has to fix: air that dry is also that cold, and blowing 50 °F air into an occupied zone would overcool it and hand the space's heating a fight it never asked for. The usual answer is reheat: after the deep coil has wrung the moisture out, a reheat coil warms the air back up to a neutral supply — around 70 °F dry-bulb, close to room temperature — so the DOAS delivers dry air that is thermally almost invisible to the space. It added no net sensible load; it just handed the room dehumidified ventilation and left the heating and cooling to the terminals, exactly as designed. (Not every DOAS reheats: a cold-air DOAS deliberately skips it and delivers the cool, dry air straight to the zone, using that coldness to knock down part of the space sensible load on purpose. Neutral is the common default; cold-air is a design choice you will meet.)
The diagram below is that process on a stripped-down psychrometric path: the outdoor-air state, the deep coil pulling it down and to the left to the apparatus dew point, and the reheat carrying it straight across — dry-bulb up, moisture unchanged — to a neutral supply. (The full air-state vocabulary — dry bulb, dew point, humidity ratio, enthalpy — lives in Psychrometrics Basics, and you can plot a real state on the psychrometric chart.)
A schematic path, not a metered chart — the point is the shape: down-and-left to dry the air, then straight across to warm it without re-wetting it. Real state points come off the psychrometric chart.
How a DOAS Is Controlled
Because the DOAS exists to control moisture, its primary control loop is not a dry-bulb loop — it is a leaving-air dew point (or humidity) setpoint on the deep coil. The unit modulates its cooling — a chilled-water valve, or compressor staging on a DX unit — to hold the supply air at a target dew point, measured by a humidity or dew-point sensor in the leaving airstream. That is the number to look for first on the BMS graphic: not "is the supply 55 degrees," but "is the supply dry enough." A separate reheat loop then trims the dry-bulb up to the neutral supply setpoint, so the two loops divide the work the same way the whole system does — one owns moisture, the other owns temperature.
Running a coil deep and then reheating is energy-hungry by nature — you are paying to cool air well past what temperature alone needs, then paying again to warm it back. That is why most modern DOAS units carry energy recovery ahead of the coil, letting the outgoing exhaust air precondition the incoming outdoor air. What actually crosses between the two streams depends on the device: an enthalpy (total-energy) device — a total-energy wheel or a membrane/enthalpy plate core — transfers both heat and moisture, while a plain sensible exchanger — a bare fixed-plate core or a sensible-only wheel — moves heat alone. For a DOAS fighting a latent load, the enthalpy kind is the one that earns its keep: in summer the cool, dry exhaust pre-cools and pre-dries the hot, humid intake before it ever reaches the coil; in winter it pre-warms and re-humidifies it. A total-energy wheel commonly recovers on the order of 70 to 80% of the energy in the air it would otherwise throw away, which is a large part of what makes a deep-coil-plus-reheat unit affordable to run — and energy codes now require recovery on many high-outdoor-air units for exactly that reason.
The rest of the control story is scheduling. Ventilation is only owed when the building is occupied, so the DOAS runs on an occupancy schedule, and on systems with variable occupancy it can lean on demand-controlled ventilation — trimming outdoor-air flow (or staging the unit) as measured or inferred occupancy falls, then restoring full ventilation as it climbs. Put together, the BMS points that tell the whole story are a short list: leaving-air temperature and leaving-air RH / dew point (the humidity control point), the cooling coil valve or compressor stage, the reheat valve or stage, and the energy-recovery wheel command and status. Read those five in order and you are watching the unit do its one job. The capstone below marks each of them on the airflow path.
When the Climate Outruns the Design
All of that is the theory working as intended. Here is what it looks like when the ground shifts under a DOAS after it is installed — a real building, told from the DOAS side. (The same building's fan-coils are the closing story on Coil Selection, seen from the terminal side; this is the unit that feeds them.)
True story. I service a building with lab spaces that have to hold humidity in a tight band. It is built exactly the way this page describes: a DOAS conditions the ventilation air and hands it to the labs, while fan-coil units — chilled water and hot water, served from the plant's chiller and boilers — handle each room's sensible load. The DOAS makes its own cold with onboard DX, and it was sized and set to a design leaving dew point chosen for the climate the drawings assumed. The whole split depends on that one number holding: the DOAS delivers air dry enough, and the FCUs — picked for mostly sensible duty — never have to become dehumidifiers. Then the outdoor climate drifted off the design. The peak summer days now land muggier and more punishing than the drawings ever budgeted for, and the outdoor air shows up at dew points past what the deep coil was built to pull down. The DOAS runs flat out and still can't reach its leaving-dew-point setpoint — the coil simply lacks the depth for the moisture load it now has to strip. So air wetter than the design ever intended reaches the labs, and the FCUs, sized for sensible and fed chilled water that was never meant to do heavy drying, can't absorb the extra latent. The lab humidity climbs. The controls tell the whole story at a glance: the DOAS leaving dew point sitting stubbornly above setpoint, space RH above its band, and the terminals pinned wide open with comfort still off. Nothing is broken and nothing is out of tune — the unit is meeting a design condition the weather stopped honoring. That is the hard lesson a DOAS teaches when its assumptions expire: the entire architecture rides on the deep coil winning its fight with the outdoor dew point, and a design dew point chosen for yesterday's climate is not a promise the coil can keep once the climate outgrows it.