Companion specification · Tier 2

MachineVision AMCM

Document-
Version1.00.0
Published2024-05-17
Namespacehttp://opcfoundation.org/UA/MachineVision/AMCM/
SectorVision

AMCM is the asynchronous machine control extension to OPC UA for Machine Vision, published 2024-05-17 as version 1.00.0. It adds request and response handling so a controller can issue vision jobs and collect results without blocking, which the 2019 base model handled only thinly.

Where it sits in the stack

AMCM is an extension, not a replacement: it carries its own namespace, http://opcfoundation.org/UA/MachineVision/AMCM/, and sits on top of the MachineVision base model, which in turn is a system-level control interface rather than a device model. A server implementing AMCM will carry both namespaces, and a client that loads only one of them cannot resolve the inherited types. The pattern it adds, an asynchronous job request with a correlated response, is the same problem Machinery Result and Machinery Jobs solve for the machinery stack, so on a modern line the same information may be reachable two ways.

Using it on a real machine

  1. Check for both namespacesRead NamespaceArray and expect the MachineVision URI and the AMCM URI together. AMCM alone is not a complete model, and a client loading only the AMCM nodeset will fail to decode inherited types.
  2. Establish that the vendor actually implements itThis extension is recent, from May 2024, and adoption is thin. Ask the vision vendor specifically whether AMCM is implemented and in which software version, rather than accepting a general statement of OPC UA vision support.
  3. Correlate requests to responses explicitlyThe whole purpose of an asynchronous model is that responses arrive out of order. Keep your own correlation table keyed on the request identifier the server returns, and never assume the next result belongs to the last request you sent.
  4. Design for the timeout caseAn asynchronous request that never produces a response is a normal failure mode. Give every outstanding request a deadline in your client and decide what a missing vision result means for the part before you go live.
  5. Keep the base model integration working without itBecause coverage is uneven, build your collection against the MachineVision base model first and treat AMCM as an improvement where present. That way one vendor's missing extension does not block the project.

What it gives you for OEE

Nothing directly. AMCM is a control and correlation mechanism, not a measurement model: it has no counters, no state machine for the machine being inspected and no stop reasons. Its indirect contribution is to the quality factor, in that reliable request and response correlation is what makes per-part inspection results trustworthy enough to count as scrap. Availability, cycle counts and reasons still come from Machinery, MachineTool, PackML or the vendor's own tags.

Try it free: There is no free public server implementing AMCM, and because it is an extension there is not much to see without a base MachineVision server either. The practical route is offline: load the AMCM nodeset alongside Opc.Ua.MachineVision.NodeSet2.xml into a local server and inspect what the extension adds. The public umati sample server does not simulate it, although the umati dashboard client lists MachineVision Part 2 among its partially covered models.

Gotchas

  • It is an extension namespace. Two nodesets must be loaded, and a partial load produces undecodable inherited types rather than a clean error.
  • Published in May 2024, so field coverage is limited. Treat vendor support as something to confirm per software version, not as a property of the product line.
  • Asynchronous means genuinely unordered. Client code written against the base model's implicit last-result semantics will appear to work in testing and mis-attribute results under load.
  • The base MachineVision model still carries no image transport, and AMCM does not change that.
Verify before you rely on this: This page deliberately lists no type inventory. AMCM is recent and we are not confident of its exact type and method names, and publishing guesses for a control interface would be actively harmful. Confirm the type names, the correlation identifiers and the method signatures against the AMCM nodeset for the version your vision system ships.

Once the machine speaks a companion spec, the OEE mapping is portable across vendors, which is the point at which analytics stops being a per-machine project.

See Jemba →

Last verified 7 September 2026 against OPC UA 1.05.07. Version and publication date read from the released nodeset. OPC UA Connect is an educational resource and is not affiliated with the OPC Foundation.

Free download
OPC UA Machine Data → OEE Mapping Kit
Which nodes to subscribe to, how they become Availability, Performance and Quality, and which companion specification already gives you the answer, so you stop inventing tags.
Get the free mapping kit →
Free, no course, no login