First Decision: Which Method Proves the Bond?
Several methods get described as “seal inspection”. They answer different questions, so be clear about which question you need answered for every bottle.
| Method | What it actually checks | Coverage | Where it fits |
|---|---|---|---|
| Inline thermal imaging (e.g. ISIVS) | The heat pattern of the foil after sealing, used to infer bond quality, liner presence and cap seating | 100% of bottles, non-destructive | Directly after the induction sealer |
| Upstream camera / cap inspection | Liner present in the cap before capping; cap present, height and skew after capping | 100%, but no bond information | Cap feed and after the capper. Complements thermal inspection, doesn't replace it |
| Offline destructive tests (peel, torque, pressure/leak) | The physical bond strength or leak tightness of the samples tested | Sampling only; the samples tested are destroyed | Process qualification, periodic audits, validation correlation |
| Manual visual check | Cap appearance only | Sampled or 100%, but blind to the seal | Not a seal check |
The physics is explained in why thermal imaging catches induction seal failures. In short: a properly bonded foil leaves a characteristic heat signature. A missing liner, a loose cap, too little or too much heat, or a cut liner each leave a different one.
Second Decision: Can It Keep Up With Your Line?
Do the arithmetic before you look at brochures:
| Line speed | Time per bottle | Bottles per 8-hour shift |
|---|---|---|
| 60 BPM | 1.0 s | 28,800 |
| 90 BPM | 0.67 s | 43,200 |
| 120 BPM | 0.5 s | 57,600 |
In that time the system must capture the image, classify it and fire the reject. Ask vendors for speed figures on your bottle diameter, cap size and conveyor pitch. Optomech publishes the ISIVS as handling up to 120 BPM, depending on the application, and we confirm the figure for each line.
Third Decision: Placement and Line Layout
- Close to the sealer exit. Residual heat fades within seconds, so the inspection point has to be near the induction tunnel. The vendor should specify the distance for your line speed.
- Stable bottle handling. Bottles should be upright and well spaced at the inspection point. Wobble and touching bottles reduce reliability.
- Room for rejection. Allow space for the ejector, a lockable reject bin and reject confirmation before the next machine.
- Sealer signals. Consider linking sealer faults and stoppages to the inspection system, so that bottles left under the coil during a stop are handled correctly.
Fourth Decision: Defect Classes and What They Tell You
A system that only says “bad seal” helps less than one that tells you why. Classified defects point straight to the upstream machine that is drifting:
| Defect class | Likely upstream cause | Where to look |
|---|---|---|
| No liner | Liner missing in the cap supply | Cap supplier and cap feed; consider upstream liner inspection |
| Loose cap | Capping torque or cap seating | Capper chucks and torque settings |
| Under heating | Too little induction energy or too much speed through the coil | Sealer power, coil height, conveyor speed |
| Over heating | Too much energy or a bottle stopped under the coil | Sealer power, line stoppages |
| Cut liner / moon cut | Liner damage from the cap or capping | Cap quality, capper tooling |
These are the six classes the Optomech ISIVS is published to recognise. A rising trend in one class, for example under heating, is an early warning to correct the sealer before rejects pile up.
Fifth Decision: Rejection, Records and Data Integrity
- Reject mechanism suited to the bottle (Optomech ISIVS uses 2-way pneumatic ejection)
- Bin-full sensor with a line stop before the bin overflows
- Defined behaviour on camera or system fault: stop or reject-all
- Reject images stored with the defect class and a timestamp
- Per-batch and per-shift reports (ISIVS: PDF and Excel)
- Multi-level user access and an audit trail for parameter changes
Validation: How to Prove It Works
For a GMP line, plan the qualification before installation:
- Challenge samples. Make controlled defects for each class: caps without liners, under-torqued caps, bottles sealed at reduced and increased power, and liners with deliberate cuts.
- Correlation. Compare thermal classification with destructive tests (peel or leak) on the same bottles. This links the thermal decision to physical seal integrity.
- Speed range. Run challenges at minimum and maximum line speed.
- Routine challenge. Define a challenge test at set points (start of batch, after changeover, after sealer adjustments) and record the results.
Optomech states that IQ/OQ/PQ documentation is supplied with ISIVS. We agree the exact scope in the quotation. For the general approach, see our IQ/OQ/PQ guide and the revised Schedule M checklist. For US OTC products, induction seals are commonly used as a tamper-evident feature under 21 CFR 211.132. Confirm the requirements that apply to your product with regulatory affairs.
Questions to Ask Every Vendor
- What detection method do you use, and what does it physically measure?
- What speed can you commit to on our bottle and cap, and with which reject method?
- Where will the station sit relative to our sealer, and what happens during a sealer stoppage?
- Which defect classes are reported separately?
- How will detection be proven: challenge samples, correlation with destructive tests, documentation?
- What are the service response, spares and calibration arrangements?
For broader vendor evaluation, use our 15 RFQ questions, and see the vision inspection price guide for Optomech starting-from price bands.