A locked isolator does not stop a fan that windmills in the draught from a neighbouring unit. Isolating an AHU means identifying every energy source, moving air included, shutting down and isolating, one personal lock per worker, releasing stored energy, and trying to start the unit before the door is opened.
A chilled water coil piped in parallel flow still cools, so the error often survives commissioning. It shows as a valve at 100%, return water colder than design and an off-coil temperature that misses setpoint. Mark inlet and outlet at the factory, check at FAT, and look again before the insulation goes on.
UV-C lamps limit microbial growth on the coil surfaces the light reaches. They do not remove dust or scale from the fin pack, and lamp output falls with running hours. After fitting them, keep coil inspection and cleaning in the maintenance plan and add lamp hours, a replacement date and an interlock test.
A standby generator run with no load each month proves that it starts. NFPA 110 sets the monthly exercise for diesel sets at a minimum of 30 minutes, at not less than 30% of nameplate kW or at the exhaust temperature the manufacturer recommends. Record load, time and temperatures at every run.
A thermal scan of a lightly loaded panel finds little, because heat at a bad joint rises roughly with the square of the current. Check the load recorded at the time of the scan, the line of sight, the emissivity setting, the comparison between phases, and that each finding has an image, a temperature difference and a severity.
EN 1822-1:2019 places E10, E11 and E12 filters in the EPA group, below HEPA. Their class rests on an efficiency determined on a statistical basis, and the standard says group E filters cannot and shall not be leak tested for classification. H13 and H14 filters are individually leak tested. Use EPA as a pre-stage.
A HEPA filter leaves the factory tested, then travels by container, forklift and scaffold. Check the cartons on arrival, match label class, serial number and rated airflow to the schedule and test report, store dry within the stacking mark, open only at a clean work area and inspect media and gasket under strong light before fitting.
EN 1822-4:2009, since replaced by EN ISO 29463-4, allowed a HEPA filter to be repaired and retested: all repairs together, whoever made them, up to 0.5% of the face area, none longer than 3.0 cm, unless buyer and seller agreed otherwise. Most purchase orders are silent. State whether repairs are accepted and have each one marked on the scan report.
Outlet dew point is the last thing to check on a new desiccant dehumidifier. First confirm rotor direction, speed and seals, set and measure the process and reactivation airflows, and prove the heater and rotor interlocks. Then set the reactivation temperature and compare inlet and outlet dew point with the selection.
N+1 on a fan wall datasheet is a claim until one fan is off. Without a backdraft damper, air runs back through the stopped fan, and with six fans the other five must each carry 20% more air with motor headroom to match. A single-fan-off step at FAT shows airflow, currents and recovery time.
An electric heater in an AHU overheats when airflow stops. At the factory test, witness each link of its safety chain: the airflow interlock, the automatic-reset and manual-reset high-limit cut-outs, the fan run-on, the clearance to filters and sensors, and a contactor that can cut power if the solid-state relay fails conducting.
The ISO 14644-3 installed leak test needs an upstream aerosol injection point, an upstream sample port and a perimeter seal the probe can reach. With a photometer, above 0.01% of upstream concentration is commonly treated as a leak. Specify terminal housings from the test backwards.