Installing cleanroom differential pressure monitoring is more than putting room-pressure values on a screen. A factory must know which room boundary it is protecting, the intended direction of airflow, how a sensor measures the relationship, and what evidence will remain when something goes wrong. This guide helps procurement, facilities and quality teams in Thailand define the scope, select sensors, specify alarms and records, compare bids, and accept the system through FAT and SAT. Values in pharmaceutical guidance need to be read in their own context; they are not universal Thai factory limits.
Start with the purpose of each pressure boundary
Differential pressure is the pressure difference between two spaces. A product-protection cleanroom may need air to move from a cleaner area toward a less clean one. A containment process may need negative pressure so hazardous dust does not escape. Neither “positive” nor “negative” is right without a defined hazard and operating mode. Record whether the system protects the product, the worker, the surrounding area, or several of these.
Draw a line across every critical boundary on the floor plan: packaging room to anteroom, anteroom to corridor, weighing room to adjacent room, and so on. For each pair, state the normal direction, door and pass-through use, HVAC mode and response to supply or exhaust failure. A long list of room names is insufficient if the relationships between rooms are not measured.
Vaisala’s technical article discusses low cleanroom pressure differences, direct room-to-room measurement, sensor uncertainty, short door transients and defensible records. Its examples of 5, 10 and 15 Pa illustrate small magnitudes; they are not universal setpoints or proof that a particular instrument fits every facility.
Pressure is not a cleanliness classification
A normal pressure reading cannot prove that a filter is intact, that air does not short-circuit around equipment, or that a person followed the intended route. Conversely, one short drop when a door opens does not by itself establish product impact. Interpret differential pressure alongside particles, airflow visualization, temperature and humidity where relevant, and process records.
ISO 14644-2:2015 concerns a risk-based monitoring plan that provides evidence of cleanroom performance related to airborne particle concentration. Its public contents include pressure-differential considerations, calibration and deviation response. It does not establish one pressure setpoint for every cleanroom. TSI’s explanation likewise distinguishes the particle concentration used for ISO 14644-1 classification from supporting environmental parameters.
Survey the room pressure cascade before selecting a sensor
Overlay the plan with clean and less clean sides, expected airflow arrows, doors, pass-throughs, supply and exhaust points, AHU zones and reduced-operation modes. Walk the line while production, material transfer, cleaning and shift changes occur. A door may remain open for a trolley, and two airlock doors may be opened together in practice even though the drawing assumes otherwise. These facts affect both setpoints and alarm timing.
Separate the value used for HVAC control from the value retained as monitoring evidence. A building management system (BMS) can control fans and dampers while an environmental monitoring system (EMS) records critical conditions. The functions can also coexist on one platform if measurement identity, calibration, data integrity, alarm ownership and access are demonstrable. Vaisala describes this distinction as design guidance, not a legal requirement for two separate boxes.
Collect short trends under identified modes: normal production, doors opening and closing, cleaning, shutdown, startup and filter service. Record the time, reference instrument, door state and HVAC mode. Do not treat an uncalibrated existing gauge as the ground truth. A mixed average across incompatible operating modes can lead to nuisance alarms or hide a sustained loss.
| Site fact | What to observe | Requirement it informs |
|---|---|---|
| Room boundary | Protected side, adjacent room, opening | Measurement point and sign |
| Operating mode | Run, idle, cleaning, startup | Mode-specific decision logic |
| Door and transfers | Frequency and duration | Alarm delay and recovery |
| HVAC | Supply, exhaust and failure response | BMS/EMS interface |
| Existing instruments | Location, tubing, calibration | Reuse or replacement |
| Records | Storage, clocks, access, audit trail | Data acceptance tests |
Select differential pressure sensors for the actual application
Choose the measuring range around observed normal and abnormal conditions. A wide maximum range alone is not an advantage when the decision concerns a small difference. Ask how zero drift, resolution, stated accuracy and uncertainty apply across the intended span. Specify calibration access and interval, rather than copying an unsupported accuracy number from another installation.
Direct measurement uses one transmitter with its high and low ports connected to opposite sides of a boundary. An alternative subtracts two room measurements taken against a common reference. The latter may fit an existing control system, but uncertainty from both instruments, sample-time mismatch and reference instability must be considered. Vaisala’s application-note page identifies low-range selection, direct measurement, tube routing and mounting location as design topics; its public summary is not a numerical specification.
Include ports and tubing in the instrument specification
Reversed high and low connections reverse the sign. Long, kinked, crushed, leaking, obstructed or wet tubing can corrupt a sound sensor. Specify the port locations, tube material and route, wall penetration finish, cleanability, labels and maintenance access. Require connection drawings, as-built photographs before concealment, calibration instructions and a post-replacement sign check.
The monitoring system must distinguish a true low reading from a missing reading. Include loss of power, broken wire, failed communication, fixed or out-of-range value and overdue calibration in the design review. State the unit, sign, timestamp and quality flag in the data interface so future room additions remain comparable.
Specify a record that can support a decision
A dashboard line is not automatically an audit-ready record. Define sampling, retention, clock synchronization, roles, audit trail, backup, restoration and export in the user requirement specification (URS) where the quality system needs them. Set retention according to the product and contractual requirements, not a generic number from a vendor brochure.
During a communication outage, can a gateway hold readings locally? On recovery, can it import them without duplicates or reversed time order? Can an investigator identify sensor failure, network failure and power loss separately? Test acknowledgement, shelving, reset and configuration changes with user, time and reason. If cloud storage is proposed, the factory’s IT and OT owners should approve connectivity, data ownership and offline procedures.
Our article on industrial data loggers versus factory IoT explains the broader collection and storage choices. Pressure monitoring adds boundary identity, sign, calibration and alarm ownership. The OT asset inventory guide helps link instrument IDs, network details and maintenance responsibility to calibration and replacement history.
Design alarms around transients and sustained excursions
A door opening causes a brief pressure change. If every transient sends the same urgent notification, operators may stop responding. A door left open, failed air handling, slow recovery or persistent reversal calls for a different response. Define alert and action conditions per boundary: threshold, duration, recovery, priority, recipient and when product-impact assessment begins. Derive delays from observed operations and risk; do not apply one timer to every room.
Retain short transients in the trend even if they do not trigger a notification. An acknowledgement must not erase an ongoing abnormal condition. Record onset, duration, acknowledgement, recovery and recurrence. Decide who receives alarms outside working hours, whether Thai-language instructions are needed and what to do if the communications channel fails.
EU GMP Annex 1, sections 4.14–4.16 addresses sterile medicinal products. It calls for critical pressure differences to be continuously monitored and recorded, and alarm delays to be assessed and justified within the contamination control strategy. Its 10 Pa between adjoining rooms of different grades is expressly a guidance value in that context, with containment exceptions. It is not a blanket Thai legal limit for all factories. The same annex also addresses airflow visualization, so a pressure number alone is not proof of airflow behavior.
| Condition | Define before purchase | Test at acceptance |
|---|---|---|
| Brief door opening | Recording versus notification | Trend remains; nuisance alerts controlled |
| Sustained loss | Limit, delay, owner | Notification and history after duration |
| Reversal | Priority and response | Correct sign and direction |
| Sensor failure | Missing-data indication | Never mistaken for normal zero |
| Recovery | Automatic or approved reset | History and cause remain available |
Translate standards into contract requirements
A line in an RFP saying “ISO compliant” does not align competing bids. ISO 14644-2 supplies a monitoring-plan framework related to particle cleanliness; access to the full standard and the factory’s quality system is needed to decide its application. WHO TRS 1010 Annex 8 concerns HVAC for non-sterile pharmaceutical products. EU GMP Annex 1 concerns sterile medicinal products. Electronics, food and general assembly facilities have different product and customer requirements. None of these international documents by itself proves compliance with Thai law.
Make a matrix separating background references from binding project acceptance criteria. Ask the factory’s quality and legal owners to confirm sector-specific local obligations. For every contractual requirement, ask bidders which drawing, test and record demonstrates it. Do not copy a pharmaceutical pressure guidance value into an unrelated negative-pressure containment room.
The point list should identify the two rooms, sign convention, modes, measurement and recording interval, display, alarms, interlocks and maintenance owner. Rank critical boundaries by risk; decide whether redundancy or backup power is justified. Uniform specifications across all boundaries can create both wasted cost and underperforming critical measurements.
Write a testable URS and comparable RFP
The URS describes what the owner must be able to demonstrate, not the features on one supplier’s data sheet. Give each requirement an ID and map it to FAT or SAT. Define who will establish any initially unknown limit and when; leaving a blank threshold in a signed order invites later disagreement. Include the measurement range and sign, uncertainty, calibration, response, fault handling, port and tubing construction, BMS/EMS interface, alarm logic, record controls and maintenance.
| Example URS ID | Acceptance requirement | Test example |
|---|---|---|
| DP-01 | Display and save a named room-to-room value with defined sign | Port check and applied test pressure |
| DP-02 | Show missing signal separately from a valid zero | Safe input-loss simulation |
| AL-01 | Notify and log sustained loss | Threshold, delay and reset simulation |
| DATA-01 | Track actor, time and before/after setting | Role and audit-trail test |
| OPS-01 | Apply defined HVAC-mode logic | Mode-change test |
| MAINT-01 | Verify sign after calibration or replacement | Record and on-site comparison |
Use one response template for all bidders. Separate transmitters, ports and tubing, panels, power, network, BMS/EMS integration, licenses, installation, clean-area controls, FAT/SAT, training, calibration and support. Identify work that requires a production stop, existing equipment changes, and what the factory supplies. Demand the room-boundary drawing, instrument and tube layout, I/O and data flow, alarm list, role matrix, screen examples, proposed tests, calibration plan and spares. Compare assumptions and exclusions side by side.
If the data will be a quality record, include software versions, restoration test, controlled configuration changes, account removal, synchronized time and access to audit trails. The needed level depends on the industry and the facility’s quality system. Explain the purpose of each requirement instead of imposing every pharmaceutical feature on a general electronics factory.
FAT: test logic before the system reaches the site
Factory acceptance testing can use safe simulated pressure and inputs to cover normal, low, reversed, out-of-range, disconnected, communication-lost and recovered states. Link calibration certificates, instrument IDs, configuration and screen versions to the same test record. If software changes before site installation, assess and repeat affected tests.
For alarms, test just before and after the delay, at recovery, and during repeated short excursions. Check whether an acknowledged but still active fault remains visible; whether a second event is logged; and whether a failed notification is reported. Never force a dangerous physical condition merely to test software. Record test conditions, reference instruments, logs, screenshots, deviations and retest dates. Carry unresolved FAT items explicitly into SAT.
SAT: verify actual rooms, doors and HVAC operation
At site, trace each pressure port through tubing to its transmitter. Confirm the room name on the display, high/low side, wall finish, cleanability and measurement against a calibrated reference under the same door and HVAC conditions. If values differ, check the port, tubing, sensor, calculation and time alignment before applying an arbitrary display offset.
Trend actual door cycles, material movement, cleaning and HVAC mode changes. Verify that the selected delay does not create constant nuisance alarms or hide a sustained loss. Test door recovery, AHU stop, network loss, power return and maintenance replacement using procedures agreed with facilities and quality teams. Activities that could affect manufacturing require an approved time, isolation and restoration plan.
The handover package should include accepted tests, open-item deadlines, calibration certificates, as-built drawings, point and tube IDs, backups, administrator access, training and maintenance instructions. Review early operation for alarm counts, causes, response and data gaps. A lower alarm count achieved only by weakening limits is not evidence that the pressure cascade is healthy.
Clarify responsibility in a Thai factory project
HVAC creates the pressure relationship; instrumentation measures it; the monitoring platform stores and communicates it; quality decides product impact. Even with a single prime contractor, name the owner of each interface and each acceptance test. Assign separate owners for correcting HVAC, evaluating product, and retaining records.
The RFP should state installation windows, wall penetrations, cleanroom access, cleaning and any requalification after work. Include Thai-language operating and training needs, spare-parts availability and local calibration support. For an existing BMS, define point names, units, sign, update rate, stale-value behavior, clocks and change control. Have the factory approve any direct cloud connection under its OT rules before installation.
Frequently asked questions
What pressure setpoint should a Thai cleanroom use?
There is no defensible universal value. Choose from the protection objective, room and HVAC design, operating modes, airflow verification, applicable obligations and customer specifications. EU GMP Annex 1’s 10 Pa guidance value belongs to sterile pharmaceuticals and is not a general Thai legal limit.
Is direct room-to-room measurement mandatory?
Not for every boundary. It is often easier to interpret for a critical relationship and avoids subtracting two independently uncertain readings. Calculated pressure may suit an existing control arrangement. Compare uncertainty, time alignment, tubing, calibration, faults and intended use of the records.
Is a BMS screen enough for monitoring?
Only if the system demonstrably meets the needed recording, alarm, access, clock, backup and investigation requirements. A separate EMS may be useful, but the requirement is the function and evidence, not two separate product boxes.
Does alarm delay hide a problem?
It should delay notification of a short normal transient, not discard the reading. Justify the delay for critical boundaries and test door transients and sustained failure. A longer delay is not automatically safer.
How do FAT and SAT differ?
FAT tests sensor, control, alarm and record logic with safe simulation before shipping. SAT checks real ports, room direction, doors, HVAC and notification at the site. Trace both to the same URS IDs.
Conclusion
A useful pressure-monitoring purchase starts with room boundaries and protection goals. Survey the pressure cascade, define sensor and tubing performance, separate control from defensible records, make alarms fit real operations, and give each URS requirement a FAT or SAT acceptance test. Read international guidance within its stated product scope and confirm Thai and customer requirements for the specific facility.
If your Thailand factory is mapping pressure points, connecting an existing BMS or drafting URS and FAT/SAT criteria, you can contact TOMAS TECH with the current plan and operating constraints. An initial discussion can begin before the equipment order.
