Ignition Power Chart can keep drawing the last historical sample after a field node becomes unreachable. The chart is showing the most recent record it received; it is not proving that the value, Remote Tag Provider, historian path, or Gateway Network connection is currently healthy. Treat freshness as a separate signal and drive an explicit stale-data indication from connection state, tag quality, or a heartbeat.
Which path supplies the value?
Follow the packet from the sensor to the operator display. The Edge node acquires the live OPC value, exposes tags to the Central Gateway as a Remote Tag Provider, and logs history. The Perspective session on the Central Gateway renders the Power Chart.
Separate the live-tag path from the historical-query path. A live binding reads the current remote tag and its quality. A historical pen requests stored samples and draws whatever records the query returns. A successful query containing an old final sample says nothing about present acquisition health.
| Path | Sender | Destination | What failure changes |
|---|---|---|---|
| Live value | Edge acquisition | Central remote tag | Value updates stop and tag quality should leave good state |
| Historical data | Configured history provider | Central Power Chart | New records stop arriving; previously returned records remain drawable |
| Health signal | Gateway connection, tag quality, or Edge heartbeat | Perspective view | Provides the explicit stale or disconnected state |
Check: Identify the configured provider for every pen and confirm whether the view also has a live binding from the same Edge node. Proceed only after the live and historical paths are documented separately.
Is layer one carrying the Gateway Network session?
Layer one first. Check the field-node link, switch port, power, cabling, and routed path before changing the chart. A Power Chart cannot distinguish a broken cable from a stopped remote service when both conditions prevent new records from reaching the Central Gateway.
| Item | Value to record | Pass condition |
|---|---|---|
| Edge address | Address configured for the gateway connection | Central Gateway reaches the intended Edge node |
| Gateway Network port | Port configured at both endpoints | Intermediate devices permit the configured traffic |
| Physical link | Observed link and error state | Link remains up without increasing physical errors |
| Route | Actual path between gateways | No failed hop or unintended address translation blocks the session |
Do not substitute a general reachability test for application-session status. Network reachability can pass while the configured port, certificate relationship, service, or Gateway Network connection remains unavailable.
Check: Confirm both gateways are reachable over the configured address and port, then observe an established Gateway Network connection before testing tags.
Does the Central Gateway see the Edge node as connected?
Open the Gateway Network diagnostics on both ends and compare connection state and recent transitions. The tested installation uses Ignition Platform v8.1.47 centrally and Ignition Edge Platform v8.1.48 in the field. Record that version difference when reviewing diagnostics or release notes, but diagnose the connection state before treating the maintenance-version difference as the cause.
| Observation | Interpretation | Next action |
|---|---|---|
| Connection down on both gateways | Failure is below the tag and chart layers | Resolve transport, trust, service, or gateway configuration |
| Connection healthy but remote tags bad | Provider or acquisition path has failed | Inspect Remote Tag Provider and OPC diagnostics |
| Connection and live tags healthy, but no recent history | Historical logging or history-query path has failed | Inspect historian status and pen provider selection |
| Fresh history exists but display does not advance | Problem is in the Perspective session, chart query, or rendering state | Reload the view and inspect the chart request |
Check: Force a controlled Gateway Network interruption and verify that the Central Gateway reports the Edge connection leaving its healthy state. Restore the link and verify reconnection before continuing.
Does the stale detector change independently of the chart?
Bind the operator indication to a health signal that changes without relying on the historical pen. The strongest design combines remote tag quality with an Edge-generated heartbeat. The heartbeat must change continuously at the source; a constant Boolean can remain cached and cannot prove freshness.
At the Central Gateway, store the time when the heartbeat last changed. Calculate age as:
heartbeat_age = current_time - last_heartbeat_changeDeclare the source stale when the age exceeds a site-selected timeout or when the remote tag quality is not good. Select the timeout above the normal heartbeat update interval plus expected network and processing variation. The evidence supplies no heartbeat interval or permitted detection delay, so derive both from the process hazard and measured network behavior.
| Setting | Too short | Too long | Selection basis |
|---|---|---|---|
| Heartbeat update interval | Unnecessary traffic and transient alarms | Slowest possible detection increases | Required fault-detection response |
| Stale timeout | Healthy jitter causes false stale states | Operators see old data for longer | Update interval plus measured variation |
| Recovery condition | A single update may clear too quickly | Valid data remains marked stale | Documented operating requirement |
Check: Stop heartbeat changes while leaving the view open. The stale state must assert after the configured timeout even though the Power Chart still contains its last sample.
How should the Power Chart show the outage?
Use the stale state to control a visible overlay, banner, status label, or pen-area treatment. Display the affected Edge node and the last update time. A frozen trend must never be the sole outage indication.
If the chart configuration provides a documented option to break lines across missing or bad-quality samples, apply it and test it. That option improves historical interpretation after data resumes, but it does not replace a live stale alarm during the outage. A line-break rule also needs an actual gap or bad-quality record; silence alone may leave the last segment visible.
- Create one health state for each Edge source used by the chart.
- Drive it from Gateway Network status, remote-tag quality, and heartbeat age as required by the application.
- Bind the state to an unmistakable stale-data indication over or beside the chart.
- Show the last successful update time so the operator can judge data age.
- Keep the warning active until the connection and fresh source updates have returned.
Check: Disconnect one Edge node while leaving other nodes online. Only pens and status indications associated with that node should enter the stale state.
How is the complete path verified?
Test the same failure modes the installation must detect. Keep the Perspective session open throughout each test so retained chart data cannot hide the transition.
- Record a changing process value, good remote-tag quality, connected Gateway Network state, advancing heartbeat, and current historical samples.
- Interrupt the network path to the Edge node. Confirm connection status changes, remote data leaves good quality or the heartbeat expires, and the stale indication appears.
- Confirm the chart does not present its retained final value without the stale warning.
- Restore the path. Confirm the gateway reconnects, the heartbeat advances, live quality returns, and new historian records appear.
- Repeat with acquisition or historian logging stopped while the network remains connected. Verify that the health design distinguishes transport, acquisition, and history failures.
Check: Compare the timestamp of the newest chart sample with the live heartbeat and connection state. The display passes only when old samples are visibly marked stale and the warning clears after a genuinely new source update.
FAQ
Why does Ignition Power Chart keep showing a flat line after disconnection?
The component retains and renders the last historical records it received. No new sample does not automatically create a bad-quality point or a visible disconnection state.
Why does good historical data not prove the Edge node is online?
Historical samples describe earlier acquisition, while Gateway Network status, live tag quality, and a changing heartbeat describe current availability. Monitor at least one current health signal separately from the pen data.
How do I verify that the stale-data warning works?
Open the chart with a changing value, interrupt the Edge connection, and verify that the warning appears when connection status or heartbeat freshness fails. Restore the link and clear the warning only after the heartbeat advances and a new source sample arrives.