Knowledge · Code reading

Industrial code reading: smart reader, PC or line scan?

The right reader architecture starts with the production task: code type, presentation, speed, data retention and the action the line must take after a failed read.

Marked metal part for DPM code reading

Architecture decision

Compare system architectures and optical constraints

Record the code format, smallest bar or cell, reading distance, position range, transport speed and machine interface. Decide whether the same station must also read printed characters (OCR), check presence, measure features or inspect defects. DPM describes how a code is marked on a part; it is an optical constraint that can apply to several system architectures.

DecisionChoose this whenWatch for
Smart code readerA bounded identification station with one or a few fixed views and direct machine I/O.Check optics, embedded tools, protocols, logging and who may change recipes.
PC-based camera and softwareIdentification is combined with OCR, inspection or measurement, or results from multiple cameras must be coordinated.Greater freedom requires software ownership, an update policy, industrial PC infrastructure and diagnostic tools.
DPM with difficult opticsLaser, dot-peen or etched codes on reflective, curved or variable surfaces.Test the mark and lighting geometry first; the DPM guide covers this optical decision.
Line-scan IDContinuous webs, circumferential views of rotating parts or long products assembled line by line.Size encoder triggering, line rate, lighting, data flow and maximum image length together.

Integration and acceptance

Choose by code type, data flow and fault handling

Define the association sequence: the controller creates a cycle ID for the part → the reader accepts its trigger → decoded text and status return with that ID or a validated handshake → the PLC checks expected content → the production record is written and acknowledged → the machine releases or diverts the part. Digital I/O commonly carries status; text needs the supported data interface. Test a no-read, unexpected string, duplicate trigger, late response and unavailable MES. A retry must not create a second production record or attach the old code to the next part. Record attempts, correct identities and invalid outcomes separately, and verify optical sampling on representative codes.

Test the architecture against your line conditions

Send representative codes or parts with the speed range, position variation, required data output and fault-handling procedure. Include reader hardware, optics, lighting, software, integration and operator training in the investment estimate. Compare ongoing maintenance with measured manual scanning time, interventions, stops and rework.