Cleanroom Cascade Pressure: A Linear Screening Model Against a Leak That Is Not Linear

The Calculation Runs Backwards From the Rest of the Series

Every airflow calculation in building services asks the same kind of question. How much air is needed to carry the heat away, to dilute the contaminant, to make up what the hoods take out. The quantity being solved for is a flow, and the difficulty lies in establishing the requirement it has to satisfy.

This calculation asks the reverse. Given a quantity of air, what pressure does it produce. The flow is the input and the pressure is the answer, and the difficulty moves with the direction.

A pharmaceutical cleanroom is held at a pressure above the space next to it so that any air crossing the boundary crosses in the intended direction. That pressure exists because more air is supplied to the room than is taken out of it. The surplus has nowhere to go but through whatever gaps the envelope has, and the pressure in the room is whatever it takes to push the surplus through those gaps.

The quantity that decides the answer is therefore the leakiness of the room. It is also the one quantity nobody measures at design stage, because at design stage the room does not exist. The calculator handles this with a four-position selector running from Very Tight to Leaky, which is an honest way of dealing with something unknowable in advance and a coarse one.

There is a second feature that makes the relation unfamiliar. Flow through a gap does not rise in proportion to the pressure across it. It rises more slowly, following a power law whose exponent for room envelopes commonly sits near two thirds, which means that the pressure rises faster than the flow. A relation that treats the two as proportional agrees with the real behaviour at one point and departs from it on both sides of that point.

What follows covers where the cascade pressure comes from, what the screening coefficient implies about the gaps behind it, how far a linear treatment departs from a power law on either side of the offset it was pinned at, what happens to the whole arrangement the moment somebody opens the door, and why a result sitting inside the calculator's own NORMAL band is not a statement about compliance with anything.

Calculator Inputs: Two Airflows and a Tightness Class

The field list is three entries long. Two of them are airflows and the third is a selection.

Field	Imperial unit	Metric unit	What it is
Supply Airflow	CFM	m³/h	Air delivered into the room
Exhaust / Return Airflow	CFM	m³/h	Air taken out of the room
Room Leakage Tightness	Very Tight, Tight, Standard, Leaky	same	An assumed class for the envelope

The relation is two steps.

Offset = max(0, Supply − Exhaust)
ΔP = Offset × K × F_leak

The calculator classifies in pascals and converts for display, so the metric coefficient is the canonical one and the imperial figure follows from it.

K       = 0.0586 Pa per m³/h
F_leak  = 1.25  Very Tight
          1.10  Tight
          1.00  Standard
          0.75  Leaky

Stated per cubic foot per minute the same coefficient is 0.0004 in. w.c., and the two statements are the same model rather than two models. Converting the imperial figure across gives 0.0004 × 0.58858 × 249.089 = 0.05864 Pa per m³/h against the published 0.0586, a difference of seven hundredths of one percent that comes from rounding the imperial figure to one significant place. The calculator runs the pascal path in both unit modes, so switching units does not move a result across a band edge.

The result is placed in a band, and the bands are these.

Cascade pressure	Class
Below 7 Pa (0.0281 in. w.c.)	LOW
7 to 15 Pa (0.0281 to 0.0602 in. w.c.)	NORMAL
15 to 25 Pa (0.0602 to 0.1004 in. w.c.)	HIGH
Above 25 Pa (0.1004 in. w.c.)	VERY HIGH

These bands are the calculator's own. The imperial edges are conversions of the pascal edges rather than separate thresholds, which is why they carry four decimal places. Section eleven covers what the bands are not.

The lower bound matters as much as the coefficient. Where the extract is equal to or greater than the supply the calculator returns zero. The model describes rooms held above their surroundings, and the negative-pressure case used when potent materials are handled falls outside it.

What the field list does not contain is most of what decides the answer. The area and character of the leakage paths, which set the actual relation between flow and pressure. The area of the doors and pass-through devices. The number of rooms in the cascade. Whether the doors are shut or standing open.

Where the Pressure Actually Comes From

The pressure is not something the air handling system delivers. It is the resistance the envelope offers to air that has nowhere else to go.

Air supplied into the room in excess of what is extracted has to leave through the envelope: through the gaps around the doors, through the pass-through hatches, through the joints between wall panels and the penetrations where services cross them. Flow through those paths requires a pressure difference to drive it, and the difference that settles is the one at which the total flow through all of them equals the surplus being supplied.

The consequence is that the pressure difference is an outcome rather than a setting. A constant-offset system holds an airflow, not a pressure. The pressure arranges itself according to how easily the air can get out.

That is why two rooms of identical volume given an identical offset can sit at different pressures. The offset is known to the designer exactly, because it is the difference between two valve settings. The tightness of the envelope is not known at all until the room is built.

The same mechanism works in both directions over the life of the installation. Door seals wear, the leakage area grows, and the same airflow produces a smaller difference. Sealing work at commissioning does the opposite: the difference rises at unchanged airflow, and the offset may need resetting to bring it back down.

This is also where the control strategy enters. A system controlling on pressure measures the difference and moves the airflow to hold a setpoint, so the tightness of the room changes the airflow rather than the pressure. A system running a fixed offset holds the airflow and lets the pressure follow the envelope. The calculator describes the second of the two.

Per ISO 14644-4 and the ASHRAE Handbook chapter on clean spaces: the pressure difference is the consequence of forcing a surplus airflow through the leakage paths of the envelope, so it follows from the tightness of the room as much as from the airflow offset.

The Model Is Linear and the Leak Is Not

Flow through a gap follows a power law, and a relation proportional to the flow reproduces it at exactly one point.

Q = C × ΔPⁿ

Q is the leakage flow in volume per unit time. C is a coefficient set by the total area and character of the leakage paths. ΔP is the pressure difference across the envelope. The exponent n is bounded by 0.5 at one end and 1.0 at the other, and measured building envelopes commonly fall between about 0.6 and 0.7, with 0.65 the usual working value.

The two ends of that range are two flow regimes. An exponent of 0.5 is the orifice case, where flow goes as the square root of the pressure, and it is a special case rather than the general law. An exponent approaching 1.0 belongs to paths dominated by viscous resistance, which is what a long narrow slot is. A room envelope contains paths of both kinds in unknown proportion, so the exponent describing the whole of it lands between them.

Inverting the relation gives the pressure as a function of the flow.

ΔP = (Q / C)^(1/n)

At n = 0.65 the inverse exponent is 1.54 and at n = 0.5 it is 2. The calculator's linear relation corresponds to an inverse exponent of 1, which is outside the range entirely.

The size of the departure shows best with all three relations pinned at one point. The imperial worked example gives 0.040 in. w.c. (9.96 Pa) at an offset of 100 CFM (170 m³/h), so that is the point the comparison uses.

Offset	Linear model	Exponent 0.65	Exponent 0.5
50 CFM (85 m³/h)	0.020 in. w.c. (4.98 Pa)	0.0138 in. w.c. (3.43 Pa)	0.010 in. w.c. (2.49 Pa)
100 CFM (170 m³/h)	0.040 in. w.c. (9.96 Pa)	0.040 in. w.c. (9.96 Pa)	0.040 in. w.c. (9.96 Pa)
200 CFM (340 m³/h)	0.080 in. w.c. (19.9 Pa)	0.116 in. w.c. (28.9 Pa)	0.160 in. w.c. (39.8 Pa)
300 CFM (510 m³/h)	0.120 in. w.c. (29.9 Pa)	0.217 in. w.c. (54.0 Pa)	0.360 in. w.c. (89.6 Pa)

All three agree at the calibration point and nowhere else, and the departure grows with distance from it in both directions. At twice that offset the linear model reads about a third low against an exponent of 0.65 and half low against an exponent of 0.5. At half the offset it reads about 1.45 times high and twice high respectively.

The calculator page describes its coefficient as a fixed empirical screening value rather than a leakage model, and the table above is what that description amounts to numerically: the result is trustworthy near the offsets the coefficient corresponds to and progressively less so away from them.

Per the ASHRAE Handbook, Fundamentals, chapter on ventilation and infiltration: leakage through a building envelope follows a power law whose exponent commonly falls near two thirds, so a relation proportional to the offset agrees with it at one point and departs from it above and below.

What the Coefficient Implies About the Gaps

Working the coefficient backwards gives an equivalent leakage area, and the figure that comes out is recognisable, which is a useful check on whether the model corresponds to a room anybody has built.

The reconstruction uses the orifice form, which means it adopts an exponent of 0.5 and a discharge coefficient, and both of those are assumptions rather than measurements.

Q = C_d × A × √(2 ΔP / ρ)

C_d is the discharge coefficient, taken here as 0.65, within the range commonly assumed for gaps and cracks. A is the equivalent leakage area. The density ρ is taken as 1.2 kg/m³ (0.075 lb/ft³) for air at ordinary room conditions.

Running the imperial worked example through it, at an offset of 100 CFM (0.0472 m³/s) and a difference of 0.040 in. w.c. (9.96 Pa), the velocity through the gaps is the square root of 2 × 9.96 / 1.2, which is 4.07 m/s (802 FPM). Dividing the flow by that velocity and by the discharge coefficient gives an equivalent leakage area of 0.0178 m², which is 178 cm² or 0.19 ft².

That number has an obvious physical comparison. A door leaf of 2.1 by 0.9 m (6 ft 11 in by 2 ft 11 in) has a perimeter of 6.0 m (19.7 ft), and a uniform gap of 3 mm (0.12 in) around that perimeter is 0.018 m², the same figure.

So the coefficient describes a room whose total leakage is about one door with an ordinary gap and no threshold seal, which for a small cleanroom is plausible. For a room with two or three doors, a pass-through hatch and a run of service penetrations the real area is larger, C is larger with it, and the same offset produces less pressure than the calculator returns.

The reconstruction is an illustration of order rather than a property of any room. It was carried out at an exponent of 0.5 and at an assumed discharge coefficient, and at 0.65 the relation between area and pressure is a different one. What it establishes is that the screening coefficient is not describing something physically strange, not what the leakage area of a particular cleanroom is.

Per the ASHRAE Handbook, Fundamentals: reconstructing an equivalent leakage area from a screening coefficient under stated assumptions gives a figure comparable with the gap around a single door, which is a useful check that the model corresponds to a physically plausible room.

Tightness Classes Carry More Than the Multipliers Show

The four-position selector stands in for a quantity that varies over a wide range, and the range the multipliers span is narrower than the range real rooms do.

What actually decides the tightness of a cleanroom envelope is a list of construction details. Door hardware, and specifically whether there is a threshold seal, how well the leaf closes onto its stops and how tightly the frame meets the wall. Pass-through hatches and the condition of their gaskets. Service penetrations where ductwork, pipework and cable containment cross the boundary. The detailing of panel joints and of the junctions at ceiling and floor. The state of every one of those seals on the day the measurement is taken.

The multipliers move the result between 0.75 and 1.25, which is a ratio of one to one and two thirds between the extreme classes. They are screening values belonging to the calculator. Translating them into a ratio of leakage areas would require choosing a power-law exponent and holding calibration data to fix the coefficient, and the model provides neither, so that translation is not made here.

Real rooms separate further than that. A room whose doors have no threshold seal and one with a full set of seals differ in leakage area by more than the multipliers differ, and a single room before and after sealing work differs in the same way.

The selection is therefore an assumption whose influence on the result is bounded by the span the multipliers cover, and the actual tightness is established by measurement on the completed room.

The practical responses are three. Provide enough capacity in the supply and extract systems that the offset can be changed at commissioning, once the real tightness is known. Prefer pressure control over fixed offset where the difference has to be held to a tolerance. Measure the difference at qualification and periodically thereafter.

Per ISO 14644-4 and the ISPE Baseline Guide for sterile product manufacturing facilities: envelope tightness depends on door hardware, transfer devices, service penetrations and joint detailing, and it is established by measurement on the completed room rather than assumed at design stage.

The Gradient Vanishes When the Door Opens

The pressure difference exists only while the boundary is closed. At the moment the room is used for its purpose, the mechanism it depends on is gone.

The arithmetic is short. The doorway of a 2.1 by 0.9 m leaf is 1.89 m² (20.3 ft²). The equivalent leakage area of the closed room reconstructed in the previous section is 0.0178 m² (0.19 ft²). The ratio is 106 to one, two orders of magnitude.

Push the same surplus through the larger area and the pressure that develops is not small, it is absent. At an offset of 100 CFM (0.0472 m³/s) the velocity through the opening is 0.0472 divided by 0.65 × 1.89, which is 0.038 m/s (7.6 FPM). The pressure corresponding to that velocity is 0.5 × 1.2 × 0.038², which is 0.0009 Pa (0.0000036 in. w.c.). No instrument used for cascade measurement resolves that, and no cascade exists across an open door.

Directional control through an open doorway is therefore not a matter of pressure at all. It is a matter of velocity through the opening, and the velocity is set by the airflow passing through it. A frequently used design figure for holding direction through a doorway is of the order of 0.5 m/s (98 FPM), and at 1.89 m² that requires 0.945 m³/s, which is 3,400 m³/h (2,000 CFM).

Set that against the 100 CFM (170 m³/h) offset holding the cascade with the door shut and the ratio is twenty to one. An offset sized to develop a pressure difference across a closed envelope does not come close to holding direction through an open doorway, and no realistic increase in it would.

What protects the room during passage is the airlock. Splitting the transition into two stages with an interlock that prevents both doors standing open at once means the boundary is never open on both sides at the same time. The difference re-establishes itself between openings, and how quickly depends on the volume of the airlock and the airflow into it.

Per ISO 14644-4 and the ISPE Baseline Guide: an open doorway presents a leakage area two orders of magnitude larger than a closed envelope, so the pressure difference collapses and directional control during passage depends on the airlock arrangement and on procedure rather than on the cascade value.

Cascades Add Along the Chain

The calculation treats one boundary. A facility is a sequence of them, and what the plant has to develop is the sum along the path.

Rooms are arranged in order of increasing cleanliness requirement, and each is held above the less clean room next to it. The difference between the two ends of the chain is the sum of the differences at every boundary in between.

Four boundaries at 15 Pa each give 60 Pa (0.241 in. w.c.) between the core and the outside. At 10 Pa each the total is 40 Pa (0.161 in. w.c.). Neither figure appears anywhere in a single-boundary calculation.

The consequences land on the plant at both ends. The air handling unit serving the core has to develop that total in addition to the resistance of its own ductwork, coils and filters. The extract system serving the least clean room in the chain works against the same total with the sign reversed.

The total also shows up at the doors. The net pressure force on a leaf is the difference multiplied by its area, so at 60 Pa across 1.89 m² it is 113 N (25 lbf) before any latch or closer force is added, and what that means at the handle depends on the leverage of the leaf. A chain long enough to accumulate a large total produces doors that are noticeably hard to open, which is an operational problem and in some arrangements a safety one.

Three responses are usual. Reduce the difference at individual boundaries so that the sum stays manageable. Use airlocks at which the sign of the cascade reverses, which breaks the chain into segments that do not accumulate. Arrange the systems so that the most heavily loaded boundaries are the least frequently used.

Per ISO 14644-4 and EU GMP Annex 1: pressure differences along a cascade add, so the total across a sequence of rooms follows from the number of boundaries as much as from the difference at each one.

Containment Reverses the Sign

A room may need to be held below its surroundings rather than above, and a facility doing both needs a boundary where the direction changes.

The two purposes pull in opposite directions. Protecting the product requires that air does not enter the clean room from outside it, which means holding the room above its surroundings. Protecting the operator and the environment from the product requires that air does not leave the room, which means holding it below.

Both requirements meet in the same building wherever potent compounds are handled under conditions that also demand cleanliness. A room held below the corridor while still being cleaner than the corridor is an ordinary arrangement, and the cleanliness comes from filtration and air change rate rather than from the pressure relationship.

The contradiction is resolved at the airlock. A positive-pressure airlock sits above both the corridor and the room, so air leaves it in both directions and neither space feeds the other. A negative-pressure airlock sits below both, so air enters it from both sides and is extracted from it. Which one is used follows from the contamination control strategy and from what is being protected first.

The calculator takes no position on any of this because it does not reach the case. Where the extract equals or exceeds the supply it returns zero rather than a negative value. The same relation applies in magnitude to a room held below its surroundings, and the calculator does not cover it.

What sets the magnitude of a negative difference is the same pair of quantities: the airflow offset and the tightness of the envelope. What sets the requirement is the contamination control strategy and the risk assessment behind it.

Per EU GMP Annex 1 and ISO 14644-4: rooms handling potent materials may be held below their surroundings while remaining cleaner than them, and a facility combining both requirements needs an airlock at which the direction of the cascade changes.

What the Bands Do Not Certify

The classification the calculator returns describes the magnitude of the pressure difference against the calculator's own scale. It is not the scale a regulator applies.

The bands reflect practical ranges over which a cascade is weak, workable or excessive, and the page states them as its own decision model. That statement is not a disclaimer added out of caution. It is the load-bearing part of how the result should be read.

EU GMP Annex 1 gives 10 Pa as a guidance minimum between adjacent rooms of different grades, subject to the contamination control and containment strategy of the facility. That value sits inside the calculator's NORMAL band rather than at its lower edge. The NORMAL band begins at 7 Pa (0.0281 in. w.c.), so a result anywhere between 7 and 10 Pa is classified NORMAL by the calculator and is below the Annex 1 guidance figure at the same time.

That is the whole of the problem with reading the badge as a verdict. A label reading NORMAL is easy to take as confirmation that the design is acceptable, when what it reports is that the number landed in the middle of a screening scale. The Leaky row of the worked example below returns 7.47 Pa, which is NORMAL and below 10 Pa.

Several other things also fail to follow from a pressure difference. The cleanliness classification of the room, which is established by airborne particle concentration under ISO 14644-1 and not by pressure. The air change rate. The recovery performance and the airflow visualisation results. None of these is derivable from a cascade value, and a design that satisfies one of them has not thereby satisfied the others.

What establishes conformity is qualification of the completed installation under the applicable procedure, with measurements at handover and a monitoring regime in operation, documented against the acceptance criteria written for that facility.

Per EU GMP Annex 1 and ISO 14644-4: guidance values for pressure difference between adjacent zones of different grades are set by the applicable document, and a screening classification does not establish conformity with them.

Measuring the Difference Is Its Own Problem

The quantity is small, the instruments that read it respond to things other than the quantity, and the reading depends on where the taps are.

A difference of 15 Pa is about 0.015 percent of atmospheric pressure. Measuring it calls for an instrument with resolution and zero stability to match, and the zero drift of a transducer over months is a real contributor to what a monitoring system reports.

Several things move the reading without the room having changed. Tap position, since a tap near a supply terminal or an extract grille reads the local velocity field rather than the room. Wind on the facade, which reaches rooms on the perimeter through the building envelope. The operation of systems serving neighbouring rooms. A door opening anywhere along the chain, which redistributes the whole cascade for as long as it is open.

The countermeasures are equally practical. Place taps away from terminals and grilles and at a height out of the local airflow. Use damped instruments so that short-lived fluctuations do not trigger alarms that carry no information. Keep continuous monitoring and qualification measurement distinct in purpose: monitoring watches for departure from a normal state, while a qualification measurement establishes the value under recorded conditions.

What has to be recorded alongside the value is the condition it was taken in: tap locations, the state of the doors throughout the cascade at the time, and what the neighbouring systems were doing. Without those, the reading cannot be reproduced and the comparison with a later reading means less than it appears to.

Per ISO 14644-3 on test methods: pressure difference measurement depends on tap location, on the state of doors throughout the cascade and on external conditions, so a recorded value is meaningful only alongside the conditions under which it was taken.

Worked Example: 100 CFM Into a Standard Room

The scenario is the imperial example the calculator page carries.

Input	Value
Supply airflow	1,200 CFM (2,039 m³/h)
Exhaust / return airflow	1,100 CFM (1,869 m³/h)
Leakage tightness	Standard

Step 1. The offset. 1,200 − 1,100 = 100 CFM (170 m³/h).

Step 2. The tightness multiplier. Standard gives F_leak = 1.00.

Step 3. The pressure difference.

ΔP = 100 × 0.0004 × 1.00 = 0.040 in. w.c.

The calculator runs the same case through its pascal path, 100 CFM × 1.699 × 0.0586 × 1.00 = 9.96 Pa, and converts that to 0.0400 in. w.c. for display.

Step 4. The classification. 9.96 Pa falls between 7 and 15, so the calculator returns NORMAL. That is its own category, and at 9.96 Pa the result is below the 10 Pa Annex 1 guidance minimum by four hundredths of a pascal.

Step 5. What the offset is as a fraction of supply. 100 out of 1,200 CFM is 8.3 percent, an ordinary figure for a cleanroom, set by the balancing dampers or terminal units of the two systems.

Step 6. What the tightness class does.

Class	F_leak	Result	Band
Very Tight	1.25	0.050 in. w.c. (12.45 Pa)	NORMAL
Tight	1.10	0.044 in. w.c. (10.96 Pa)	NORMAL
Standard	1.00	0.040 in. w.c. (9.96 Pa)	NORMAL
Leaky	0.75	0.030 in. w.c. (7.47 Pa)	NORMAL

All four land in the same band at this offset, because the multipliers span a range narrower than the band is wide. The selection standing in for the least known quantity in the problem does not change the badge. It does change whether the result clears 10 Pa: the top two do and the bottom two do not.

Step 7. What the offset does. Doubling it to 200 CFM (340 m³/h) takes the linear model to 0.080 in. w.c. (19.9 Pa), which is HIGH. A power law at an exponent of 0.65, pinned to the same calibration point, gives 0.116 in. w.c. (28.9 Pa), which is VERY HIGH. The divergence between the two treatments is enough to move the result across a band boundary.

Step 8. What an open door does. At a doorway area of 1.89 m² (20.3 ft²) the same offset produces about 0.0009 Pa, which is no cascade at all. Holding direction through the opening instead takes something of the order of 2,000 CFM (3,400 m³/h), twenty times the offset.

Step 9. What the result does not establish. Conformity with the difference required between zones of different grades in the applicable document. The cleanliness classification of the room. The behaviour of the cascade while neighbouring rooms run or while any door in the chain stands open.

Step 10. What to do with it. Compare it against the difference the project requires, not against the band. Provide capacity to change the offset at commissioning. Check the door forces against the accumulated total along the chain rather than against this single boundary.

Metric Example and the Offset Behind a Target

The scenario is the metric example the page carries, and it is a different room in a different class.

Input	Value
Supply airflow	2,000 m³/h (1,177 CFM)
Exhaust / return airflow	1,800 m³/h (1,060 CFM)
Leakage tightness	Tight

The offset is 200 m³/h (118 CFM), the multiplier is 1.10, and the result follows directly.

ΔP = 200 × 0.0586 × 1.10 = 12.89 Pa

That is 0.0518 in. w.c., it rounds to 13 Pa for reporting, and it falls between 7 and 15, so the classification is NORMAL. Unlike the imperial case it also clears the 10 Pa Annex 1 guidance figure, which is a separate statement from the band it landed in.

Running the same case the imperial way is a check on the two coefficients. 118 CFM × 0.0004 × 1.10 gives 0.0519 in. w.c., which is 12.93 Pa against 12.89 Pa from the metric path. The gap of three tenths of one percent comes from rounding the offset at 118 CFM in the conversion, not from a disagreement between the coefficients.

The more useful direction is the inverse problem. Given a target difference, what offset does it take.

Offset = ΔP / (K × F_leak)

At a target of 15 Pa the answer depends entirely on the tightness assumed.

Assumed class	Offset for 15 Pa
Very Tight (1.25)	205 m³/h (121 CFM)
Tight (1.10)	233 m³/h (137 CFM)
Leaky (0.75)	341 m³/h (201 CFM)

The extremes stand in the ratio of the multipliers, 1.25 to 0.75, which is 1.67 to one; the offsets of 341 and 205 m³/h are that ratio after rounding for display. It is the ratio of the multipliers and nothing else. It is not the ratio of the leakage areas of a tight room and a leaky one, which would require choosing an exponent and holding calibration data the model does not carry.

For design this has one practical consequence. The offset is fixed before the tightness is known, and every part of the difference between the assumption and the fact shows up in the pressure that is actually achieved. Capacity margin on the supply and extract systems, enough to move the offset by something like the span of that table, is the usual protection against it.

Per ISO 14644-4: the offset required for a target pressure difference scales inversely with the tightness assumed, and the assumption is tested only when the completed room is measured.

Application Boundaries: Model, Operation, Qualification

The calculator estimates a pressure difference from an airflow offset and an assumed tightness class, for a single boundary, in steady state, with the boundary closed. Nine things fall outside that and need treating separately.

The form of the relation. The model is proportional to the offset while the underlying leakage follows a power law with an exponent commonly near 0.65, so the two coincide only near the offset the coefficient corresponds to. At twice that offset the difference is about a third; at half of it, about 45 percent the other way.

Envelope tightness. It is represented by a four-position selector spanning 0.75 to 1.25, and it is established by measurement on the completed room rather than by selection at design stage.

Open doors. The difference collapses to the order of a thousandth of a pascal when a door opens, and direction is then held by velocity through the opening at something like twenty times the offset, or not at all.

Summation along the chain. The result covers one boundary. Four boundaries at 15 Pa put 60 Pa (0.241 in. w.c.) between the core and the outside, and the plant and the doors both see that total.

Rooms held below their surroundings. The calculator returns zero for a non-positive offset, so containment cases are outside it entirely.

The classification bands. They belong to the calculator. The EU GMP Annex 1 guidance minimum of 10 Pa between adjacent rooms of different grades sits inside the NORMAL band, not at its edge, so a NORMAL result can be below it.

Measurement. The reading depends on tap position, on the state of doors throughout the cascade and on wind and neighbouring systems, so a value without its conditions is not reproducible.

Transient behaviour. Door openings and closings, system changeover and airlock cycling produce excursions and recovery times that a steady-state relation does not describe, and monitoring alarms are frequently set by those transients rather than by the steady value.

Qualification. Conformity is established by the applicable qualification procedure with documented results. A screening estimate made before the room exists does not substitute for it.

Per ISO 14644-4, ISO 14644-3 and EU GMP Annex 1: estimating a pressure difference from an airflow offset and an assumed tightness is the scope of this model, while the form of the leakage relation, envelope tightness, door operation, cascade summation, measurement and qualification each require separate treatment.

Pharmaceutical Cleanroom Cascade Pressure Calculator

Standards and References

• ISO 14644-4:2022, Cleanrooms and Associated Controlled Environments, Part 4: Design, Construction and Start-up (International Organization for Standardization, 2022). The current edition, which replaced the 2001 first edition. Design and construction of cleanrooms and clean air devices, including the arrangement of pressure differences between adjacent spaces and the performance parameters considered at each stage of a project.
• ISO 14644-3:2019, Cleanrooms and Associated Controlled Environments, Part 3: Test Methods (International Organization for Standardization, 2019). The current edition, which replaced the 2005 first edition. Test methods for verifying cleanroom performance, including the measurement of pressure difference between rooms and the conditions under which such measurements are recorded.
• ISO 14644-1:2015, Cleanrooms and Associated Controlled Environments, Part 1: Classification of Air Cleanliness by Particle Concentration (International Organization for Standardization, 2015, reviewed and confirmed 2021). The classification of a cleanroom is set by airborne particle concentration, which is a separate quantity from the pressure difference and is not derivable from it.
• EU GMP Annex 1, Manufacture of Sterile Medicinal Products (European Commission, 2022 revision, published August 2022 and in effect since 25 August 2023). Requirements for sterile manufacture, including the contamination control strategy, the arrangement of airlocks, and the guidance minimum of 10 Pa between adjacent rooms of different grades.
• ASHRAE Handbook, HVAC Applications (American Society of Heating, Refrigerating and Air-Conditioning Engineers, 2023), chapter on clean spaces. Design of cleanroom air systems, the arrangement of a pressure cascade and the relationship between room airflow and the differences it develops.
• ASHRAE Handbook, Fundamentals (ASHRAE, 2025), chapter on ventilation and infiltration. The power-law representation of flow through building envelopes, Q = C ΔPⁿ, and the range within which the pressure exponent of measured envelopes falls. This is the source for the treatment of exponents in this article.
• ISPE Baseline Guide Volume 3: Sterile Product Manufacturing Facilities, Third Edition (International Society for Pharmaceutical Engineering, 2018). Facility design practice for sterile manufacture, including personnel and material flows, airlock arrangements and the treatment of rooms requiring containment as well as cleanliness.
• Manufacturer data for cleanroom doors, airlocks and pass-through devices (current published editions). Seal arrangements including threshold seals, interlock behaviour and cycle times, and the contribution each device makes to the leakage area of the envelope, which is what the four tightness classes in this calculation stand in for.

FAQ

Related Calculators

• Cleanroom Air Change Rate Calculator: the air change rate of the room, which is what the cleanliness classification actually follows from, while the cascade decides only the direction of flow across the boundary.
• Static Pressure Calculator: the pressure the supply unit has to develop, which includes the accumulated cascade along the chain on top of the resistance of the system itself.
• Velocity Pressure Calculator: the velocity pressure behind an air velocity measurement, which is the quantity that holds direction through an open doorway once the cascade has collapsed (article).
• Laboratory Fume Hood Diversity Factor: the adjacent case in which the airflow follows from a protection requirement rather than from a thermal balance, and in which the design assumption is likewise untested by the calculation.
• Hospital Operating Room Airflow Calculator: another controlled environment with its own pressure cascade and its own reasons for holding a room above or below the space next to it.
• CFM Calculator: the airflow of a space in the general case, which is where the supply and extract figures entering this calculation come from.
• Fan Power Calculator: the power drawn by plant working against the accumulated cascade, which is where a decision about the number of boundaries turns into an operating cost.
• Duct Size Calculator: sizing the supply and extract ductwork for airflows that differ deliberately by the offset rather than balancing to the same figure.