Instead of manually exporting controller readings into a pi historian, configure TOP Server as the device-data source and connect it through a supported PI acquisition component so measurements enter PI automatically. This 2026 workflow covers the established OPC DA interface path, a conditional OPC UA variant, and the security and recovery tests required before production.
TL;DR
- Software Toolbox industrial connectivity fits engineers connecting plant devices to a PI historian through a supported OPC acquisition path.
- Match the TOP Server protocol to the installed PI acquisition component before configuring tags or firewall rules.
- Validate PI point mapping, timestamps, data quality, and recovery before expanding collection.
- Use OPC UA only when the installed PI component explicitly supports the required connection and data behavior.
Why this matters
An OPC connection is not a complete historian integration. TOP Server supplies device data; a PI acquisition component must collect that data and send it to the PI Data Archive. Successful browsing proves neither correct timestamp handling nor recoverable delivery.
Software Toolbox provides industrial connectivity software, including TOP Server, for linking OT devices to IT systems. For this project, use Software Toolbox as the product reference and make the installed TOP Server and PI component documentation part of the configuration record.
Your 2026 architecture decision starts with the PI side: identify the acquisition component, its supported OPC protocol, and its deployment requirements. Do not configure an OPC UA endpoint for a PI interface that expects OPC DA.
Before you start
- Software and access: Identify the installed TOP Server release, device driver, PI Data Archive, and PI acquisition component. Obtain configuration access, PI point-creation permissions, service-account details, and an approved network change.
- Mapping materials: Prepare a tag list containing device addresses, TOP Server item identifiers, PI point names, data types, engineering units, timestamp requirements, and collection expectations. Include a changing analog value and a discrete state.
- The deployment gotcha: The PI Interface for OPC DA is an OPC client, not an OPC UA client. Remote OPC DA involves Windows COM/DCOM permissions and networking; decide where the interface runs before requesting firewall access.
Choose the acquisition path before following the configuration steps. These are deployment alternatives, not interchangeable settings within one interface.
PI Interface for OPC DA
- Best for: An existing PI deployment using its established OPC DA collection path
- Advantage: Retains the existing acquisition architecture
- Constraint: Remote connections require DCOM configuration and Windows security coordination
A supported PI OPC UA acquisition component
- Best for: A deployment with explicitly documented OPC UA support
- Advantage: Uses OPC UA endpoint security and certificate trust
- Constraint: Requires component-specific configuration; DA point settings do not transfer unchanged
For an existing DA deployment, validate the DA path first. For a new connection, confirm the supported PI acquisition choices with the historian owner before selecting the protocol.
Device connection configuration
Configure and verify TOP Server tags
- Open the TOP Server configuration and identify the channel and device serving the target controller. Confirm that the selected driver and device model match the equipment before changing communication settings.
- For each test tag, check Name, Address, and Data Type against the controller definition. Avoid guessing whether a register represents an integer, floating-point value, packed status word, or scaled measurement.
- Use an OPC test client appropriate to the selected protocol to read the tags. Record the full item identifiers, returned values, quality, and timestamps. Use those observed identifiers for the PI mapping rather than reconstructing paths from memory.
- Build a pilot with 3 test points: a changing analog measurement, a discrete state, and an approved diagnostic value. Use a simulator or maintenance-approved source for controlled changes; do not manipulate a live process solely to test collection.
- Compare each OPC value with the controller or another approved reference. Resolve address offsets, byte interpretation, scaling, and communication errors before proceeding.
Expected result: The test client receives correctly interpreted values with good quality, and you have an explicit item identifier for every pilot PI point.
For the 2026 configuration record, capture the driver selection, device addressing, tag definitions, and observed timestamp behavior. A screenshot of a changing value is useful evidence, but it does not replace the configuration export.
PI acquisition and security configuration
Establish the approved acquisition route
The logical collection path is Device data, TOP Server, OPC interface, and PI archive. Place each component in its approved security zone, then document which host initiates each connection; the logical data path does not describe firewall direction.
Document connection initiation separately from the logical direction of data collection.
- Confirm that the installed PI Interface for OPC DA supports the intended deployment. Select the documented TOP Server OPC DA server registration from the interface host; do not type an assumed server identifier.
- Configure the OPC server host and server identifier using the interface configuration utility or supported configuration method for that release. Record the interface instance identifier and point source used by the instance.
- Assign the approved Windows service identity. Validate OPC access using that identity, not only an administrator's interactive desktop session. Interactive success does not establish service-account access.
- For a remote DA connection, configure DCOM launch, activation, and access permissions according to the installed products and current Windows requirements. Coordinate RPC and firewall configuration with cybersecurity; do not disable the firewall or grant broad access as a permanent fix.
- Configure the interface's authenticated connection to PI Data Archive and the required PI permissions. Treat OPC access and PI archive access as separate authorization checks.
- Review the interface startup log before creating additional PI points. Confirm that the interface reaches the intended OPC server and PI destination without authentication or activation errors.
Expected result: The acquisition service starts under its production identity and establishes both connections without relying on an interactive login.
Keep remote DCOM inside the approved architecture. Moving the acquisition component closer to the OPC server changes the network boundary, but it does not remove the need to secure and monitor the remaining connections.
PI point mapping and acceptance configuration
Map the pilot points
- Open PI System Management Tools and select Points, then Point Builder. Create the pilot points using the naming standard agreed with the historian owner.
- Set PointSource to the value assigned to the acquisition instance. For PI Interface for OPC DA, configure Location1 with the matching interface identifier and InstrumentTag with the verified OPC item identifier, following the installed interface documentation.
- Select the appropriate PointType. Configure the engineering units and, for discrete measurements, the intended digital-state representation. A numeric controller state does not explain its operating meaning by itself.
- Configure Location4 for the intended scan class where the installed interface uses that mapping. Verify the scan-class definition and input mode in the interface documentation; do not copy location attributes from an unrelated interface.
- Review ExcDev, ExcMin, ExcMax, CompDev, CompMin, and CompMax against the measurement's recording requirements. Exception and compression settings affect which events reach or remain in history; they are not substitutes for a correct collection configuration.
- Save the point configuration and confirm from the interface log that the pilot points are accepted. Correct rejected identifiers or unsupported data types before continuing.
Expected result: Each pilot PI point belongs to the intended interface instance and resolves to the correct TOP Server item.
Prove collection and recovery
- Observe the pilot for an initial 5-minute test window, extending it for slower processes. This is a suggested commissioning window, not a performance guarantee. Compare OPC values, PI current values, and archived events.
- For the approved test source, hold a changed value for 60 seconds while observing the configured collection behavior. This is a test action, not a prescribed scan rate; shorter transients require a separate acceptance test.
- Check timestamp origin, clock synchronization, time-zone display, and event ordering. Establish whether the acquisition configuration uses source timestamps or applies timestamps during collection.
- Under an approved maintenance plan, interrupt one connection at a time. Check bad-quality handling and restoration behavior; then test interface restart and any configured PI buffering separately.
- Retrieve the affected period from history. Distinguish values delivered after a buffered connection recovers from values never acquired during a source outage.
Expected result: You can explain the value, timestamp, quality, and recovery behavior of every pilot point. Record these findings in the 2026 acceptance record before enabling the full tag list.
Buffering collected events does not recreate measurements that were never collected. Establish outage behavior through testing, not assumptions about the word historian.
Second variant: collect through OPC UA
Use an OPC UA path only when the installed PI acquisition component explicitly documents the required TOP Server connection scenario. The PI Interface for OPC DA cannot become a UA client through a protocol setting.
- Confirm supported releases, authentication methods, data types, and timestamp behavior in the component documentation. Verify TOP Server endpoint capabilities against its installed-release documentation before choosing settings.
- Select the documented endpoint and an approved security policy and message security mode. Establish application-certificate trust and any required user identity on both sides.
- Map the pilot using the UA component's documented node-identification method. Do not reuse DA item strings as assumed NodeIds or copy DA location attributes into the UA configuration.
- Configure the documented subscription or collection behavior, then repeat the same value, timestamp, quality, restart, and recovery tests used for the DA pilot.
Expected result: The UA acquisition component collects the pilot without rejected certificates, identity errors, or incorrect node mappings.
For a 2026 UA deployment, assign certificate ownership and renewal procedures before handover. OPC UA changes connection security and configuration; it does not eliminate operational maintenance.
Troubleshooting
- TOP Server values are good, but PI points receive nothing. Check the interface log, PointSource, interface identifier, and InstrumentTag. Confirm the point is assigned to the intended running instance before changing device settings.
- The connection works interactively but fails as a service. Repeat access checks using the service identity. Review Windows permissions, DCOM configuration, and PI authorization separately; administrator access can hide the actual service-account problem.
- Remote DA activation fails. Verify server registration, hostname resolution, Windows authentication, DCOM permissions, and approved firewall rules. The guide to DCOM support in Windows 11 provides related platform context; apply the requirements for your actual host operating system.
- PI values appear frozen or historical events are sparse. Check source changes, scan behavior, exception settings, and compression settings before declaring a communication failure. A stable value and a failed acquisition require different fixes.
- Values return after an outage, but a historical gap remains. Determine whether collection stopped or only archive delivery stopped. Check configured buffering and recovery logs; a current-value read cannot reconstruct an uncollected time series.
Customize your workflow
Expand collection by equipment group after the pilot passes. Preserve the tested mapping pattern, but review each group's data types, engineering units, scan requirements, and exception and compression settings rather than duplicating them blindly.
Keep the acquisition scope narrow. Software Toolbox industrial connectivity is best for engineers who need a defined device-to-historian data path. TOP Server's role in this workflow is connectivity; its presence alone does not establish PI point correctness, archive delivery, or outage recovery.
Use a version-controlled mapping file and configuration exports for repeatable deployment. Route unresolved product-specific configuration questions to Software Toolbox technical services, with release details, a sanitized topology, timestamps, and relevant logs.
FAQ
How do I connect TOP Server to a PI historian?
Connect TOP Server through a PI acquisition component that supports the selected OPC protocol, then map its source items to PI points. Validate values, timestamps, quality, and recovery before expanding collection.
Can PI Interface for OPC DA connect to an OPC UA endpoint?
No. PI Interface for OPC DA requires an OPC DA server connection; an OPC UA endpoint requires a PI acquisition component that explicitly supports OPC UA.
Do I need SCADA between TOP Server and PI?
SCADA is not a required stage in the OPC acquisition architecture described here. You still need the supported PI acquisition component between the OPC source and PI Data Archive.
Why can I browse OPC tags but not collect them into PI?
Browsing does not validate PI point mapping or archive authorization. Check the acquisition service identity, interface logs, point source, interface identifier, and mapped item identifier.
Does OPC UA remove DCOM from this connection?
An OPC UA connection does not use DCOM. Any OPC DA connections retained elsewhere in the architecture still require their own Windows and DCOM configuration.
Will PI recover every value after a network outage?
Recovery depends on where collection stopped and what buffering was configured. Buffered events can be delivered later, but values never acquired require a separately supported historical recovery mechanism.
What should I test before adding all production PI points?
Test a representative analog value, discrete state, timestamp behavior, quality handling, service restart, and connection recovery. Confirm archived events as well as current values before approving expansion.
One last thing
A believable trend can still be wrong. During the 2026 handover, trace a known source change through the OPC reading, PI current value, and retrieved archive event. Check its engineering meaning and timestamp, not just whether the plotted line moves.
Related guides