The NEBB Cleanroom Performance Testing discipline covers a set of field tests that constrain one another: airflow delivers dilution, filters remove particles, pressure directs flow, and recovery measures the combined result. Studying these as isolated facts makes scenarios and field decisions harder than they need to be. Build a dependency map instead: for each test, write down its purpose, instrument, acceptance logic, and which other tests it depends on. Then practice paper scenarios where one test's result changes how you interpret another. Administrative details such as eligibility and scheduling sit with NEBB at nebb.org.
Classification versus performance tests: which number answers which question
Classification (ISO 14644-1) answers how clean the air was; performance testing (methods described in ISO 14644-3, applied in NEBB's cleanroom discipline) answers whether the room delivers and maintains that cleanliness reliably.
The ISO 14644 series splits cleanroom evaluation into different questions. Part 1 defines cleanliness classes by airborne particle concentration; Part 3 describes test methods such as airflow, filter installation leak, pressure differential, and recovery used to demonstrate that a room performs; Part 2 addresses monitoring to show the room continues to comply. A classification result tells you what the air contained during the test; a performance test tells you why the air contains what it does. Confusing these roles leads to testing the wrong thing or accepting a result that answers a different question.
NEBB certifies firms and individuals across disciplines including cleanroom performance testing, and the CPT discipline applies these measurement concepts to built rooms rather than laboratory theory. When you study, translate every topic into the pair: what question does this test answer, and what would make me distrust this result? A filter leak and an inadequate air change rate both raise particle counts, but they call for completely different responses. Naming the question first is what separates a test result from a diagnosis.
| Test | Question it answers | Typical instrument | Key interpretation decision |
|---|---|---|---|
| Particle concentration (classification) | What class of cleanliness did the room show? | Optical particle counter | Is the comparison to the limit made per the cited edition's method? |
| Airflow / velocity | Is enough air moving, and is it distributed as designed? | Anemometer / flow hood | Velocity criterion for unidirectional flow versus volumetric criterion for mixed flow |
| Filter integrity | Do installed HEPA/ULPA filters and seals leak? | Aerosol photometer with challenge generator | Is an exceedance a real leak or a setup artifact, and is it media or seal? |
| Pressure differential | Does flow move from clean to less clean across every boundary? | Differential pressure gauge | Does the cascade hold across all adjoining spaces and door states? |
| Recovery | Does the room return to its cleanliness state after a challenge? | Particle counter (time-based) | Is slow decay a distribution problem, a filtration problem, or a measurement artifact? |
Getting particle concentration math right when class limits span orders of magnitude
Class numbers are logarithmic: limits follow C = 10^N × (0.1/D)^2.08 with D in micrometres. Learn one anchor value and derive others, and check how your standard's edition sets sampling locations and statistical treatment.
Worked example: the Class 7 limit at 0.5 µm equals 10^7 × (0.1/0.5)^2.08 = 352,000 particles per cubic metre. The classic mistake here is treating class numbers as a linear scale, expecting Class 7 to be 'seven-tenths' of Class 10, or halving the limit with each class step. The scale is roughly tenfold per class. Derive the same value for 0.3 µm or for Class 8 and compare against a published table; if your derived figures match, the formula and your arithmetic are both solid.
Sampling locations and result treatment changed between editions of ISO 14644-1: older editions computed the number of locations from the square root of room area, while recent editions use a tabulated count, and the statistical handling of individual location results also differs. This matters in practice: a 36 m² room can demand a different number of locations depending on which edition your procedure cites. Before any calculation drill, confirm which edition your study materials and project procedures reference, then practice finding the right rule rather than memorizing one method.
Choosing between velocity and volumetric airflow criteria
Unidirectional-flow zones are judged on velocity and its uniformity; mixed-flow rooms are judged on volumetric airflow and air change rate. Choosing the wrong criterion makes even a correct measurement uninterpretable.
For unidirectional flow, such as heavily filtered ceilings or work benches designed to sweep particles away in parallel streams, the useful questions are average velocity and how evenly it is distributed. Measure across a grid on the filter face or work zone, then compare individual readings against the specified range, not only the mean. A uniform 0.45 m/s average with one corner reading half that value can still fail a uniformity criterion, because a slow patch lets particles linger instead of carrying them out of the zone.
For mixed or non-unidirectional rooms, air change rate is the working metric: total supply airflow divided by room volume. A single-point velocity reading in such a room describes that point's turbulence, not the room's dilution capacity. The decision rule is simple: if the design intent is sweeping directional flow, test velocity; if the design intent is dilution, test airflow volume and compute air changes per hour. Because scenarios state design intent in their wording, read carefully for phrases like 'laminar flow zone' versus 'dilution ventilation' before choosing a criterion — the measurement may be correct and still answer the wrong question.
Filter integrity scanning: separating a real leak from a test artifact
Integrity testing challenges filters with aerosol and scans downstream with a photometer; a reading above a defined fraction of the measured upstream concentration indicates a leak, but only after you verify the setup itself.
Worked scenario: during a PAO scan of a HEPA ceiling, the photometer reads 0.004 µg/L against a 20 µg/L upstream challenge, which is 0.02 percent and above a 0.01 percent threshold stated in the procedure. The tempting call is to condemn the filter. The better sequence is to first confirm the upstream challenge concentration is adequate and stable, verify the photometer's reference setting, and check scan speed and probe-to-face distance, all of which inflate downstream readings when wrong. Only then localize the exceedance: media leaks and housing or gasket leaks sit in different places and lead to different repairs.
Why this matters: condemning a sound filter wastes a validated component, while passing a marginal seal leak leaves a contamination path that routine particle counting may not catch until product is affected. Scan technique, including overlapping passes, consistent speed, full coverage, and the perimeter, is procedure-driven, so rehearse the sequence on paper: verify upstream challenge, set reference, scan, localize, record location and reading. An exceedance reported without a location is nearly useless to the repair crew that follows you.
Pressure cascades that hold on the gauge but fail at the doorway
A pressure cascade protects by flow direction, not magnitude alone. Verify the differential across every adjoining space and at door openings, because a room can show a healthy gauge number and still lose its cascade.
Worked scenario: a cleanroom reads +15 Pa against its corridor, comfortably above the specified minimum, and the steady-state checklist appears complete. The overlooked check is the room on the other side: the material airlock reads only +2 Pa relative to the cleanroom, and with doors cracked during a transfer, airflow direction at the opening is ambiguous. The better decision is to map the full cascade across cleanroom, airlocks, gowning area, and corridor, and verify direction at each boundary rather than certifying one large differential against the corridor.
The underlying reason is that pressure is a network property: throttling a return raises a room's gauge reading but may starve its air change rate, and adding makeup air upstream rebalances every adjacent room simultaneously. A cascade that holds at rest can still reverse during door operation if interlock timing or transfer practice is wrong. Paper scenarios reward the habit of asking what each room's differential is to every neighbor, and what changes at each door state, before accepting a single gauge reading as evidence.
Recovery testing: reading the room as one system after a contamination event
Recovery testing measures how quickly a room returns toward its defined cleanliness after a particle challenge. Because it combines dilution, filtration, and airflow pattern, it is an end-to-end check on the whole system, not one component.
The method concept: artificially elevate particle concentration in the room, record the decay curve, and compare the time to reach a defined fraction of the starting concentration, such as a 100:1 recovery, against the criterion in the procedure. A slow recovery points toward an insufficient air change rate, short-circuiting where supply air bypasses the occupied zone, or filters performing poorly. Interpreting direction matters: recovery failure alongside good filter integrity suggests a distribution problem, while the reverse combination suggests filtration.
Guard against two interpretation traps. First, a room can pass a classification test at rest yet recover poorly after a contamination event, because classification captures a snapshot while recovery captures dynamics. Second, background instrument counts or an incomplete cleanup of the added challenge can flatten the decay curve artificially. A monitoring plan in the spirit of ISO 14644-2 helps here: trending routine particle data after perturbations such as door openings or equipment startup gives you real-world recovery observations without staging a formal test.
A dependency-map exercise, study sequence, and readiness checks
Run three passes: concepts and instruments first, calculations second, linked paper scenarios third. Finish with a dependency map scored against a rubric; treat those scores as learning milestones, not pass predictions.
Practical exercise: pick a paper scenario, say a Class 8 mixed-flow cleanroom containing a unidirectional filling zone, one material airlock, and a gowning room. Draft a one-page dependency map listing every test, its instrument, its acceptance logic, and the tests it depends on. Expected observations: filter integrity feeds the classification result, air change rate feeds both pressure and recovery, and the cascade diagram must show all four spaces. Self-check rubric, 0 to 2 points per row: purpose stated; instrument and setup sequence named; acceptance logic correct; dependencies on other tests identified. With four rows, the maximum is 8 points; a score of 7 or higher means the map is internally consistent.
Concrete readiness checks: derive a Class 7 limit at 0.5 µm from the formula without notes; explain in two sentences why velocity uniformity matters before trusting a particle count; sequence a photometer integrity setup from memory; justify a pressure cascade across a three-room chain including the airlock state. If any check stalls, return to that section and rework its scenario rather than rereading passively. Eligibility, scheduling, and fee details are administrative matters handled by NEBB; confirm them directly at nebb.org.
- Week 1: concepts — the ISO 14644 division of labor and the question each test answers.
- Week 2: calculations — class limits from the formula, air change rates, and pressure differentials, all on paper.
- Week 3: paper scenarios — the leak-localization and cascade cases from this guide, then fresh ones you write yourself.
- Week 4: rebuild the dependency map from memory, audit it against your notes, and redo any rubric item scored below 2.
References and further reading
Use these references to explore the concepts and check the latest information from the relevant organizations.