1.1 × 1.1 × 1.1 = 1.33, which is how three modest 10% margins leave a cooling coil a third bigger than its load.
10% on the cooling load, 10% more airflow than that load needs, 10% on the coil for fouling. Nobody chose 33%.
Each engineer added 10% at their own stage, on a number that already carried the previous 10%, so the margins compound rather than add. Put an owner's allowance for future expansion on top and a fourth 10% makes it 1.1⁴ ≈ 1.46, about 46%.
The cost shows up after handover. An oversized chilled-water coil under temperature control runs at part load for most of the year. The valve closes down, the coil surface warms up, and it condenses less water out of the air. Room temperature sits on setpoint while RH climbs, and in a humid climate RH was the reason you needed that coil. With DX the same oversizing means short cycling, and the water left on the coil evaporates back into the airstream each time the compressor stops.
Our preference is one margin, in the design basis, with the uncertainty it covers written next to it. Then check the selected coil at part load as well as at peak, at the dew point the rooms have to hold.
Humiley writes the design basis so the same numbers govern the drawing, the factory acceptance test and the maintenance plan, and a margin added later stands out against it.
On your last AHU, could anyone say which stage added the most margin?




