Frequently Asked Questions

Answers on acoustic vs machine vision inspection, noisy workshop environments, modeling without defective samples, MES/PLC integration, thresholds and data security.

Top questions

Q1. How is acoustic inspection different from machine vision?

Machine vision only sees external surfaces. Acoustic inspection detects internal and dynamic faults — loose assembly, missing components, damaged gears and bearings — through the sound a unit radiates while running. The two are complementary, not interchangeable; for interior defects that emit sound, acoustics is often the only practical online method.

Q2. Our workshop is noisy (80–90 dB(A)). Does acoustic inspection still work?

Yes, with proper engineering: a fixed acoustic enclosure or sound-absorbing station design combined with signal gating. Feasibility is confirmed by an on-site survey — we do not promise that "it works everywhere".

Q3. We have no defective samples. Can we still build a model?

Yes. ISS-ACS-100 uses One-Class learning: the model learns a baseline from 20–30 confirmed good units and flags deviations from it. No defective samples are required to go live.

Q4. Can it integrate with our PLC or MES?

Yes. The station outputs OK/NG verdict signals to the production-line PLC for automated sorting; MES/HTTP reporting is available for traceability.

Q5. How is the pass/fail threshold determined?

Thresholds are calibrated on your own good-unit samples during commissioning and verified on site. Suspected units are routed to manual recheck instead of being auto-released.

Q6. Does our audio data leave the factory?

No. Inference runs locally on the edge unit and audio data stays in your workshop.

All 96 FAQs (Chinese) →

ISSAUTO has no affiliation with issautomotive.com (North America). Resonance is a market competitor only.