This machine tutorial explains how to operate and troubleshoot building a defect library for 100% web inspection on web inspection and strobe synchronization on converting lines. It is written for shift supervisors, maintenance technicians, and application engineers who need repeatable procedures—not theory alone.
Machine scope and operating context
Yaoshg field teams use this discipline on presses and converting lines built in Wenzhou—from early stack flexo units through CI, gravure, laminating, slitting, bag making, and paper container equipment. The steps below assume normal safety lockout rules, OEM manual limits, and documented substrate specifications for each job.
Automated 100 percent inspection systems are only as good as the defect taxonomy behind them. If classification rules are vague, operators get alarm fatigue and critical defects are buried in noise.
Start with a practical library: register miss, ink void, contamination spot, streak, haze, edge nick, and splice anomaly. Each class should include threshold definitions by size, contrast, and recurrence pattern.
Step-by-step machine procedure
Thresholds must be SKU-aware. A defect that is acceptable on industrial film may be rejectable on premium food packaging where visual consistency and barcode readability are strict.
100% web inspection requires stable strobe or line-scan synchronization. Build a defect library with real images from your press—not generic samples—so false rejects stay low.
Link defect classes to hold tags and splice procedures so downstream knows whether to stop or trim.
Operator shift checklist
- Calibrate strobe frequency to line speed and repeat length.
- Confirm camera exposure for defect type library in use.
- Test false-reject rate on known good and known bad samples.
- Link defect codes to downstream hold-release procedure.
Common defects and corrective adjustments
Link defect events to machine-state data such as speed, tension, and dryer zone temperature. Correlation across these signals converts inspection from passive detection into active process control.
Review false positives and escapes weekly with cross-functional teams. Continuous refinement keeps the library relevant as substrates, inks, and customer quality standards change.
False rejects from glare or wrinkle mimic print defects. Adjust lighting angle and web spreader before lowering detection sensitivity.
Maintenance records and when to call service
Clean optics and verify encoder coupling on PM cycles. Export lines increasingly tie inspection data to customer quality portals—keep clock sync and defect IDs consistent.
If mechanical adjustment, drive parameter changes, or repeated defects exceed on-site scope, log serial number, job recipe, and photos before contacting Yaoshg service. Commissioning engineers can remote-review HMI trends when VPN or data export is available—faster resolution when shift records are complete.