0.075 pounds mass per cubic foot. That is the air density behind almost every fan curve you will be handed, and it describes clean air at sea level and room temperature. If your process gas leaves a dryer at 400 degrees, or your plant sits at five thousand feet, or the airstream carries enough moisture and particulate to change its weight, the curve in front of you was drawn for somebody else's air.
Reading a fan curve well starts there. The lines describe what the fan did under known conditions. Your job is working out how far your conditions sit from those, because that distance is where selection mistakes live.
What the lines are, and where the duty point sits
Airflow runs along the horizontal axis, usually in CFM. Pressure runs up the vertical axis, usually inches of water gauge. A single curve represents one fan model at one constant speed, and most published sheets stack several curves for different speeds or wheel diameters. Brake horsepower and efficiency usually appear as their own lines on the same chart.
Fan curves come out of a laboratory. The aerodynamic performance test method is ANSI/AMCA Standard 210, published jointly with ASHRAE as Standard 51, and the current edition is 210-25. Standardizing the test is what makes two manufacturers' curves comparable at all.
The duty point is where the fan curve crosses your system resistance curve. That intersection, and not the fan curve on its own, tells you the airflow and pressure the installation will produce.
The system curve is the half nobody hands you
A manufacturer publishes the fan curve. The system curve comes from your ductwork, and it has to be calculated or measured on your side of the transaction.
System resistance rises with flow, roughly as the square of it, because the losses through duct, elbows, filters and coils scale that way. Plot that curve against the fan curve and the intersection moves as the system changes. A filter bank loading up over its service interval steepens the system curve and walks the duty point back along the fan curve toward less flow and more pressure. Nothing about the fan changed.
AMCA's own teaching material on fan and system curves is worth an hour if this is new. Hartzell's engineering guide covers the same ground in its Fans and Systems section, on page 13.
The fan laws, applied to one real change
The fan laws let you move a known operating point to a new one when speed or density changes. Page 12 of Hartzell's engineering guide states them plainly. When fan speed varies, air delivery varies directly as the RPM ratio, developed pressure varies as the RPM ratio squared, and horsepower absorbed varies as the RPM ratio cubed.
Run that on a real decision. Slow a fan to 90 percent of its original speed and you get 90 percent of the airflow, about 81 percent of the pressure, and roughly 73 percent of the shaft power. A 10 percent speed reduction takes about 27 percent off the power. That relationship is the whole argument for variable speed on a fan that spends most of its life below design flow.
Density works differently and catches people more often. A fan running at constant speed is essentially a constant volume machine, so the volume it moves does not change with density. What changes is everything you pay for and everything the duct sees: absorbed power and the static, velocity, and total pressures all vary directly with the density ratio, which falls as absolute temperature rises. A fan selected on a standard air curve and installed on a hot process gas will move its rated volume and develop less pressure than the sheet says.
The guide's density, temperature and altitude corrections are on page 14, and safe operating speeds are on page 15. The full document is the Hartzell Engineering Guide (A-108-Q-1).
Three standards, three different questions
A fan curve carries more authority once you know which standard produced it and which standards check it. ANSI/AMCA 210 is the laboratory test that generates the published curve, run under uniform, idealized duct conditions a real job site rarely provides. AMCA Publication 201 is the design-stage document that explains why installed performance drifts from that curve, giving system designers the System Effect Factors that account for how the fan and duct system interact when an inlet elbow sits too close or a straight run is too short. AMCA Publication 203 is the field measurement method, used to verify a fan's actual aerodynamic performance once installed so a shortfall can be confirmed as real and quantified. For industrial process and power generation fans, AMCA Standard 803 is the site performance test written for that industry.
Where the chamber and the plant diverge
It helps to picture the room the curve came from. Hartzell's laboratory in Piqua, Ohio runs a twelve-foot chamber that moves 50,000 CFM at 24 inches water gauge, a smaller chamber rated to 11,000 CFM at 80 inches, and an overspeed cell for fans up to 200 horsepower. A fan on that chamber is ducted straight into instrumentation with clean, uniform approach conditions and nothing upstream disturbing the inlet.
Your installation has an elbow four diameters from the inlet, a damper that someone set during startup in 2019, and a duct run that was value-engineered after the fan was ordered. Those differences have a name in this industry, system effects, and AMCA 201 exists to account for them. They are the reason a fan that meets its published curve can still underperform in the field. We wrote up six such cases presented at AMCA International, and in all six the fan met specification.
Hartzell runs an AMCA-accredited laboratory and will still tell you that the curve it produces describes the fan rather than your system. Those are two different questions and the sheet only answers one of them.
There is a practical ceiling on the chamber side too. The largest metal centrifugal fan Hartzell has mounted on its own chamber and run at full speed is still smaller than what one target application needs, and larger units are on order for delivery in November and December of 2026. Test capacity is a real constraint on any manufacturer, and it is worth asking how a large selection was rated.
The left end of the curve, where selections go wrong quietly
Every fan curve has a region on the left side, at low flow and high pressure, where the flow through the wheel stops behaving. Air separates from the blades, the curve flattens or dips, and the fan enters stall. A fan operating there moves less air than the sheet suggests, runs rough, and puts cyclic loading into the wheel and bearings that shows up later as a vibration complaint nobody connects back to the selection.
The awkward part is that stall rarely announces itself at startup. A system oversized on paper, a damper closed further than design intended, or a duct run that got shortened during construction can all push the operating point left over time. When a fan that ran fine for two years starts sounding wrong, the duty point is worth replotting before anyone opens the bearing housings.
Performance envelopes and what a fan class tells you
Individual curves get compared one at a time. Performance envelopes let you compare a whole product family at once, and Hartzell publishes them on pages 8 and 9 of the engineering guide. An envelope shows the flow and pressure territory a series can cover, which is the faster way to work out whether you are looking at the right family before you go hunting for the right size within it.
Fan class sits alongside the curve and answers a different question. Class describes the pressure and speed envelope the fan is built to handle structurally, and on the fiberglass centrifugal line the published ceilings run to 5 inches static pressure for Class I, 8.5 inches for Class II, and 20 inches for Class III. A duty point that falls inside a curve but outside the class rating is a construction problem rather than an aerodynamic one, and the curve alone will let it slip past you.
Two things printed on the curve that decide whether it applies
The density basis, and the speed. Both appear on a properly published sheet and both get skipped.
If the sheet says the ratings are based on standard air and your airstream is hot, wet, dusty or at altitude, the curve needs correcting before it means anything for your duty point. If the sheet stacks curves by speed, confirm which one the selection was made on and whether that speed is achievable with the drive being supplied. Our walkthrough on centrifugal fan selection covers how blade design changes the shape of these curves. If you are still working out what airflow you need before you get to the curve, start with CFM.
When the curve is right and the field is wrong
A fan can meet its published curve and still miss in the field. The causes fall into a short list worth knowing before you open a bearing housing looking for the problem.
Inlet and outlet conditions. An elbow too close to the inlet, a damper at the outlet, or a fan ducted into a plenum with no straight run feeds the wheel non-uniform air, and the wheel does not see the airstream it saw on the test stand. The fix is straight duct length, turning vanes, a reoriented elbow, or an evasé on the outlet.
A mis-estimated system curve. The fan curve was accurate and the system resistance calculation was not, from underestimated duct friction, dirty filters left out of the math, dampers not fully open, or fittings added late. The operating point slides to less flow than intended. The fix is to re-commission and verify actual system resistance by measurement.
Fan speed below assumption. Belt slip, wrong sheave sizing, or a motor turning at a different RPM than design. Flow scales with RPM and pressure with RPM squared, so a small shortfall compounds fast. Verify actual RPM in the field and correct the sheave ratio or belt tension.
Density correction skipped. Curves are published at standard air. Hot, humid, or high-altitude field air moves the same volumetric CFM but less mass flow and pressure than a naive read of the curve suggests. Correct the selection using actual air density.
Recirculation. Discharge air finds its way back to the inlet, common on roof-mounted or closely spaced installs. The fan moves its rated CFM, but some of it is used air being re-ingested. Increase separation, add a stack or hood, or reorient inlet and outlet.
Fans interacting in parallel or series. Without accounting for the combined curve near stall, one fan can push another into an unstable region, or flow imbalance means neither reaches its rated point. Verify the combined system curve crosses both fans on the stable part of their curves, or manage staging with interlocks and dampers.
Before you accept a selection
Ask for three things and the conversation gets a lot more useful. Ask for the density and speed the curve was drawn at. Ask for the system resistance calculation the selection was based on, not just the resulting duty point. Ask where that duty point sits relative to the peak of the efficiency curve, because a point far to the right of peak buys you power consumption you will pay for every hour the fan runs.
If you have a curve in front of you and a system that does not look like a test chamber, send us the arrangement and the airstream conditions. Our application engineers will work the duty point with you: 1-800-336-3267 or info@hartzell.com.
Frequently Asked Questions
Why don't two manufacturers' fan curves for a similar fan match my expected performance?
A fan curve describes the fan under standardized lab conditions, not your installation. The duty point that matters is where the fan curve crosses your system's resistance curve, and that system curve depends on your ductwork, not the fan.
Does a fan curve account for hot or high-altitude air?
No. Most published curves are based on standard air, 0.075 lb/ft³. If your airstream is hot, humid, dusty, or at altitude, the actual density differs from that basis, and pressure and power shift accordingly even though the fan still moves the same volume at constant speed.
If I slow down my fan, how much does that save on power?
Power drops with the cube of the speed ratio. Cutting speed to 90 percent of the original cuts power to roughly 73 percent, about a 27 percent reduction, while airflow only drops to 90 percent. This is the core case for variable speed on a fan that runs below design flow most of the time.
What does "fan class" mean, and is it different from the curve?
Class describes the pressure and speed envelope the fan is built to handle structurally, separate from the aerodynamic curve. A duty point can sit inside the curve but outside the fan's class rating, which is a construction limit rather than a performance one.
What should I ask for before accepting a fan selection?
Three things: the density and speed basis the curve was drawn at, the system resistance calculation behind the duty point (not just the resulting number), and where that duty point sits relative to peak efficiency.