Follow our business 👇 Scan to subscribe
Pitfall 1: Signal Reversal – The Basic Mistake of Turning “OK” into “NG”
Over 200 defective screens flowed into the next process, with rework losses exceeding 50,000 yuan. Whether it is micron-level defect detection in 3C electronics or dimensional measurement of components in the automotive industry, the rationality of hardware selection directly affects the final detection results. An electronic factory’s mobile phone screen inspection line once caused an accident due to signal reversal: the vision system detected defects and output an “NG” signal, but the PLC judged it as “OK”, resulting in over 200 defective screens flowing into the next process, with rework losses exceeding 50,000 yuan. The root cause was that the vision side assumed “high level = OK”, while the PLC programmer set “high level = NG” based on experience, and both parties connected the wiring without written confirmation, directly leading to a logical misalignment of signals.
This type of error can lead to undetected defective products, production line downtime for screening, increasing costs and damaging reputation. To avoid this, two steps are necessary: first, establish a “Signal Definition Confirmation Form” before debugging, clarifying core logic such as “DO1=OK, DO2=NG”, and have both parties sign and archive it; second, add comments in the PLC program for the vision signal module, indicating the meaning of the signals and corresponding actions to avoid confusion during later maintenance.
(High Level) (High Level) Vision and PLC Team (e.g., “DO1 receives vision OK signal, triggers the production line to release”)
Pitfall 2: Protocol Trap – “All Support” Does Not Mean “Interoperable”
“All support Modbus TCP” is a common misleading statement in production line integration. When a certain automotive parts factory introduced a new vision system, it fell into trouble by trusting this promise: the equipment claimed protocol consistency, but data could never communicate, leading to a 4-hour production line halt.
The core of avoiding this pitfall is to penetrate the protocol name and confirm implementation details. Organize a special docking between both parties in advance to form a “Communication Protocol Details” document, clarifying key parameters such as address offsets and byte order; during the integration phase, conduct full-scene read/write tests, using Wireshark to capture and analyze data packet structures, ensuring that data fields and parsing rules match completely, eliminating the risk of “superficial compatibility”.
Upon investigation, three hidden differences were found: different address counting rules (vision starts from 0, PLC starts from 1), reversed byte order (vision big-endian, PLC little-endian), and mismatched data types (vision boolean, PLC integer parsing), which led to data garbling, production misjudgment, and qualified items being mistakenly rejected while defective items might pass unnoticed.
(e.g., Vision address +1 = PLC address)
Pitfall 3: Signal Loss – “Sent” Does Not Mean “Received”
The typical oversight in communication design is the assumption that a signal “ends” once it is sent. When a certain engine factory’s vision system detected a defect in a component, it sent an “NG” signal to trigger a shutdown, but there were multiple instances of “visual alarm, production line still running”, resulting in defective products being mixed with qualified ones across three batches.
The problem lies in the unidirectional communication mode: the vision system lacks a confirmation mechanism, and when the PLC misses or delays receiving due to data congestion, both sides cannot form a closed loop. This logic of “sending a WeChat message does not mean a reply” renders critical alarm signals ineffective, potentially leading to after-sales risks and secondary issues such as equipment collisions.
The solution is to establish a handshake mechanism: after the vision sends a signal, it starts a timer, and the PLC must return a confirmation signal upon receipt; if not received within 500ms, it automatically resends three times, and if it still fails, it triggers an audible and visual alarm. At the same time, set a 200ms signal hold time to ensure stable collection even when the PLC is busy, avoiding signal omissions.
Pitfall 4: IP Collision – The Hidden Killer of “Disconnecting” Upon Powering On
IP address conflicts are a common cause of production line paralysis upon powering on. After manually setting the IP of the vision device in a certain appliance factory, they found that communication between the vision and PLC frequently interrupted, and multiple reboots did not resolve the issue. Upon investigation, it was discovered that the IP completely duplicated that of the workshop’s stamping equipment, akin to a “house number conflict” causing data transmission chaos.
The network environment can also exacerbate the problem: the workshop network was not segmented, leading to frequent device broadcast storms; using consumer-grade switches resulted in severe packet loss under high load. Such issues are insidious—normal during low load in the daytime, but errors occur when multiple devices operate at night, causing fluctuations in production line qualification rates.
A three-pronged approach is needed to avoid pitfalls: in IP management, vision devices should uniformly use DHCP for automatic allocation; in special cases of manual settings, the network administrator should establish a “Device IP Allocation Table” to ensure uniqueness; in hardware, replace with industrial-grade switches to enhance anti-interference capability; in network planning, segment the vision and PLC into independent VLANs. During debugging, use packet capture tools to monitor latency and packet loss rates, and only proceed to production after meeting standards.
And synchronize to the cloud
Pitfall 5: Program Chaos – “Talking Past Each Other” After Changing Models
Changing models and modifying programs on the production line can trigger a communication “avalanche” with the slightest oversight. When a certain plastic bottle factory launched a new product, the PLC team updated the program without notifying the vision side, directly leading to a complete failure of the vision detection—the vision system continued to send data to the old register addresses, while the PLC had already enabled new addresses for reception, resulting in both sides “talking past each other” and causing a 3-hour production line halt.
Summary: Communication is No Small Matter, Details Determine Success or Failure
The communication link between vision and PLC, seemingly a “small detail” in production line transformation, is actually a “key hub” that determines the success or failure of the system. Signal reversal, protocol incompatibility, signal loss, IP conflicts, and program desynchronization are five major pitfalls, each of which can lead to production interruptions, quality decline, and cost increases. The communication between vision and PLC is a “key hub” in production line transformation, and these five major issues can lead to serious consequences such as batch defects and production line stoppages, increasing enterprise costs and operational risks.
The core logic of avoiding pitfalls lies in “clarifying in advance, verifying in progress, and tracing afterwards”: solidifying signal definitions and protocol details through written documents in advance to avoid “verbal agreements”; verifying communication reliability through testing and packet capture during the integration phase; and establishing change management and backup mechanisms in the later stage to ensure traceability of issues. Only by solidifying every detail of the communication link can vision and PLC truly achieve “seamless collaboration”, providing a solid guarantee for improving efficiency and preventing errors in the production line.
For project evaluation, please scan to add WeChat: note the evaluation project
Recommended Popular Articles:Click to directly access related articles:
001: After five years in the industry, a brief discussion on how to master machine vision
002: Practical PPT sharing | Application of 3D machine vision technology in intelligent manufacturing!
003: Main parameters of industrial cameras [Machine Vision Hardware]
004: Senior vision engineer organizes 39 questions about industrial cameras (Part 2) Like and save!
005: Machine vision accelerates the transition from 2D to 3D
006: Common solutions for frame loss in industrial cameras
007: Ten popular algorithms in artificial intelligence, explained in simple terms
008: Eight lighting techniques from Japanese lighting experts (in-depth article)
009: Analysis of common causes of misjudgment in machine vision
010: Methods and applications of three-dimensional imaging in machine vision