How Air and Water Behave Differently in the Same Run

Why this matters

Techs cross between air work and water work constantly, and the instincts do not transfer cleanly. A duct velocity that would be silent is an erosion problem in copper. A leak that is a nuisance on one side floods a ceiling on the other. A blocked high point stops a water loop dead and does nothing at all to a duct. Almost every one of those differences traces to two properties: water carries roughly three and a half thousand times more heat per cubic foot per degree than air, and water does not compress. Hold those two and the rest stops being trivia you have to memorize.

Before you break into either one

• Water side: relieve pressure and confirm zero on a gauge before any joint comes apart, and shut the heat source and let the fluid drop below 120 F, or discharge through a hose into a drain while you stand clear of the opening with a face shield and heat-rated gloves. Water held above 212 F under pressure flashes to steam the moment the joint opens, so read the temperature first.
• Air side: lock the blower disconnect open and confirm the wheel has stopped before any part of you passes the cabinet opening (29 CFR 1910.147). A coasting wheel looks stopped from three feet away.
• Disturbing duct interiors, liner or accumulated debris is an inhalation route. Fibers and biological growth go airborne when you brush, cut or probe them. Use a fitted particulate respirator under a program meeting 29 CFR 1910.134; gloves do nothing for this one.
• Never use a shop compressed-air line to blow out a water line, clear a drain or chase debris out of a duct. That is a different fluid at a different energy: compressed air can inject through skin and drive debris at eye level. OSHA restricts compressed air used for cleaning to under 30 psi with effective chip guarding and personal protective equipment, at 29 CFR 1910.242(b). Low-pressure duct air is nearly harmless; a shop air line at line pressure is not, and the two get confused because both are "air."

The one number

Heat carried per cubic foot per degree is density times specific heat.

• Water: about 62.4 pounds per cubic foot times 1.0 Btu per pound per degree F, which is about 62.4 Btu per cubic foot per degree F.
• Standard air: about 0.075 pounds per cubic foot times 0.24 Btu per pound per degree F, which is about 0.018 Btu per cubic foot per degree F.

The ratio is about 3,470 to 1. Note the conditions attached to the air figure: standard air is roughly sea level, near 70 F, dry. Hot supply air is less dense and carries less per cubic foot, altitude air less still, and both effects are large enough to matter on real equipment selections.

Note also that per pound the ratio is only about 4 to 1, not 3,470 to 1. Air is not a poor heat carrier by weight. It is a poor heat carrier by volume, and volume is what your ductwork and your fan have to move.

The two field constants both fall straight out of that:

• Air, sensible heat only: Btu/h = 1.08 x CFM x temperature difference, where 1.08 is 60 minutes per hour times 0.075 times 0.24, at standard air conditions.
• Water: Btu/h = 500 x GPM x temperature difference, where 500 is 60 minutes per hour times about 8.33 pounds per gallon times 1.0, for water at ordinary service temperature. Substitute the mix's actual density and specific heat for glycol.

The same duty, both ways

Move 60,000 Btu/h with a 20 F temperature difference on the fluid.

Water: 60,000 divided by (500 x 20) = 6.0 GPM. Six gallons per minute is 6 x 0.1337 = 0.80 cubic feet per minute.

Air: 60,000 divided by (1.08 x 20) = 2,778 CFM.

So the same job needs 2,778 cubic feet of air per minute against 0.80 cubic feet of water per minute, a volume ratio of about 3,460 to 1, which is the property ratio arriving exactly where you would expect it.

Now put each in a pipe or a duct.

• Water in 3/4 in. type L copper, inside diameter about 0.785 in., cross section about 0.484 square inches, which is 0.00336 square feet. Velocity is 0.80 divided by 0.00336 = 238 feet per minute, which is 4.0 feet per second.
• Air at a conventional trunk velocity of 900 fpm needs 2,778 divided by 900 = 3.09 square feet, which is about 445 square inches, roughly a 20 in. by 22 in. duct.

Read those side by side. The duct cross section is about 920 times the pipe cross section, and the air is still moving about 3.8 times faster. Multiply those two together and you land back within rounding at the 3,460 volume ratio, which is a useful check that the whole example closes.

This is the number behind every "why is the duct so big" conversation, and behind the fact that a hydronic retrofit into a building with no room for ductwork is often the only option that physically fits.

Why the two use different pressure units

Duct pressures are quoted in inches of water column and hydronic pressures in psi or feet of head, and it is not a convention gap. One inch of water column is 0.036 psi. A whole residential duct system operating at 0.5 in. w.c. total external static is running on about 0.018 psi. A hydronic loop at 34 ft of pump head is running on 14.7 psi.

That is a factor of roughly 800 in the pressure the two systems work against. It is why a duct joint that leaks is a capacity problem and a pipe joint that leaks is a building problem, why duct is sheet metal and pipe is not, and why an instrument that reads hydronic pressure usefully cannot resolve duct pressure at all.

What incompressibility costs you

Water does not compress meaningfully. A moving column of it that gets stopped abruptly has nowhere to put its momentum, so the pressure spikes. The magnitude scales with the velocity change you impose, which is the practical lever: halve the design velocity and you roughly halve the spike a given valve closure produces. That is why quick-closing solenoid valves on high-velocity lines are the classic hammer source, and why the fix is usually arrestors plus a look at whether the line is undersized, not a slower valve alone.

Air compresses, so a duct system self-damps. Slam a damper and you get noise and a transient, not a pressure wave that splits sheet metal. The exception worth naming is compressed air: a shop or control air line at line pressure stores real energy and behaves much more like the water case, which is exactly why the cleaning restriction cited above exists.

The same property drives the two systems' blockage behaviour, and it is completely asymmetric:

• Air trapped in a water line collects at high points and stops flow. The pump can be running perfectly and deliver nothing.
• Water in a duct is condensate. It runs to the low point, and if the drainage is right, it leaves. It does not stop airflow.

Which side controls a heat exchanger

On an air-to-water coil, the film resistance on the air side is far higher than on the water side. That is why coils have fins on the air side and bare tube on the water side, and it has a direct service consequence: the same thickness of fouling costs you far more on the air side of an air-to-water coil than on the water side, because the air side already dominates the total resistance.

Hedge that properly, because the family matters. It holds for air-to-water and air-to-refrigerant coils. It does not hold for water-to-water or water-to-refrigerant exchangers, where neither side is overwhelmingly dominant and a fouled water side is a first-class problem. Check which two fluids you have before you decide which side to look at first.

The velocity ceilings are set by different things

Both fluids have a practical velocity limit and the limits come from unrelated physics, which is why the numbers look backwards.

• Air is limited by noise, mostly at the outlets and at fittings. Conventional residential practice keeps trunk velocities in roughly the 700 to 900 fpm band with runouts lower, and the friction penalty is a secondary consideration at those speeds. Confirm the target against the design guidance you work to, because commercial and duct-lined applications carry different numbers.
• Water is limited by erosion-corrosion, which accelerates with both velocity and temperature. Common practice caps copper at roughly 8 feet per second in cold water lines and lower, around 4 to 5 feet per second, in hot recirculating lines. Confirm against the tubing manufacturer's limit for the alloy and service, because it varies by material.

Convert the water limit to the same units as the air one and the contrast is stark: 4 feet per second is 240 fpm. Air runs about four times faster than water and is the quiet one, because the energy carried in the stream is set by density times velocity squared, and the density term is 830 to 1 in water's favour.

Where the two behave identically

A fan and a pump are the same machine. Both are rotodynamic, both produce a curve of falling head against rising flow, both obey the same affinity relationships (flow with speed, head with the square of speed, power with the cube of speed), and both find their operating point where their curve crosses the system's. The system curve is k x flow squared in both cases for a closed circuit.

This is worth stating explicitly because it is the part that does transfer. A tech who learned to read a pump curve and locate an operating point already knows how to read a fan curve. The reasoning is portable even though almost none of the numbers are.

References

• ASHRAE Handbook, Fundamentals, for fluid properties, sensible heat relationships and duct and pipe design velocities
• Tubing manufacturer data for maximum design velocity by alloy, temperature and service
• OSHA 29 CFR 1910.242(b), compressed air used for cleaning; 29 CFR 1910.147 for blower isolation; 29 CFR 1910.134 where duct interiors or liner are disturbed
• See related: Static, Velocity and Total Pressure; What a Closed Loop Does That an Open One Does Not; The Airlock That Looked Like a Dead Pump