Before the supply fan is selected, check whether the coil sheet quotes a wet or a dry air pressure drop, and read it with the entering condition: dry- and wet-bulb at the peak dew point case. Then check leaving dew point against room RH, face velocity below about 2.5 m/s and a realistic chilled-water ΔT.
A cleanroom is qualified on its supply airflow: air change rate, recovery time and hood velocity. Controlling room pressure by trimming supply moves those numbers with every door opening. Holding supply constant and modulating extract or return keeps the qualified airflow and still holds the pressure cascade.
For the same excess airflow, room pressure differential varies with the inverse square of the leakage area. On a finished cleanroom that area hides in door gaps, ungasketed light fittings, open socket boxes, panel joints and penetrations. Compare measured excess air with the design allowance at commissioning.
Standing condensate in an AHU drain pan sits upstream of the final filters. ASHRAE 62.1 asks for a pan sloped at least 10 mm per metre toward the outlet so it drains completely. Check the slope with a level at FAT, prime the trap after long shutdowns and inspect the pan on a schedule.
When several use points open together, a purified water loop's return flow can fall close to zero. Turbulent flow needs a Reynolds number above about 4,000; a 22 mm bore at 0.1 m/s gives about 2,200. Size for the realistic maximum draw, trend return flow and alarm before it turns laminar.
Transmitter accuracy is usually quoted as a percentage of full scale. A 1% device ranged 0–1,000 Pa carries ±10 Pa of error, the whole of a 10 Pa cascade step; one specified 0–50 Pa at 1% of full scale is within ±0.5 Pa. Re-ranging a wide unit down keeps its base error, so specify the range.
Long-lead AHUs are often ordered before a single end-of-design DQ report is signed, so a casing leakage class of L2 against a URS that asked for L1 surfaces at IQ as a deviation. Splitting DQ along the procurement packages, with each section signed before its order is released, finds it while the order can still change.
During start and stop, whichever fan moves first decides which way air flows through the door gaps. Positive suites start supply first and stop extract first; containment suites do the opposite. Interlock the order, define the response to a single fan trip and test it at OQ with a real power cut.
IEEE 519-2022 allows 8% total voltage distortion at 1 kV and below, but only 5% total and 3% per harmonic from 1 to 69 kV. A plant supplied at 22 kV has its point of common coupling on the MV side, so the stricter limits apply, and fan-wall harmonics pass straight through the transformer.
With the motor in the airstream, fan power becomes heat: about 2.1 K at 1,500 Pa and 60% efficiency, roughly 14 kW on a 20,000 m³/h unit. After the coil it reaches the room; counted twice, it oversizes a chilled-water coil that then dehumidifies poorly at part load.
A passive LC filter is tuned to one harmonic order and adds leading kvar at light load. An active filter injects the opposite current, is sized in amps and follows the load. Cleanroom fan walls run at part speed most of the day, which favours the active filter.
At 50% airflow the cube law gives a fan 1/8 of its full-flow power, but the motor is then at about 25% of rated torque and away from its nameplate efficiency. Ask for motor and drive losses, or EC fan input power, at the setback point and list it on the fan schedule.