Connecting a sortation control system to an optimization layer is not a novel idea. But doing it cleanly, without introducing latency that defeats the purpose, requires choosing the right protocol and understanding exactly what data lives where inside the PLC. This guide walks through how Sortwyre connects to common sorter control systems, what data streams we pull, and what you should realistically expect in terms of latency and security posture.
Why OPC-UA for Sorter PLC Integration
OPC-UA (Open Platform Communications Unified Architecture) is an industrial communication standard developed by the OPC Foundation. It defines both a transport layer and an information model, which matters more than most engineers expect when they first approach sorter integration.
Older alternatives like Modbus TCP or proprietary vendor protocols expose raw register maps. You get a flat address space with no semantic context. Register 40017 means something, but only if you have the vendor documentation open in another window. OPC-UA solves this with a structured address space: every data point is a node with a defined type, namespace, and browse path. You can enumerate what a server exposes without guessing register offsets.
For sortation specifically, this matters because a typical sorter PLC exposes dozens of distinct data streams. Induction scanner read rates, belt speed feedback, divert confirmation signals, motor current draw, photoeye trip counts — each one lives in a named node. We found early on that trying to maintain Modbus register maps across firmware updates was a maintenance nightmare. OPC-UA's self-describing address space cuts that entirely.
That said, not every site has OPC-UA available on the sorter controller. For those cases, Sortwyre falls back to a Modbus TCP sensor bridge — a lightweight edge process that polls the Modbus registers and re-exposes them as OPC-UA nodes to the optimization engine. Same interface, different acquisition path. The latency penalty is small: typically under 200ms additional round-trip.
Connecting to Dematic and Honeywell Intelligrated Systems
The two control systems we see most often in mid-size fulfillment centers are Dematic's iQ Sorter Software and Honeywell Intelligrated's Momentum WCS. Both support OPC-UA server functionality, though the configuration steps differ.
Dematic iQ
Dematic exposes an OPC-UA server at the iQ Sorter Software layer, not directly at the PLC. The server endpoint is typically opc.tcp://[sorter-host]:4840 using the default OPC-UA port. Security mode is configurable; most sites we've worked with run SignAndEncrypt with Basic256Sha256. Node IDs are structured under the urn:dematic:iq:sorter namespace.
Key nodes for Sortwyre's optimization loop include the induction gap target register, per-zone divert confirmation counts, scanner read-rate telemetry by induction station, and belt speed feedback from the tachometer input. All four update at the PLC scan rate, typically 100ms cycles on a Siemens S7-1500 base.
Honeywell Intelligrated Momentum
Momentum's OPC-UA implementation sits at the WCS controller level. The namespace structure is flatter than Dematic's, and node browse names follow a zone-machine-parameter hierarchy. We typically connect to the urn:hii:momentum:wcs namespace and subscribe to nodes covering induction rate targets, crossbelt tray occupancy, scan confirmation rates, and the dimensioner output feed when a Dimensioning, Weighing, and Scanning (DWS) unit is in the line.
One practical note on Momentum: the OPC-UA server license is a separate line item in Honeywell's configuration tooling. We've arrived at sites where the WCS was installed without the OPC-UA server activated. Verify the license is present before scoping integration timelines.
Data Streams Available
Here's what Sortwyre actually reads from a fully instrumented sorter line:
| Data Stream | Source Node | Update Rate |
|---|---|---|
| PLC event logs | EventNotifier on zone controllers | Event-driven |
| Induction scanner reads | Scanner station read-rate nodes | 100ms (polled) |
| Divert confirmation sensors | Per-divert confirmation registers | 100ms (polled) |
| Dimensioner outputs | DWS system OPC-UA nodes | Per-parcel trigger |
| Belt speed feedback | Tachometer input nodes | 500ms (polled) |
| Motor current draw | Drive health registers | 1s (polled) |
Not every site has all six. Dimensioner data in particular is missing from a surprising number of otherwise well-instrumented facilities. When DWS data is absent, the optimization engine infers parcel size distribution from scan-rate variability and divert timing patterns. Less precise, but functional.
OT/IT Security Considerations
Connecting an optimization system to a sorter PLC crosses the OT/IT boundary. This deserves more attention than it usually gets during initial scoping calls.
Sortwyre runs entirely inside the customer's OT network segment. There is no cloud dependency for the real-time optimization loop. The optimization engine process runs on a hardened edge node — either a rack-mounted server in the MCC room or a DIN-rail mounted industrial PC near the sorter controller. All OPC-UA traffic stays on the sorter LAN. The only outbound communication is the shift-level performance log pushed to Sortwyre's dashboard, which goes through an outbound HTTPS connection on a separate network interface with customer-controlled firewall rules.
For OPC-UA security, we require at minimum Sign mode and recommend SignAndEncrypt. Certificate management is handled through Sortwyre's edge agent, which generates a self-signed certificate on first boot and presents it to the OPC-UA server for trust-list enrollment. The facility's controls engineer approves the certificate once through the PLC programming interface. After that, the connection is automatic on restart.
One thing that catches teams off guard: OPC-UA sessions have a configured session timeout. If the optimization engine restarts or loses network briefly, the PLC will close the session after the timeout window (default 30 seconds on most Siemens configurations). Sortwyre's connection manager handles automatic session re-establishment with exponential backoff, so a brief network hiccup doesn't stop the optimization loop — it just creates a short data gap in the event log.
What to Expect from the 90-Second Optimization Loop
Sortwyre's optimization engine runs on a 90-second rolling window. Every 90 seconds, it pushes updated induction rate and divert timing parameters back to the sorter PLC via the same OPC-UA connection it reads from. Here's what that latency profile looks like in practice.
Data acquisition latency from PLC to optimization engine is typically under 300ms total, including the OPC-UA subscription delivery, local processing, and write-back to the parameter nodes. The 90-second cycle time is not driven by data acquisition speed — it's driven by how quickly the sorter line stabilizes after a parameter change. Push a new induction gap target and the line needs roughly 60-80 seconds to reach steady-state throughput at the new setting. Updating more frequently than that creates oscillation. 90 seconds is the right cadence. Not a technical limitation. A deliberate design choice.
In our deployment experience, the biggest latency source is not the OPC-UA connection itself but the write-back path. Some WCS implementations require parameter updates to flow through the WCS application layer rather than being written directly to PLC nodes. When that's the case, the Sortwyre edge agent uses the WCS REST API or vendor-specific write interface instead of direct OPC-UA writes. The cycle time remains 90 seconds; the write path is just longer.
Getting Started
The integration process for a new site typically takes 4-8 hours from first OPC-UA connection to live optimization. The majority of that time is certificate enrollment, node discovery, and verifying that the data streams are correctly mapped to the optimization engine's input schema. The actual optimization start is anticlimactic — parameters start updating, throughput numbers on the dashboard start moving, and the sorter line keeps running the whole time.
If you're evaluating whether OPC-UA integration is feasible at your facility, the first question to answer is whether your sorter control system has an OPC-UA server license active. If yes, integration is straightforward. If not, the Modbus TCP bridge path works for most older Dematic and Intelligrated hardware. Either way, no changes to the sorter control logic are required. Sortwyre reads and writes to defined parameter nodes — it never modifies ladder logic or sequence programs.
Contact us if you want to walk through the specific node map for your control system before committing to an integration project.