You reduce a fan from 3,000 RPM to 2,400 RPM and expect airflow to fall by 20%. That part is usually a reasonable first estimate. The surprise comes when pressure falls much faster and the measured electrical power refuses to match a perfect cube-law prediction.
Fan affinity laws are useful because they turn one tested operating condition into a quick estimate at another speed. They are not a substitute for the fan curve, the system curve or a prototype test. Used with those tools, however, they help you screen a speed change before you commit to a new fan, controller or power supply.
What decision can the fan affinity laws help you make?
The laws answer a narrow but valuable question: if the same fan operates under dynamically similar conditions, how should airflow, pressure and absorbed power change when speed changes? They can also be used to scale geometrically similar fan sizes, although that comparison carries more assumptions.
For an enclosure, UPS, inverter or telecom cabinet, this lets you estimate whether a lower speed could still meet the cooling duty, how much pressure margin you may lose and whether a higher speed could overload the motor or power electronics. It also shows why a small RPM change can create a much larger power change.
The laws do not tell you the final installed airflow by themselves. That result still comes from the intersection of the fan curve and the system resistance curve. If you are not comfortable with that intersection, review how to read a fan curve before applying the ratios below.
The three fan laws for a speed change
For the same fan diameter, similar flow pattern and essentially constant air density, define N1 as the original speed and N2 as the new speed.
Airflow changes in direct proportion to speed
Q2 = Q1 x (N2 / N1)
Q is volume airflow. If speed falls to 80% of the original value, the corresponding point on the scaled fan curve has about 80% of the original airflow.
Pressure changes with the square of speed
P2 = P1 x (N2 / N1)^2
P may represent fan static pressure or fan total pressure, but you must use the same pressure definition at both conditions. At 80% speed, the corresponding pressure is about 64% of the original value.
Absorbed fan power changes with the cube of speed
Power2 = Power1 x (N2 / N1)^3
At 80% speed, the ideal absorbed power becomes about 51.2% of the original value. This cubic relationship explains why variable-speed operation can reduce fan power quickly. It does not guarantee that measured electrical input power will fall by exactly the same percentage, because motor, drive and control-electronics losses do not disappear in perfect proportion.
A worked example: reducing a cooling fan from 3,000 to 2,400 RPM
Assume a fan produces 600 CFM at 150 Pa and absorbs 120 W at a defined point on its curve. You want a first estimate at 2,400 RPM.
Speed ratio = 2,400 / 3,000 = 0.80
| Quantity | Calculation | Estimated result |
|---|---|---|
| Airflow | 600 x 0.80 | 480 CFM |
| Pressure | 150 x 0.80^2 | 96 Pa |
| Absorbed power | 120 x 0.80^3 | 61.4 W |
These three results describe a corresponding point on a scaled fan curve. They do not say that your cabinet will automatically operate at exactly 480 CFM and 96 Pa. The system will choose a new operating point where the lower-speed fan curve intersects the same system curve.
If the system contains a clean filter, the prediction may look reasonable. If the filter is loaded, a damper is partly closed or the inlet is crowded, the real intersection may sit farther left. The component temperatures, not the ratio calculation alone, decide whether the speed reduction is acceptable.
Note: Never apply the airflow, pressure and power ratios to three unrelated catalog endpoints. Start from one real operating point, then scale that point or the complete curve.
Why the system operating point moves
A fixed duct or enclosure often has resistance that rises approximately with airflow squared over its normal turbulent operating range. The fan pressure capability also scales with speed squared. In an ideal fixed system, that similarity is why airflow may track speed closely.
Real cooling systems introduce complications. A filter can have different clean and loaded characteristics. A backdraft shutter may not open proportionally at low flow. A heat exchanger can redistribute air unevenly. Control dampers, leakage paths and parallel fans can change state as speed changes. Once the system geometry or resistance behavior changes, the new operating point no longer follows a simple scaled path.
Use the affinity laws to create a candidate curve, then intersect that curve with the actual clean-filter and dirty-filter system curves. The article on fan speed versus airflow explains why RPM alone cannot define installed airflow.
How air density changes the prediction
Volume airflow is primarily set by geometry and speed, so a fan can move a similar volume of thinner air at altitude. The mass of air in that volume is lower, however, and that changes pressure capability, absorbed power and heat removal.
For the same fan, speed and similar flow condition:
Pressure2 = Pressure1 x (rho2 / rho1)
Power2 = Power1 x (rho2 / rho1)
Here, rho is air density. If density falls by 15%, the corresponding pressure and absorbed fan power also fall by about 15% under the similarity assumptions. More importantly for electronics cooling, the mass flow and heat capacity carried by the air fall. A catalog CFM value therefore cannot prove equal cooling at sea level and at high altitude.
Temperature, altitude and humidity all affect density. Use the density stated for the fan curve or test standard, correct only within a defensible range and validate the equipment at the worst expected ambient and altitude. Do not use a density correction to hide insufficient thermal margin.
What changes when fan diameter changes?
The full similarity relationships include fan diameter D as well as speed N. For geometrically similar fans operating at similar points:
Q2 / Q1 = (N2 / N1) x (D2 / D1)^3
P2 / P1 = (rho2 / rho1) x (N2 / N1)^2 x (D2 / D1)^2
Power2 / Power1 = (rho2 / rho1) x (N2 / N1)^3 x (D2 / D1)^5
The diameter exponents make scaling look powerful, but compact cooling fans with different frame sizes are rarely perfect geometric copies. Hub ratio, blade count, tip clearance, motor blockage, frame depth and guard geometry may all change. Use diameter scaling to understand a trend, not to manufacture a performance curve for an untested model.
Where fan affinity laws become unreliable
| Condition | Why the estimate can drift | What to use instead |
|---|---|---|
| Near stall or an unstable curve region | Flow separation changes the aerodynamic pattern | Measured curves and a stable operating-range limit |
| Very low speed | Motor, bearing and electronic losses become a larger share of input power | Measured electrical power and airflow at the commanded speed |
| Different fan models or blade geometry | Dynamic similarity is not established | Individual model curves |
| Changing filter, damper or leakage state | The system curve is no longer fixed | System measurements in each configuration |
| Large density or temperature change | Cooling mass flow and component limits may change | Density correction plus thermal validation |
| Speed above the rated limit | Motor current, blade stress, bearing load and noise may exceed limits | Manufacturer-approved operating range |
The U.S. Department of Energy fan-system sourcebook uses fan laws as part of system assessment, while also warning that higher speed can require substantially more power and a review of motors and rotating components. That is the right level of caution for an equipment design: calculate first, then confirm the electrical, mechanical and thermal limits.
Applying the laws to AC, DC and EC cooling fans
EC and controlled DC fans
For a controllable EC or brushless DC fan, use measured RPM rather than assuming that control percentage equals speed percentage. A 60% PWM command or a 6 V analog command does not universally mean 60% RPM. Minimum-speed limits, dead bands, internal control maps and load limits are model-specific.
Once you know the actual speeds, the laws provide a useful first projection. Then compare the projected point with the correct speed curve and measure electrical input power. This keeps the aerodynamic estimate separate from motor and controller efficiency.
AC fans
Do not assume supply frequency and fan RPM remain in perfect proportion for every AC fan. Motor slip, control method and load matter. Voltage reduction can also change motor torque and temperature without producing a clean affinity-law speed change. Use an approved speed-control method and measure RPM before applying the ratios.
Multiple fans
Affinity laws scale the performance of each fan; they do not replace the rules for combining fans. Parallel fans share pressure and add airflow at a common pressure, while series stages add pressure at a common airflow. If one fan stops or the fans interact through a small plenum, the array may not behave like a simple multiplied curve. See fans in parallel versus series for the array calculation.
A practical validation sequence
- Choose a tested baseline point for the exact fan model, voltage, frequency and air density.
- Use measured RPM to calculate the speed ratio.
- Scale airflow, pressure and absorbed fan power with the three laws.
- Create or obtain the scaled fan curve rather than relying on one calculated point.
- Intersect it with the clean and worst-case system curves.
- Check current, electrical input power, speed limits and control stability.
- Test the production airflow path with the real filter, grille, heat exchanger and covers.
- Record critical component temperatures at maximum heat load and worst ambient conditions.
A useful test report keeps calculated and measured values in separate columns. If they disagree, investigate the reason instead of adjusting the calculation until it looks right. The difference may reveal an incorrect RPM assumption, installation loss, density error, unstable operating point or measurement problem.
FAQ
What are the three fan affinity laws?
For the same fan at similar conditions, airflow varies with speed, pressure varies with speed squared and absorbed fan power varies with speed cubed.
Does reducing fan speed by 20% reduce power by 20%?
No. The ideal absorbed fan power falls to about 51.2% because 0.80 cubed equals 0.512. Measured electrical input may not follow the cube exactly because motor and controller losses remain.
Do fan affinity laws predict the installed airflow?
They predict how the fan curve may scale. Installed airflow is the new intersection between that curve and the system curve, so filters, ducts, grilles and leakage still matter.
Can affinity laws compare two different fan sizes?
Only when the fans are geometrically and dynamically similar. Compact fans with different hub ratios, blade designs or frame depths should be compared with their own tested curves.
Can I use PWM duty cycle as the speed ratio?
Not safely. Use actual RPM or a verified command-to-speed map for the exact model. PWM frequency, duty range, minimum speed and internal control logic vary by fan.
Why is measured power higher than the cube-law estimate?
The calculation concerns ideal absorbed fan power at corresponding aerodynamic points. Electrical losses, control electronics, motor efficiency, low-speed friction and a changed operating point can all raise the measured input above that estimate.
