EN 779 F-classes were replaced by ISO 16890 ePM1, ePM2.5, ePM10 and ISO Coarse, with no exact conversion, so a spares line reading F9 leaves each supplier to choose. Re-specify every pre-filter position in the asset register and on the purchase code, and keep HEPA on EN 1822.
Operating theatres run positive, airborne isolation rooms negative (ASHRAE 170: at least 2.5 Pa and 12 air changes per hour), protective environment rooms positive, and hazardous drug compounding rooms negative. The common failure is drift, so continuous pressure monitoring with a local alarm matters.
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.
Percent complete is an opinion; earned value is a measurement. Stage gates from concept to PQ, monthly SPI and CPI reporting, a live FMEA risk register reviewed weekly and a RACI the owner's QA team has seen give a pharma project governance that is used, not filed.
21 CFR Part 11 and EU GMP Annex 11 overlap on validation, access control and audit trails but differ in emphasis: Part 11 on electronic signatures, Annex 11 on a risk-based system lifecycle. Fixing the destination market in the design basis means the URS and validation plan for BMS, EMS and SCADA are written once.
Fan power scales roughly with the cube of speed, so right-sizing air change rates against recovery is the first lever. Dew-point control that cuts overcooling and reheat, and energy recovery where cross-contamination allows, follow. Humiley's energy modeling targets 30% lower HVAC energy with GMP limits untouched.
IQ verifies installed equipment against its records, so a file reconstructed after delivery delays qualification. Humiley compiles material certificates, weld and passivation records, the witnessed FAT report, O&M manual and spares list while each unit is built, so IQ starts from a complete record.
"Like-for-like" is a conclusion that needs evidence. Before a replacement fan, FFU or AHU is fabricated for a validated suite, compare the measured duty point, filters and seals, connections, control points and clean-zone materials with the installed item, and let QA decide in change control.
Airflow visualization shows flow patterns over the critical zone, direction at openings and the effect of interventions. It does not show particle levels, filter integrity, velocity or microbial control, and a study filmed at rest says little about operation. Plan interventions and acceptance criteria first.
A single AHU replacement in a live GMP plant carries every constraint of a large project: a short shutdown window, existing connections, a pressure cascade to re-prove and change control before fabrication or any site work starts. Planned well, requalification stays short. Planned badly, a weekend job becomes a two-week deviation.
Personnel, material, product and waste flows decide much of a facility's cross-contamination risk long before sampling starts. Stepped changing, dedicated material airlocks, forward-only product routes and separate waste exits keep them apart; time segregation is the weakest fallback. Check crossings on the layout before room data sheets freeze.
EU GMP grades map to different ISO 14644-1 classes at rest and in operation: Grade B is ISO 5 at rest and ISO 7 in operation. A URS line that quotes only one invites the wrong design. State both classes, the maximum number of people and the clean-up period.