Back to all posts
Plant Data7 min read·October 25, 2026

How Can You Collect Data From an Existing PLC?

TB
ThynkBlox Automation Team
Automation Engineering

You can collect data from an existing PLC through a supported read interface, either directly or through an existing supervisory system or gateway. Start with the smallest useful tag set, then verify that the proposed connection and polling rate fit the controller's documented limits and current workload. Set read-only access where supported, separate the collection path from business networks, and prove that collection leaves machine operation within agreed limits before expanding it.

What should you check before connecting?

Begin with an inventory, not a dashboard purchase. Record the controller model, firmware, communication modules, available interfaces and existing connections. Include operator panels, engineering tools, supervisory systems and any equipment already exchanging data.

Do not assume that an unused network socket means spare communication capacity. Confirm the connection limits and communication restrictions for the exact controller and module combination.

Ask maintenance for the current program backup, network drawing and any known communication faults. Agree who can authorise access and when testing is permitted.

If the controller cannot be identified, documentation is missing or production is already experiencing unexplained faults, make assessment the first job. An additional connection should not become an experiment during normal production.

Which supported read interface should you use?

Compare the available routes against the installed equipment, access controls and operational constraints.

Possible routeWhat to establish before choosing it
Existing supervisory system or historianWhether it already holds the required values and permits supported access without extra PLC reads
Existing gatewayWhether its supported driver, capacity and permissions cover the required data
Controller's native data interfaceWhether the installed hardware and firmware support it, with sufficient connection capacity
Supported communication moduleWhether installation needs downtime, configuration changes or program changes

Prefer a route that reuses an established collection point when its capacity, data quality and support arrangements are suitable. Verify how that route obtains its values; adding a consumer should not silently create another polling session against the PLC.

If OPC UA is available, confirm where its server runs and which data it exposes. The server location and available features depend on the installed product and configuration.

What is the smallest useful tag set?

Choose one operational question first. For example: “When did this machine stop, and what production count had it reached?”

For that question, propose a machine state, a cumulative production count and a fault code if one is already available. Do not start by collecting every accessible address.

For each proposed tag, record:

  • Its address or name and plain-language meaning.
  • Its data type, unit and scaling.
  • Which machine or production stage it describes.
  • How quickly the information is needed.
  • What to record when the value is unavailable.
  • Whether it requires interpretation outside the PLC.

Check each definition with someone who understands the machine sequence. Specify whether “running” means a motor command, confirmed motion or active production. Define whether the counter represents total pieces, accepted pieces or completed cycles.

For production reporting, read what OEE hides behind the number before treating those signals as a complete performance measure.

How do you set polling limits without overloading the controller?

Treat polling frequency as a requirement to justify, not a setting to maximise.

Separate slow reporting values from signals that need quicker observation. Avoid assigning one fast polling interval to the entire tag list. Where the supported interface permits grouped reads or subscriptions, assess those options against the manufacturer's guidance.

There is no universal safe polling interval for an existing PLC. Establish a baseline before enabling collection: controller diagnostics, communication errors, operator-panel response and any available measures of controller load.

Then enable the smallest proposed collection job. Compare the same observations during representative operation, including startup, changeover and fault recovery where practical.

Agree stop conditions beforehand. These should include new communication faults, delayed operator response or departure from the controller's documented operating limits. Stop collection and investigate if a condition is met.

Set limits on retries and reconnection attempts too. Require the collector to back off during a fault rather than repeatedly opening connections or issuing requests without restraint.

What would a limited pilot look like?

The following example is illustrative, not a tested configuration or a recommended polling rate.

Suppose a packaging machine needs a shift report showing production count and broad machine state. Propose three tags: cumulative count, state and active fault code. Begin by evaluating a five-second collection interval against the reporting need and equipment limits.

That interval is unsuitable if the requirement includes reliably capturing every brief stop. A state change can occur and disappear between polls.

Resolve that gap before building the report. Investigate whether an existing event history, maintained counter or supported event interface provides the required record. Do not present sampled observations as a complete event history.

Accept the pilot only after checking tag meaning, collection behaviour, operational impact and recovery from a lost connection.

How do you enforce read-only access?

Write down exactly what the collector may read. Exclude writes, program downloads, mode changes, resets, recipe transfers and remote commands from its scope.

Where the interface supports it, use a dedicated identity restricted to the approved data and read operations. Verify the effective permissions through an authorised commissioning procedure.

Older interfaces may not provide account-based permissions or separate read and write privileges.

In that case, selecting “read only” in the collector is not sufficient evidence of an enforced boundary. Establish which restrictions a gateway or other control can actually enforce, document the remaining exposure and obtain a specific access decision.

Keep data collection separate from remote engineering access. Each should have its own purpose, access rules and approval.

How should you separate the collection network?

Define an approved path from the controller to the collector and from the collector to the reporting destination.

Specify the required endpoints, connection directions and permitted services. Have the plant's network owner review the design, including how maintenance will reach the equipment.

Do not give a dashboard unrestricted access to the control network. If a collector connects to both plant and business networks, explicitly review its routing, forwarding and administration settings.

Require a collection failure to leave machine control independent. Test what happens when the reporting destination is unavailable, including retry limits, local storage limits and recovery. Define how reports distinguish missing data from a genuine zero.

Put these decisions in the functional design specification, alongside ownership and acceptance criteria.

What common mistakes should you avoid?

Assuming every legacy PLC has spare capacity. Verify support and headroom before choosing direct collection.

Polling everything at the same rate. Match each data group to its purpose and documented limits.

Treating a successful read as acceptance. Also verify meaning, access restrictions, operational impact and failure recovery.

Turning missing data into zero. Preserve an unavailable or stale status so readers can distinguish a collection problem from a production result.

Adding program logic without assessing the change. New counters, timestamps or buffers need their own review and commissioning plan.

Extending collection into safety functions. Any proposed safety-related change needs assessment with a qualified assessor.

What if another PLC connection is not suitable?

Stop the direct-connection proposal. Assess an existing supervisory data source, a supported gateway arrangement or a planned communication upgrade.

If none is suitable, document the limitation and plan the work around an agreed shutdown. Our guide to planning a PLC retrofit around a shutdown window covers preparation and commissioning.

The useful outcome may be a clearly defined upgrade, rather than immediate collection.

Related reading: choosing plant data sampling intervals, aligning timestamps across plant data systems, buffering plant data during network outages, using machine runtime for maintenance scheduling and adding sensors to machines without data outputs.

When a protocol converter is the answer, see how to choose an industrial protocol gateway. See RPA or a proper interface for legacy plant software.

Frequently asked questions

Can you collect data without changing the PLC program?

Possibly. First establish whether the required values are already exposed through a supported interface. Treat any additional logic or configuration change as separate work requiring review.

How often should you read the tags?

Choose intervals from the reporting requirement, then validate them against equipment documentation and observed operation. Use a suitable event-recording method when periodic sampling cannot meet the requirement.

Can the collector send data to a remote dashboard?

Make that a separate design decision. Define the approved network path, access controls, outage behaviour and data ownership before enabling it.

What should the handover include?

Request the approved tag list, interface settings, polling and retry limits, access rules, network drawing, acceptance results and recovery procedure. Name the owner responsible for future changes.

For help defining a limited collection scope, see our automation services or contact the team.

Next step

Want to apply this to your plant?

Every project starts with a fixed-fee assessment, and you can stop after the report.