Fan selection software has made it easier than ever to generate a specification. Enter the airflow, enter the pressure, and the program returns a model number. The problem is that the software doesn't know whether the inputs are correct. It will happily size a fan for a system that doesn't exist. Most selection errors aren't software errors-they're input errors. And they're almost always discovered only after the fan is installed, balanced, and found to be deficient.
Density Correction Oversights
The most persistent error in fan selection is ignoring air density. Catalog performance data is typically published at standard air density-1.2 kg/m³ or 0.075 lb/ft³, corresponding to 20°C and 101.325 kPa. When the actual operating density differs, the fan's pressure development and power consumption change.
For altitude applications, the error is straightforward. At 1,500 meters above sea level, air density is roughly 85% of sea-level value. A fan selected without density correction will deliver the correct volumetric flow but only about 85% of the design pressure. The system doesn't get the airflow it needs. The correction is simple-multiply the required pressure by the density ratio-but it's frequently skipped because the software default is standard air.
Temperature compounds the problem. A process exhaust system at 150°C operates at roughly 75% of standard density. Combine altitude and temperature-a high-altitude industrial facility with hot exhaust-and the density ratio might be 0.65 or lower. The uncorrected selection could be undersized by 35%.
Confusing Static Pressure with Total Pressure
Selection programs ask for pressure. The user enters a number. But is it static pressure or total pressure? The difference matters.
Static pressure is the usable force for overcoming system resistance. Total pressure includes both static and velocity pressure. If a system requires 3.0 inches of static pressure at the fan discharge and the velocity pressure is 0.4 inches, the fan must develop 3.4 inches of total pressure. Selecting a fan on 3.0 inches of total pressure when the system needs 3.0 inches of static pressure means the fan is undersized by the velocity component.
The error goes the other way too. Specifying total pressure when static pressure is what the system actually requires results in an oversized fan, higher cost, and potentially motor overload if the operating point shifts.
System Curve Addition Errors
When multiple branches connect to a common main, the system curve isn't simply the sum of the branch resistances. Each branch has its own resistance, and the shared main carries the combined flow. The correct approach is to add the pressure drops of series components at a common flow, then combine parallel paths by adding flows at common pressures.
The common mistake is adding all resistances as if they were in series. This produces a system curve that is too steep, leading to fan selection at too high a pressure point. The installed system then delivers more airflow than designed-which may sound like a good problem until the increased velocity causes noise, erosion, or terminal device malfunction.
Filter and Coil Resistance at Wrong Conditions
As covered in our filtration discussion, filter resistance changes dramatically as the filter loads. The error is selecting the fan based on clean-filter resistance and ignoring the end-of-life value. The system performs adequately when new but degrades over the filter service interval.
The same logic applies to coils. Coil resistance is often specified at the design face velocity, but as airflow varies-particularly in variable speed systems-the actual coil pressure drop follows the square law. Using a single-point coil resistance value across the operating range leads to incorrect system curve construction.
Unit Conversion Errors
The seemingly trivial task of unit conversion causes a disproportionate number of field problems. Inches of water column to Pascals, CFM to m³/h, and horsepower to kilowatts each introduce opportunities for error. A factor of 249 or 0.004-depending on direction-is easy to mistype. A decimal point error in a pressure specification can shift the entire selection.
The more insidious version is mixing units within the same calculation. A designer calculates resistance in Pascals, then adds a component specified in inches of water column without conversion. The result is a system curve that bears no relation to the actual system.
Ignoring System Effect in the System Curve
The system curve should represent the actual installed system, including system effects at the fan connections. Most designers construct the system curve from duct and component losses alone, then compare it to the catalog fan curve. The system effect is applied later-if at all-as a separate correction.
The correct approach is to include system effect in the system curve from the start. The combined curve represents what the fan actually sees. Comparing the catalog curve to an incomplete system curve and then applying a system effect correction to the fan produces a different-and less accurate-result than comparing the fan curve (with system effect) to the complete system curve.
Motor Power Margins
Selection software typically calculates the required brake horsepower based on the fan's mechanical efficiency at the operating point. The error is treating this calculated value as the motor size. Motors require service factors-typically 1.15 for continuous duty. A fan requiring 9.5 kW at the operating point needs a 11 kW motor, not a 10 kW motor.
The margin also needs to account for motor efficiency degradation over time, voltage variations, and the possibility of system changes that increase resistance. Undersizing the motor by ignoring the service factor leads to overheating, nuisance tripping, and premature motor failure.
MIDI Industrial Selection Process
MIDI Industrial's selection process addresses these common errors systematically. Density corrections for altitude and temperature are applied as standard. Pressure specifications are verified as static or total before selection. System curves are constructed using proper series and parallel combination methods, with system effect included from the start. Filter loading and coil resistance are evaluated at operating conditions, not just design points. The result is a specification that holds up in the field, not just on paper.

