Sachet date-code and seal checks: what vision sees on a VFFS line

Where should the camera look?

A mark inspected before sealing can be damaged later. Define whether the acceptance requirement concerns the printed film at the camera or the finished pack after all handling.

How should the date and batch check work?

Read the code and compare it with the approved recipe or production data. Reject a missing, unreadable or incorrect string under the agreed rules. Include film shine, print movement and wrinkling in the sample test. A no-read caused by a fold is still an inspection failure to manage; it must not be silently treated as a correct code.

Keep line clearance and first-off approval at reel and product changes. FSSC 22000’s version 6 controls for labels and packaging are relevant to the changeover procedure; they do not make a camera a substitute for approved artwork and master data.

Can vision detect powder in a sachet seal?

It can detect contamination that produces a visible difference in the imaged seal region. Opaque or metallised films can hide material inside the seam, and a weak seal may look normal. Use representative contaminated, wrinkled and acceptable samples to establish what the camera actually distinguishes. Seal strength and leak integrity need the appropriate additional process or test method.

Do not label every unusual seam image as a proven leak. Record the observed defect and compare it with the agreed quality specification and, where needed, the functional test.

What should the reject test cover?

• Normal speed, speed changes, stops and restarts.
• Touching or unusually spaced sachets where the line allows them.
• A deliberate bad code and a visible seal fault, followed to the actual reject point.
• Reject confirmation, a full reject bin and the response to a lost trigger or missing result.

Pouch inspection explains the complete station. PLC and reject integration explains tracking. If underweight packs are also a concern, specify the weighing task separately; a surface view does not measure net mass.

Separate print, forming and sealing faults

A sachet can leave the coder with a readable mark and still become unacceptable after forming or cutting. The web can wander, the cut position can change, and a fold can hide characters. Sealing adds another independent requirement: the seal region must be formed and free of the visible faults defined by quality. Checking only a printed web upstream does not establish the condition of the final individual sachet.

For the date code, distinguish ink deposition from image formation. Missing dots or ribbon voids originate in printing; glare and motion blur can make intact print appear missing in the camera. Save an original image before applying contrast filters and inspect the physical sachet. If a code looks correct when stationary but fails during movement, examine exposure and trigger position before changing the expected-character rule.

HIKROBOT’s lighting glossary describes ways to control reflection on shiny material. A diffuse arrangement can improve print visibility while another geometry may be needed to expose a wrinkle in the seal band. (HIKROBOT Light Source Catalog, 2025 Q3, glossary) Do not assume the image best suited to OCR is also the best image for every seal defect. Define each required feature and trial the views together.

Treat every lane and finished pack as an identifiable item

A multi-lane machine needs an explicit relationship between an image region, a lane and a finished sachet. If the web is inspected before cutting, the integration must preserve that relationship through the cut and transfer. If it is inspected afterwards, presentation must keep the code and seal visible without overlap. A lane with a fault should not disappear into an overall pass rate dominated by the other lanes.

In South African food production, the dated R146 text supplies the batch and date-marking context, while R638 addresses food-premises hygiene. (South African food labelling regulations R146, published 1 March 2010) (South Africa: food premises and transport hygiene regulations R638, 2018) Neither sets a universal camera exposure or sample size. Keep the approved artwork, date logic and cleaning procedure under site control; the camera recipe implements those requirements rather than inventing them.

A hypothetical challenge after a reel change

Hypothetical example: a new film reel has the correct artwork, but its print-to-cut registration differs from the previous reel. A code-presence tool still finds dark characters, while the final cut removes part of the date. Test a complete finished sachet, a sachet with the code crossing the cut boundary and a sachet with an obsolete but readable date. These exercise position, completeness and content as separate decisions.

Also include an acceptable wrinkle outside the functional seal band and a visible contaminant inside it. The first should follow the agreed cosmetic specification; the second should follow the defined seal-region fault policy. This prevents an overly broad wrinkle detector from rejecting harmless packaging variation while giving a misleading impression of seal-integrity coverage. The example is a challenge design, not a measured line result.

Freeze the recipe after the trial, then run reserved samples under normal operation, restart conditions and the permitted speed range. Confirm the physical sachet reaches the intended reject path. If several sachets are held because individual tracking is not possible, document that containment scope in the acceptance criteria. Do not claim individual rejection unless it has actually been demonstrated.

Keep appearance evidence separate from integrity evidence

A visible seal-band check can identify product, folds or incomplete-looking regions that the selected image reveals. It cannot prove that every seal withstands the required load or is leak-tight. ASTM F88/F88M-23 scope 1.1 addresses seal strength, while ASTM F2096 addresses gross leak detection using internal pressurisation. (ASTM F88/F88M-23: seal strength of flexible barrier materials) (ASTM F2096-11(2019): gross leaks by internal pressurisation (bubble test)) The public scope descriptions do not supply an acceptance force, pressure or leak threshold for your sachet. Select and validate the relevant package test with the responsible packaging team.

Correlate image findings with the selected integrity tests using identified samples. Some visible faults may consistently fail the physical test, but that relationship must be demonstrated for the material and process in scope. A change in laminate, sealing conditions or product can alter it. Keep unexplained disagreement samples for investigation rather than redefining the vision threshold until the results appear to agree.

For operational reporting, retain separate counters for unreadable code, wrong code, misplaced code, visible seal fault, tracking failure and failed removal confirmation. Assign each to an owner: printing, packaging setup, quality data or integration. Report false accepts and false rejects against their relevant test populations, without presenting a successful finite trial as zero risk. NIST’s proportion guidance explains that statistical limit. (NIST e-Handbook, 7.2.4.1 Confidence intervals for a proportion)

The pouch-inspection solution covers the station scope and the PLC/reject guide covers product tracking. This article focuses on why sachet forming changes the image and how to preserve the distinction between correct information, acceptable visible packaging and independently validated package integrity.

Define the pouch station and its inspection limits

Use the sachet forming process and representative packs to decide where each check belongs, then agree which faults need a separate integrity test.