Cleanroom testing is often bought late and on price, and QA first meets the tester as a signature on a report. Before the order, hold evidence of tester training, calibrated instruments, method statements to ISO 14644-1 and ISO 14644-3, a sample report and agreed raw data practice, then list the tester as an approved supplier.
A log entry of 'fan tripped on overload, reset, running' names no cause. Five whys, with evidence for each answer, takes the trip back through full fan speed and loaded filters to an alarm that notified nobody. Stop at a cause the site can change, assign owners, and check later that it worked.
PDE, permitted daily exposure, is a substance-specific dose unlikely to cause an adverse effect if a person is exposed at or below it every day for a lifetime. In a shared facility it is the basis for cleaning limits and for how much HVAC separation the design needs. Ask for it at URS stage.
Most calibration certificates are filed unread. Check that the laboratory is accredited for this measurement, that tag, serial number and calibrated points match the instrument and its range, read as-found before as-left, compare the uncertainty with the tolerance, and look for traceability and the decision rule behind any pass statement.
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.
A capture hood throttles the terminal it measures and assumes an even flow across its sensing grid, so on a swirl diffuser it can read well off the true airflow, in either direction. Name the method per terminal type in the protocol, use back-pressure compensation, and check the hood against a duct traverse.
ICH Q1A(R2) defines a pilot scale batch for solid oral dosage forms as generally at least one-tenth of production scale or 100,000 tablets or capsules, whichever is the larger. Two of three primary stability batches should reach it. At 500 mg a tablet that is 50 kg of blend, which sets the equipment size.
A hot WFI tank pushes saturated vapour through its vent filter. In an unheated housing the vapour condenses and chokes the cartridge, and the next fast draw-down can pull a vacuum on the vessel. Heat the housing, size the filter for the fastest inbreathing case, and fit vacuum protection with an alarm.
A pressure monitor that reads the same value for a month may be right, or stuck. The CDC guidelines ask for monitoring preferably daily, with documentation. A short round covers the monitor and its alarm, smoke at the closed door, the door closer and latch, the anteroom doors and the exhaust fan.
Adjusting each diffuser to its design flow in turn goes round the room without settling, because terminals on a branch share one duct pressure. Proportional balancing sets every terminal to the same ratio of design, with the lowest-ratio terminal left fully open, and only then brings the fan to the design total.
ICH Q1D (2002) describes two reduced stability designs. Bracketing tests only the extremes of a factor such as strength or container size, at all time points. Matrixing tests a selected subset of all the combinations at a given time point. Either should be justified, and either can end in a shorter shelf life.
EU GMP Annex 15 (2015) says a formal release for the next qualification stage should be authorised, and that conditional approval to proceed can be given where there is a documented assessment of no significant impact on the next activity. That assessment is made item by item, against the tests that follow.