IEEE 519 is measured at the point of common coupling, and a fab's tools sit far downstream of it. Harmonics from fan walls, pumps and chiller drives distort the shared LV bus. Separate tool transformers, DC chokes in the drive spec, filters sized from a study and a voltage THD check at each tool bus keep it off the tools.
A gap marked closed needs evidence someone else can check: a revised drawing and calculation for a design gap, the same test repeated for a performance gap, an effective SOP with training records for a procedural gap. Agree the evidence before the work starts, and have someone other than the fixer sign it.
When one location in a Grade C room recovers far slower than the rest, raising a 60 m³ room from 20 to 30 air changes adds 50% more air and may barely help. A smoke study usually shows short-circuiting to a return; moving or adding a low-level return is the cheaper fix.
"OK" on a FAT sheet records a judgement. "118 Pa, acceptance ≤ 150 Pa", with the gauge ID and calibration date, records a measurement that IQ, maintenance and an inspector can use years later, and shows when a pass was close. It is how the Excellence value looks on a test bench.
At 40 °C condensing, the Carnot COP rises from 7.95 at 5 °C evaporating to 8.21 at 6 °C, about 3.3% per kelvin. Real chillers differ but follow the direction, so a GMP plant should not run every coil at the temperature only the fresh-air dehumidification coil needs.
Filter transmitters get bought twice, or not at all, when the tender never says who supplies them. An AHU/BMS interface schedule lists every sensor, drive signal, actuator and fire alarm contact with who supplies, mounts, wires and configures it, attached to both tenders before they go out.
316L caps carbon at 0.03% in plate (ASTM A240) and 0.035% in sanitary tube (A270), against 0.08% for 316. Low carbon limits chromium carbide beside the weld; heat tint from a poor purge and weak passivation are more common starting points for rouge and pitting in orbital-welded tube, so read the mill certificate against the product form.
The 2022 Annex 1 dropped the 1 m³ per location rule and points to ISO 14644-1, where Grade A at 0.5 µm needs about 6 litres a location (at least 2 L and 1 minute). Classification time goes on the number of locations including critical points, at-rest and in-operation runs, and 5 µm counting if the CCS calls for it.
EEMUA 191 and ISA-18.2 treat about one alarm per 10 minutes per operator as manageable. A month of EMS alarm log, split by shift and sorted by tag, shows the busiest and the standing alarms. Rationalise those first, and turn alarms with no operator action into logged events.
At EMS OQ, change a setpoint and check the entry holds old and new value, user, time and reason; try to pause the trail from operator and supervisor accounts; move the workstation clock. Archive readability after upgrades has to be planned, because OQ cannot show it.
The sterilisation route sets the filling room grade and most of a sterile line's HVAC cost. Terminally sterilised products typically fill in Grade C; aseptic filling needs Grade A over a Grade B background or an isolator, with monitoring and process simulations. Fix the route before the layout.
The installed HEPA scan covers media, frame and the filter-to-housing seal, but not the flange between housing and ceiling. Where the ceiling void runs at higher pressure than the room, unfiltered air enters there. Seal it with an evenly compressed gasket and sealant bead, then probe the flange and file the result.