Filling Line FAT Acceptance Protocol for Pharmaceutical, Veterinary and High-Viscosity Products
Executive answer
A filling-line FAT should prove the complete operating chain against an approved user requirement specification—not merely show that the machine runs. For pharmaceutical, veterinary and high-viscosity products, the protocol should challenge dose accuracy across the specified fill range, product behavior at representative viscosity and temperature, container and closure variation, cleaning and changeover steps, recipe security, alarms, rejection logic and the records needed for traceable release. Shipment should depend on objective evidence and an agreed punch-list classification.
The practical rule is simple: every critical requirement needs a test method, an acceptance limit, a data record and a named disposition. If a requirement cannot be demonstrated at the supplier’s factory, the protocol should state why, identify the substitute evidence and assign the remaining test to site acceptance or qualification.
For equipment context, review King Pack’s filling machine solutions, the engineering discussion of why FAT matters more than brochure speed, and a documented customized filling-line FAT example.
1. Quick FAT map: from URS to shipment release
| Gate | Question to answer | Primary evidence |
|---|---|---|
| Readiness | Are approved requirements, samples and utilities available? | URS traceability matrix, test plan, calibration status |
| Completeness | Is the contracted machine and documentation set present? | Configuration check, drawings, manuals, certificates |
| Performance | Can the line repeatedly fill and close the defined product/package range? | Raw measurements, calculations, reject records |
| Control | Do recipes, alarms, access levels and recovery logic behave as specified? | Challenge scripts, audit-ready screenshots, event records |
| Hygiene | Can the product path be drained, cleaned and changed over as designed? | Coverage observations, rinse evidence, timed changeover |
| Release | Are deviations closed or formally accepted with owners and dates? | Punch list, approved FAT report, shipment authorization |
FAT is one stage in a lifecycle. FDA process-validation guidance describes process validation as collecting and evaluating data from process design through commercial production, while EU GMP Annex 15 frames qualification and validation as planned, documented activities. A useful FAT therefore creates evidence that can be reused—but not misrepresented—as input to installation, operational and performance qualification.
2. Freeze the inputs before the test
A strong protocol begins weeks before witnesses arrive. The supplier and user should approve a test boundary: machine modules, formats, fill volumes, product simulants or actual product, nominal output, utilities, environmental assumptions and interfaces excluded from the factory setup. The latest approved URS and functional design specification should be cross-referenced line by line.
- Product data: density, viscosity range and measurement conditions, shear sensitivity, foaming tendency, particulate content, temperature window and acceptable hold time.
- Package data: minimum/nominal/maximum dimensions, tolerance drawings, closure specification, supplier lots and known cosmetic or functional defects.
- Process targets: fill range, accuracy basis, speed by format, reject philosophy, allowable stoppage, product recovery and batch reconciliation expectations.
- Quality limits: sampling plan, calculation method, measurement-system resolution, pass/fail criteria and deviation handling.
- Utilities and interfaces: air quality and pressure, electrical supply, vacuum, water, extraction, upstream/downstream handshakes and data connections.
Do not let the test start with phrases such as “approximately accurate” or “runs well.” Acceptance criteria should be numeric where possible and unambiguous where qualitative. Record whether fill accuracy is assessed by mass or volume and how density correction is handled.
3. Documentation and configuration review
Before wet testing, verify that the physical machine matches the purchased configuration. Record nameplates, contact-material grades, surface-finish certificates where contracted, elastomer lists, motors, sensors, guards, safety devices and change parts. Confirm that drawings reflect the as-built system and that measuring instruments used during FAT have current calibration evidence traceable to the supplier’s quality system.
The document review should include operating and maintenance manuals, electrical and pneumatic drawings, software and recipe version identifiers, spare-parts lists, recommended consumables, lubrication schedules, certificates, weld or passivation records when in scope, and draft IQ/OQ documentation when contracted. Document gaps should enter the punch list; verbal promises should not disappear after the factory visit.
4. Verify the complete product path
Trace product from the supply or holding vessel to every filling nozzle. Confirm dead legs, low points, drainability, hose routing, clamps, seals, filters, pumps, manifolds, recirculation and temperature-control points. For pharmaceutical or veterinary products, the line should support the user’s contamination-control and cleaning strategy. For viscous products, avoid assuming that a water run proves performance: water does not reproduce pressure loss, stringing, air entrapment or refill behavior.
Use a representative product or justified simulant. If actual product is unavailable, document the matching rationale using measurable attributes such as viscosity at defined shear rate and temperature, density, surface tension, particle behavior and foaming. Record batch preparation and any changes made during testing.
5. Challenge dosage accuracy across the operating range
A defensible accuracy study uses predefined samples at minimum, nominal and maximum fill volumes, with start-up, steady-state and restart conditions. Include every filling head. Weigh empty and filled containers using a suitable balance, account for tare and density when converting mass to volume, and report individual results rather than only an average. An average can hide one biased nozzle.
| Condition | Why it matters | Suggested record |
|---|---|---|
| Minimum fill | Small errors become a larger percentage of dose | Individual results by nozzle; mean, range and deviation |
| Nominal fill | Represents routine production setpoint | Consecutive samples at target speed |
| Maximum fill | Challenges pump capacity and refill time | Cycle stability and overflow observations |
| Start/restart | Air and pressure may redistribute after stops | First-off samples after defined dwell times |
| Low hopper level | Tests inlet consistency and air pickup risk | Level, pressure, weights and alarm response |
| Speed challenge | Confirms the quoted rate does not degrade control | Accepted units per minute with rejects and downtime |
Agree how outliers are handled before the run. Repeating a failed sample until it passes destroys the value of the evidence. A deviation should be logged, investigated, corrected when appropriate, and followed by a defined retest that preserves the original data.
6. Test viscosity, bubbles and cutoff—not just volume
High-viscosity products create interacting risks: incomplete cylinder refill, pressure variation, product heating, nozzle stringing, tailing, splashing and air pockets. Challenge the specified viscosity and temperature envelope. Observe pump inlet conditions, refill consistency, nozzle closing, suck-back settings, container presentation and product appearance after filling.
For aeration-prone creams or suspensions, assess whether mixing, transfer and pumping introduce bubbles and whether the holding system provides a controlled path for deaeration. A clean cutoff test should define acceptable drip, thread or smear behavior and evaluate the downstream consequence—not only the nozzle. Product on a tube seal area, bottle neck or closure interface can cause later failures.
7. Challenge container and closure variation
Run approved samples representing dimensional extremes and, where practical, multiple component lots. Confirm feeding, orientation, detection, filling position, closure application, reject logic and recovery after a jam. If the factory cannot obtain statistically representative extremes, document the limitation and reserve the challenge for SAT or qualification.
The FAT should not treat a container as an isolated drawing. Verify the system: container, closure, product, change parts and transport logic at the intended speed. Record scuffing, deformation, tipping, cross-threading, poor seals, cap torque variation and false rejects.
8. Cleaning, changeover and recipe control
A factory test cannot validate the user’s final cleaning process, but it can verify that the equipment design and automated sequence are capable of executing the specified strategy. Review access, drainage, removable parts, cleaning connections, sequence parameters and operator prompts. When CIP is contracted, challenge sequence steps, valve states, flow or pressure monitoring, temperature inputs, alarms and report generation with a justified test medium.
Time a representative format change from last good unit to first approved unit. Include line clearance, change-part identification, tool needs, adjustments, recipe selection and verification. Recipe tests should confirm authorized creation and editing, protected critical parameters, correct format recall, version identification and behavior after power loss.
9. Alarms, interlocks and records
Test safety and quality-related alarms deliberately: guard opening, emergency stop, loss of air, low product level, missing container, no-container/no-fill, missing closure, jam, reject-bin full, sensor failure where simulatable, and communication loss at interfaces. Record the trigger, annunciation, machine response, reset conditions and product disposition.
Where electronic records are in scope, verify user roles, password rules defined by the specification, event or audit records, time synchronization, backup/restore behavior, recipe traceability and data export. FAT does not itself establish full data-integrity compliance, but it should prove the contracted functions and expose gaps before shipment.
10. Punch-list control and shipment release
| Class | Typical meaning | Release treatment |
|---|---|---|
| A — Critical | Safety, product-quality or core contractual function not acceptable | Shipment blocked until closed and witnessed or formally re-approved |
| B — Major | Required function incomplete but controlled correction is possible | Closure evidence and owner/date required; shipment by approved exception only |
| C — Minor | Documentation, finish or low-risk item with no immediate performance effect | May remain open with dated action and verification plan |
The final report should identify equipment and software versions, participants, calibrated instruments, materials, executed procedures, raw data, deviations, retests, photographs or screenshots, punch-list status and signatures. Release is a governance decision: an unfinished issue should never vanish because the machine is needed urgently on site.
Supplier evaluation: questions that reveal FAT maturity
- Can the supplier map each URS requirement to design evidence, FAT, SAT or qualification without leaving orphan requirements?
- Will representative product and packaging be run at the specified operating envelope, not only at an easy demonstration point?
- Are raw data, failed runs and corrective actions retained in the final evidence package?
- Can the supplier explain what the FAT does not prove and how remaining risks transfer to site?
- Who owns software changes after FAT, and how are versions, backups and retest needs controlled?
FAQ
Is FAT the same as qualification?
No. FAT is supplier-site acceptance evidence against the purchase specification. It can support later qualification, but site installation, utilities, procedures, trained operators and actual production conditions still require their own approved verification.
Must actual product be used?
Use actual product when feasible and safe. Otherwise choose a justified simulant based on relevant physical behavior and document limitations and follow-up testing.
How many samples are enough for accuracy?
There is no universal number. The sampling plan should reflect fill range, nozzle count, variability, risk, measurement capability and the purpose of the test. Approve it before execution.
Should the machine run at maximum speed?
Test the contracted production rate for each relevant format and product condition. A mechanical dry-cycle maximum is not equivalent to accepted filled output.
Can open punch-list items ship?
Only under an approved risk-based exception that defines classification, owner, due date, evidence and the point of verification. Critical unresolved issues should block release.
What should be sent to King Pack before protocol drafting?
Send the URS, product properties, samples and drawings, fill range, target output, acceptance limits, utility data, interface scope and validation expectations.