Why Do Plant Systems Show Different Event Times?
Why do PLC, SCADA and historian timestamps not match?
When PLC, SCADA and historian timestamps do not match, start with two checks: whether their clocks agree, and where each system assigns its timestamp. One record may show when a controller detected a change, while another shows when software received or stored it. Different time zones, clock drift and daylight-saving settings can also make records appear inconsistent.
Before comparing event records, establish what each time represents. Matching the displayed clock format is only the beginning.
What does each timestamp actually represent?
Ask the controls team or system supplier to trace one event from its origin to the screen. For every timestamp, identify the component that creates it and the action that triggers it.
Use this worksheet to record the answers:
| Record | Questions to resolve |
|---|---|
| PLC event | Does the controller timestamp the detected change, or provide only a value? |
| SCADA event | Is the time supplied by the controller or assigned when SCADA processes the change? |
| Historian entry | Does the historian preserve the incoming timestamp, assign another, or retain both? |
| Dashboard or export | Which stored field is displayed, and what conversion is applied? |
Event time describes when the source detected or recorded the event. It is not automatically the exact instant of the physical change: establish how the source detects that change.
Collection time describes when a downstream component receives or reads the record. A separate storage timestamp may describe when it is written to the database. Confirm the product's definitions rather than relying on field names.
For alarm investigations, record which stage each entry represents. Our guide to why operators ignore alarms covers the wider work of making alarm records useful to operators.
How should you check clock synchronisation?
Create an inventory of the clocks involved in the disputed records. Include controllers, SCADA servers, collection gateways and historian servers. Include the display computer if it performs local time conversion.
For each component, record:
- Its configured time source.
- Whether synchronisation is enabled and currently healthy.
- Its observed difference from the agreed reference.
- When synchronisation last succeeded, if available.
- What happens after a restart or loss of the time source.
A configured time source does not prove that a device is synchronising successfully. Verify the status and compare actual clocks. Clocks can drift apart when they operate without effective synchronisation.
Take comparisons close together and record how you measured them. A manually refreshed screen is a weak basis for judging a small difference.
Agree the required timing accuracy before changing settings. Relating production totals to a shift boundary and reconstructing a rapid equipment trip are different requirements. Specify an acceptable clock difference and collection delay for the investigation you need to support.
How do time zones and daylight saving affect comparisons?
Separate the stored timestamp from its displayed representation. Establish whether each field contains UTC, local time with an offset, or local time without an offset.
A UTC offset tells you the difference from UTC for that timestamp. A time zone also defines the rules used to translate between UTC and local civil time. Those rules may include daylight-saving changes.
Check the application, server, browser and export settings wherever they participate in conversion. Ask explicitly whether conversion has already happened upstream.
Where daylight saving applies, the clock change can produce repeated or missing local clock times. A local timestamp without an offset may therefore be ambiguous.
Choose a documented storage and display policy. For example, require unambiguous timestamps for exchange and clearly labelled local times for operators. Include exports in the policy so that a downloaded record remains understandable outside the application.
Can the clocks agree while event records still differ?
Yes. Clock synchronisation and data collection answer separate questions.
Polling, processing, buffering and transmission can introduce time between source detection and downstream receipt. Following an interruption, buffered records may arrive later while retaining their earlier source timestamps, where the system supports that behaviour.
The following example is illustrative; it is not a measured plant result.
| Stage | Illustrative recorded time | Meaning in this example |
|---|---|---|
| PLC detects a stop | 10:15:00.200 | Source event time |
| SCADA receives the change | 10:15:00.650 | Collection time |
| Historian stores the entry | 10:15:01.100 | Storage time |
These records could describe one stop without a clock fault. The investigation must establish whether the historian's event field contains the original source time or a later assigned time.
If only collection time was recorded, do not present it as a precise reconstruction of source event time. Document what evidence is available and the limits it places on the investigation.
How can you test the complete timestamp path?
Use an approved test signal or a planned operational event. Agree the method with the plant team, and keep the test within existing operating procedures.
Follow this checklist:
- Define the question. State which records must be comparable and the timing accuracy required.
- Confirm event identity. Establish that all records refer to the same signal transition.
- Capture clock health. Record synchronisation status and observed clock differences.
- Capture original records. Retain source, receipt and storage fields wherever available.
- Compare representations. Check raw records, screens and an export using the same reference basis.
- Inspect precision. Record whether each stage retains, rounds or hides fractional seconds.
- Test recovery where authorised. Establish how buffered events are timestamped after a collection interruption.
- Record the result. State what passed, what failed and what remains unknown.
Displayed precision does not establish clock accuracy. A field showing fractional seconds still needs evidence that its clock and timestamp assignment meet the requirement.
Put the agreed rules into the functional design specification, including timestamp ownership, conversion and interruption behaviour.
What common mistakes should you avoid?
Adding a fixed correction too early. First distinguish a display conversion from a clock error or collection delay. Correct the responsible setting or interface rule before adjusting historical records.
Comparing different events. Confirm the tag, transition and event identity before interpreting nearby entries as the same occurrence.
Sorting only by arrival time. If an investigation needs source order, establish whether that information was preserved. Arrival order alone may differ from source event order when delivery is delayed.
Correcting every mismatch as a clock fault. A valid event timestamp and a valid receipt timestamp can differ because they describe different stages.
Changing historical data without retaining evidence. Preserve the original records and document any correction, including its basis.
Applying the same timing requirement everywhere. Define what each use needs. For production reporting, also review what an OEE number hides before treating timestamp alignment as proof that the underlying calculation is sound.
Related reading: buffering plant data during network outages.
What are the frequently asked questions?
Should the historian use PLC time or its own time?
Decide according to the record's purpose and the source clock's reliability. For event reconstruction, require a documented source timestamp and consider retaining receipt time alongside it. Verify that the interface preserves both as agreed.
Will synchronising every clock fix the problem?
It addresses clock disagreement. You must still check timestamp assignment, time conversion and the collection path. Validate the complete record after making changes.
Can old records be corrected?
First determine whether the original meaning, time basis and correction are supported by evidence. If they are unclear, retain that uncertainty instead of assigning a precise corrected time.
What should we ask an integrator to deliver?
Request a clock inventory, timestamp ownership map, conversion rules, interruption behaviour and acceptance-test evidence. Our services page is a starting point for discussing the scope of a plant data assessment.