Allot Tech · FAT / SAT
Filling machine FAT: trace the evidence before shipment
A water demonstration can show that a filler moves. It does not prove that your detergent, oil or sauce meets the agreed quantity and output. A factory acceptance test (FAT) should compare the equipment being shipped with a written, testable requirement and preserve the observations behind the decision.
Agree the test boundary before booking the visit
Identify the machine serial number, installed heads, software recipe, supplied accessories and drawing revision. List connected equipment and anything simulated or absent, including upstream product feed and downstream closures. Separate the filler trial from whole-line performance. Agree test duration, sampling, limits and who can approve each result before the machine runs.
Prepare representative product and packaging
For every selected SKU, record the product batch and temperature, target dose, bottle and cap revision, and sample quantity. Include the difficult fill range, narrow neck or foam-prone product rather than only the easiest package. Agree delivery, storage, safe handling and disposal with the receiving party. A substitute liquid needs a written explanation of what it can and cannot establish; keep the untested product requirement open.
Write pass criteria that can actually be measured
| Check | Evidence to keep | Acceptance decision |
|---|---|---|
| Fill quantity | Individual results by head, tare method, instrument and product temperature | Compare each result and the agreed spread with defined limits; the average alone is insufficient |
| Saleable output | Elapsed time, accepted bottles, rejects and every stop | Use the agreed product, package, operators and connected equipment |
| Foam and cutoff | Settling time, overflow, drops and contaminated necks | Define observable limits with the actual liquid |
| Closure, if supplied | Cap version, application settings and agreed closure checks | Confirm the included closure scope; do not assume a filler-only test covers it |
| Changeover and cleaning access | Parts, tools, personnel and time until the next good bottle | Observe the agreed procedure; visual access is not cleaning validation |
| Safety and faults | Qualified staff's approved alarm and interlock checks | Follow the manufacturer procedure; never defeat guards for a demonstration |
Test operating changes, not just steady running
- Observe startup and retain the initial doses instead of removing them from the report.
- Run the agreed sustained trial and record stops as they occur.
- Repeat at a lower product supply level within the approved operating range.
- Stop and restart using the normal recipe; check the first containers again.
- Change to the agreed second format and identify the first acceptable bottle.
Count good containers against elapsed time
Illustrative calculation, not a machine claim: 2,850 acceptable bottles over a complete 30-minute agreed trial equals 95 good bottles/minute. If 150 other bottles are rejected, a counter reading 3,000 does not establish 100 good bottles/minute. Keep the start and finish times and reject reasons. If a contract excludes particular stops, report those exclusions separately so buyers can reproduce the calculation.
Make a remote FAT traceable
Use a continuous view of the identified equipment and a second view of the measurement when needed. Read the SKU, settings and clock into the record; link each sample number to its raw measurement. Keep uncropped result files and a list of interruptions. Video cannot replace measurements or prove components outside the camera and test boundary.
Close deviations before releasing shipment
Give each failure an ID, requirement reference, result, responsible person, correction deadline and retest method. Separate shipment-blocking failures from explicitly accepted open items. Retain both failed and repeated results. The agreed authorized parties should record the release decision; a short successful rerun does not erase the original deviation.
Carry the remaining tests into site acceptance
A site acceptance test (SAT) checks the installed equipment with actual utilities, product feed, conveyors and operators. Carry forward anything not demonstrated at FAT, including integration and sustained site output. Equipment acceptance alone does not validate shelf life, hygiene or destination compliance; those need their own applicable process review.
Link each acceptance decision to its source evidence
Keep a result traceable from the sample and filling head through any deviation and retest to the release decision. A video montage or a single average cannot provide that chain on its own.
Keep individual readings identifiable
Assign sample IDs and record the SKU, filling head, operating phase, units, individual reading, tare method, instrument ID and relevant product conditions. Keep startup, steady running and restart results distinguishable under the agreed plan. Do not pool heads or remove inconvenient values without a documented reason and reviewer.
Define the impact of a change before retesting
After a part, setting or software change, record the old and new revision and the affected requirements and formats. The designated technical owner should approve the retest scope, including related functions when needed. Retain both failed and repeated results. This guide supplies no universal sample count, precision limit or automatic pass rule.
Four links to check in the evidence register
- Substitute material
- Identify the intended liquid and substitute, and which properties the substitute does not represent. Keep unsupported product tests open rather than marking them passed.
- Remote witness
- Link the identified machine, sample label, instrument view and raw file. Agree observation requests within the approved test plan; an unreadable result remains unverified.
- Correction and retest
- Connect the deviation ID to the corrective work, revision, responsible party and new test files. A later passing value does not erase the first result.
- Release and packing
- Match the accepted serial number, configuration and included format parts to the dispatch records. Record the release authority and any agreed open items; a packing list is not a performance certificate.
Questions about FAT records and repeat tests
Can a good average hide one poor filling head?
Yes. Review individual results by head and operating phase against the agreed limits, not only the combined average. Keep the sample order, instrument and method traceable so a reviewer can identify a local issue without inventing or discarding readings.
Must the entire FAT be repeated after a repair?
Not necessarily. The responsible technical parties should assess which requirements, formats and linked functions the repair may affect, then approve repeat coverage. Document what was retested and what remains open; testing one successful bottle does not establish every affected function.
Is a remote FAT video enough to release a machine?
The release decision must follow the agreed acceptance process. Use video with identifiable raw measurements, configuration records and deviation closure evidence. If a file or measurement cannot be checked, record that limitation and agree further evidence instead of assuming a pass.
Download a working acceptance record
Working template: fill in your own limits, sampling plan, results, evidence and reviewer. Blank cells do not mean a test passed. This is a procurement aid, not a certification or a statutory sampling plan.
Questions buyers should settle
Can water be used for the FAT?
It can support agreed mechanical checks. It cannot establish viscosity, foaming, particles, cutoff or cleanability for a different liquid. Document the substitute and carry the unsupported tests into an actual-product trial.
Must every bottle size be tested?
Agree coverage from the actual risk and contract. Include the smallest and largest doses, difficult container geometry and critical product differences. Record any excluded SKU and how it will be verified later; a result for one format is not automatic approval for all formats.