Understanding the OPC Server Connected But Not Updating Problem
OPC (Open Platform for Communications) bridges control system networks. An OPC client reads tags from a server, which reads data from PLCs and DCS controllers. When the OPC server shows a connected status but tags go stale, the problem is usually in the data path between the server and the controller. The server may be connected to the network but unable to poll the controller successfully.
In mixed-vendor plants, this scenario is common. An Emerson Ovation DCS operator station runs an OPC server to expose tags to a Rockwell FactoryTalk View HMI client. Everything shows green on the server dashboard. Yet three quarters of the tags show a timestamp from hours ago. What went wrong? The answer lies in DCOM security permissions, network routing tables, and controller session limits.
Root Cause 1: DCOM Security Misconfiguration on the OPC Server Host
OPC DA relies on Distributed Component Object Model (DCOM) for remote procedure calls. When the OPC client runs on a different machine from the OPC server, DCOM handles the communication between them. If the user account running the OPC client lacks proper DCOM launch and access permissions on the server host, the client can establish a TCP connection but receive no data.
Typical symptom: OPC client shows server as online. Browsing the server address space returns all items. However all items report quality as "Uncertain: Last Usable Value" with a stale timestamp. The root cause is almost always DCOM Authentication Level and Named Pipe permissions.
On a Windows Server hosting the OPC server, open Component Services (dcomcnfg). Navigate to Component Services > Computers > My Computer > DCOM Configuration. Find the OPC Server application. Set the Authentication Level to Packet Privacy. Grant Launch and Activation permissions to the OPC client user account. This forces encrypted DCOM communication and resolves most cross-domain permission failures.
Root Cause 2: Ethernet/IP Session Limit Exceeded on Allen-Bradley Controllers
Allen-Bradley controllers enforce a maximum number of simultaneous EtherNet/IP connections. For a CompactLogix 5380, the maximum number of owning connections (exclusive owner) is limited per controller. When the OPC server opens too many connections, the controller rejects new connection requests with CIP error code 0x0107 (Connection resource unavailable).
RSLinx Classic uses a "SoftLogix" or "Remote Device" driver to connect to controllers over EtherNet/IP. Each driver instance opens a connection. If you have multiple RSLinx driver instances trying to connect to the same controller, each consumes a separate connection. The controller accepts the TCP three-way handshake. However the CIP session establishment fails silently.
The fix requires reducing the number of owning connections. Set all OPC tag groups to use the same RSLinx driver instance. Increase the poll rate to share the connection efficiently. For large tag counts, use a Bridged connection rather than individual point-to-point connections to the controller. Configure the controller’s CIP Connection Manager to allow more connections if the controller model permits.
Step-by-Step: Diagnosing EtherNet/IP Connection Saturation
- Step 1: Open RSLinx Classic. Select Communications > Who Active. Right-click the controller. Choose Module Properties. Look at the Connections tab. Note the number of current connections. Compare this to the controller’s maximum owning connection limit from the controller’s EDS file or product datasheet.
- Step 2: Open RSLogix 5000 (or Studio 5000 Logix Designer). Go online with the controller. Navigate to Controller Properties > General > Connection. Check the "Connections Handled" counter. If this equals the maximum and new connections are being rejected, you have a connection saturation issue.
- Step 3: Open the OPC server configuration. Find the tag group properties. Set the "Requested Packet Interval" (RPI) to the same value for all tags to the same controller. This allows the OPC server to use a "Multiple" connection type instead of multiple "Exclusive Owner" connections, which is more efficient.
- Step 4: Reduce the total number of RSLinx driver instances. Consolidate all OPC tag subscriptions onto a single driver instance where possible. Use RSLinx Enterprise or FactoryTalk Linx for better connection management if your system supports it.
Root Cause 3: OPC UA Endpoint Mismatch and Certificate Trust Issues
OPC UA (Unified Architecture) uses a certificate-based security model. The OPC server presents an X.509 certificate. The OPC client must trust this certificate before any session is established. If the client does not trust the server’s certificate, the session establishment fails at the Hello message level. The client may show the server as online because the TCP connection succeeds. But no data flows because the security handshake never completes.
Emerson Ovation systems running Ovation UFC (Universal Function Controller) with OPC UA integration are particularly sensitive to certificate expiry. The Ovation OPC UA server certificate has a 5-year validity period. After renewal, all clients must re-accept the new certificate. A certificate expiry causes all OPC UA tags to go stale simultaneously while leaving the TCP connection intact.
To diagnose certificate issues, enable OPC UA stack trace logging on the client. Set the trace level to Verbose in the OPC client configuration. Look for log entries containing "CertificateRejected" or "BadCertificateUntrusted". These entries confirm a certificate trust failure. Then either renew and re-import the server certificate on the client, or configure the client to accept untrusted certificates for testing purposes.
Step-by-Step: Complete OPC Tag Update Diagnostic Workflow
- Step 1: Verify physical network connectivity from the OPC server host to the controller. Ping the controller’s IP address. If ping fails, check VLAN configuration, subnet mask, and router ACLs. EtherNet/IP requires UDP port 2222 (CIP) to be open between the server and controller.
- Step 2: Check the OPC server diagnostic view. Most OPC servers display per-tag quality codes. Good quality means the server successfully read the tag. Bad quality means the server encountered an error. Sort tags by quality to identify a pattern. If all tags from one controller are bad, the problem is at that controller or network path.
- Step 3: Verify that the controller’s program is running. A controller in Program mode stops executing tag updates. Use RSLogix to go online. Confirm the controller mode is Run. Check the controller event log for fault codes that might indicate why tag access is blocked.
- Step 4: Test with a direct client. Install the OPC Quick Client or Matrikon OPC Explorer on the OPC server host. Connect directly to the local OPC server. If tags update correctly locally but fail remotely, the problem is network or DCOM related. If tags fail locally, the problem is the controller or OPC server configuration.
Preventive Measures for Mixed DCS and PLC OPC Environments
Plants with both Emerson Ovation DCS and Allen-Bradley PLCs face integration challenges that pure-vendor environments avoid. The OPC bridge is a critical integration point that requires proactive maintenance. Schedule an annual review of all OPC certificates, DCOM permissions, and EtherNet/IP connection counts.
Use OPC UA instead of OPC DA whenever possible. OPC UA eliminates DCOM configuration entirely and provides certificate-based security with explicit trust management. Many modern OPC servers, including RSLinx Gateway and Emerson Ovation OPC UA, support OPC UA natively. Migrate legacy DA-only systems during a planned maintenance window.
Document all OPC tag configurations in a configuration management database. When tags go stale, the first question is always: what changed? An up-to-date configuration record lets you compare current state against the baseline and identify unauthorized changes before they cause production disruptions.
Conclusion & Action Advice
OPC server connection without tag updates is a layered problem. It can stem from DCOM security misconfiguration, EtherNet/IP connection saturation, certificate trust failures, network routing errors, or controller mode issues. The key diagnostic skill is to isolate which layer of the data path has failed. Start at the controller and work outward.
The action engineers must take now: build an OPC diagnostic checklist for your specific system. Document the expected DCOM permissions, connection counts, and certificate expiry dates. Run a preventive audit before the next unplanned stop. In a modern plant, OPC is the nervous system connecting every operator interface. When it fails, the entire plant loses visibility.