IMDTouchKnowledge BaseShop main store

Troubleshooting

Why an HMI Trend Shows No Curves but the PC Web Page Does

HMI and touchscreens Published 8 min read

Short answer

The panel's browser requests the live trend window using its own OS clock while the WebIQ server stamps samples in UTC from its host's clock. When the two differ by more than the visible window, the request contains no samples and the chart stays empty. Put the server host, controller and panel on one NTP source and the curves return.

On this page
  1. Is the panel failing to render the trend component?
  2. Why does a clock difference empty the trend on the panel?
  3. Which WebIQ version warns about the time difference?
  4. How do you prove the clocks differ?
  5. How do you fix it permanently with NTP?
  6. What can you do when the network has no NTP server?
  7. What else blanks a trend when the clocks match?
  8. Questions asked

A web HMI trend that scrolls its Y-axis on the panel but draws no curves, while the same page in a PC browser shows data, is a time mismatch between the system that records the samples and the panel that displays them. WebIQ stamps every sample in UTC using the clock of the server host, and the panel's browser requests the live window relative to its own OS clock. When those two clocks disagree by more than the visible window, the request lands where no samples exist and the chart stays empty even though recording continues. Put the server host, the controller and every panel on one time source and the curves return. The steps below rule out the panel's rendering first, then find the offset, prove it and remove it permanently.

Is the panel failing to render the trend component? #

No — the scrolling Y-axis shows the panel's browser renders the trend component correctly, which rules out resolution, screen size and unsupported web components, and points at the data request itself.

The PC browser and the panel load the same project page. The PC draws curves, so the page, the trend object, the tag binding, the driver link to the controller and the recording all work. The panel draws the trend frame and scrolls the axis, so its browser engine handles the component. Editing pixel values in the browser inspector without any effect removes resolution and layout as causes.

What you see What it rules out What it points to
Variables update in simulation and in the PC browser Tag addressing, driver, controller link, trend setup The panel as a client
The Y-axis scrolls on the panel Unsupported component or a failing render engine A requested time window that contains no samples
Pixel edits in the inspector change nothing Resolution or layout Data or time, not geometry

Open the same page on the PC and the panel at the same moment. Only the panel blank: continue below. Both blank: the fault is on the recording side, and the clock checks still apply to the server host and the controller.

Why does a clock difference empty the trend on the panel? #

Because the live trend asks for samples relative to the browser's clock, and the samples are stamped by the server's clock, so a large enough offset makes the requested window fall where no samples exist.

A web HMI is a distributed system: the WebIQ Server records the trend data, the browser on the panel displays it, and each machine follows the time set in its own operating system. Time zones do not matter, because every sample carries a UTC timestamp. The absolute time does: a wrong clock or a wrong date breaks the display while recording continues.

A worked example, using illustrative clock readings:

  • Server host clock: 14:00:00 UTC, recording continuously. The newest fifteen minutes of samples run 13:45:00 to 14:00:00 UTC.
  • Panel clock: 13:40:00 UTC, twenty minutes behind, because 13:40:00 + 20 min = 14:00:00.
  • Live trend window: the last fifteen minutes.
  • The panel therefore requests 13:40:00 − 15 min = 13:25:00 through 13:40:00.
  • The overlap of [13:25:00, 13:40:00] with [13:45:00, 14:00:00] is zero, since 13:40:00 is earlier than 13:45:00. The chart is empty while the axis keeps scrolling.

Run the same reading with the panel only five minutes behind: its clock reads 13:55:00, the request becomes 13:40:00 through 13:55:00, the overlap is 13:45:00 through 13:55:00, and ten of the fifteen minutes draw as a curve. The rule that falls out of the arithmetic: an offset at or above the live window empties the chart, an offset below it shortens the curve.

Which WebIQ version warns about the time difference? #

From WebIQ 2.17 onward the trend object displays a warning when the times differ too significantly, and the same warning may also be present in a late 2.16.x release; on 2.16.x builds without it the trend stays silently empty.

Read the exact version string of the installation before anything else and keep it with the service ticket.

Version in use Behavior under a clock difference What you do
2.16.x without the warning Empty trend, no message Compare the clocks by hand, next section
2.17 and newer, possibly a late 2.16.x release Trend shows a time-difference warning Treat the warning as confirmation and fix the time source

Upgrading adds the warning but does not repair the offset; only matching clocks bring the curves back. One caution belongs here: on a build that fails silently, an operator watching a blank trend for a process limit gets no indication the evidence is missing, so treat the empty chart itself as a fault to clear. If the case goes to support, collect the exact version, the connect.log with the system in trace mode, and the times read from both systems; trace mode serves the case file only, because the clock comparison decides the fix.

How do you prove the clocks differ? #

Read the OS date and time on the panel, on the WebIQ Server host and on the controller at the same moment, compare all three in UTC, and the one that disagrees by more than the live window is the cause.

  • The panel's own OS date and time settings: the clock its browser uses for the live request.
  • The WebIQ Server host: the clock that stamps the recorded samples.
  • The controller, read in its own web interface: it matters when the project or the logged variables depend on controller time.

Compare all three against one trusted reference, in UTC where the devices show different zones. Then run the confirmation test: set the panel clock by hand to the server time once and reload the page. Curves at the first load confirm the diagnosis and you move to the permanent fix. Power-cycle the panel and read its clock again: a panel that falls back to a factory-default date on power loss reproduces the fault at the next start, and that panel needs a working time source at every start, not a one-time correction.

How do you fix it permanently with NTP? #

Point the WebIQ Server host, the controller and every panel at one Network Time Protocol source, ideally the correct time, because one time base restores the curves and keeps every device's logs alignable during fault analysis.

  1. Choose the source: an external NTP server where the machine network reaches one.
  2. On a closed machine network, run the NTP service on a local device — the controller or a compact computer in the cabinet — and use that address.
  3. Enter the address in the time settings of the panel, the controller and the server host.
  4. Restart the sync service or reboot each device where the OS requires it, and confirm each reports a synchronized state.
  5. Read all three clocks again: they must agree within a second. Power-cycle the panel and read once more before closing the ticket.

Zones can differ from device to device; only the absolute time must agree. Panels also fail to sync for panel-side reasons of their own — troubleshooting SNTP clock sync on a Windows CE panel works through that fault path. Where no time server can exist, writing the clock from a controller on a schedule is a workable pattern, and setting HMI clock time from PC and PLC daily covers the scheduling, as long as the shared time also matches the operator's watch.

What can you do when the network has no NTP server? #

Run the NTP service on a local device, and only where that is impossible use a startup UI-Actions script that sets the controller time to the panel's time through the controller's REST interface.

The script carries real limits:

  • It works with one HMI at a time.
  • It synchronizes only the HMI and the controller; the shared time can be wrong against the true time.
  • It corrects a small offset, not a panel that starts on a factory-default date.
  • It needs no extra hardware and is fast to implement; a typical version calls the REST interface with the axios library.

Build it against the REST documentation and authentication settings of the controller in front of you, and never carry credentials or endpoints over from an example. Decide it this way: NTP whenever more than one operator device exists or whenever alarm and event logs must carry the true time; the script only on a single-panel machine where true time does not matter. The acceptance test is a cold start of controller and panel, then opening the trend page without touching any clock — the curves must be present at first load.

What else blanks a trend when the clocks match? #

When all three clocks agree, move to the recording side: check that the trend is actually collecting, that its retention covers the window you are requesting, and that the same page is still live on the PC at that moment.

Re-run the comparison from the first section with the clocks now matched. If the PC page is also blank, the recording is the suspect, not the display. A requested window that reaches further back than the recorder keeps samples is empty for a data reason rather than a clock reason; verifying HMI historical trend retention before committing covers that check.

What you find The decision
Only the panel is blank; clocks differ by more than the live window Correct the panel clock, then put every device on one NTP source
The trend shows a time-difference warning Same fix; the warning only reports the offset
Closed network with a single panel NTP service on a local device; the startup UI-Actions script as the last resort
Clocks match and the PC page is blank too Recording side: collection stopped, or retention shorter than the window

Questions asked

Does the panel's time zone setting affect the WebIQ trend?

No. Every sample is recorded with a UTC timestamp, so the zone the panel displays does not decide what appears. The absolute time does: a panel set to the wrong zone but keeping the correct UTC time draws the curve, while a panel in the correct zone with a wrong clock stays blank.

How far apart can the clocks be before the trend goes blank?

Compare the offset with the live window. Twenty minutes of offset against a fifteen-minute window requests a range that ends before the newest sample starts, so the chart is empty; five minutes of offset leaves ten of the fifteen minutes drawn. No numeric tolerance is published, so the window is the yardstick, and anything visible reading two clocks side by side is worth removing.

What does a WebIQ support case for this symptom need?

Three items: the exact version string of the WebIQ installation, the connect.log collected with the system in trace mode, and the date and time read from the server host and the panel. Read and write down the times before correcting anything, because the first manual fix destroys the evidence that confirms the diagnosis.

Is it worth upgrading from 2.16.x just to get the time-difference warning?

Yes, but not as the fix. The warning turns a silent blank chart into a visible message, which cuts the next occurrence from a fault hunt to a clock correction, while the curves themselves still only return when the clocks match. Correct the time first and schedule the upgrade with other maintenance.

Can the startup script replace NTP on a line with several panels?

No. The script sets the controller time from one panel's clock, so a second panel with its own offset breaks the agreement again. Use it only where exactly one HMI faces the controller and nobody needs the true time on alarm and event logs; otherwise a time server, local or external, is the fix.

Keep reading

  • Troubleshooting

    Why HMI Data Logs Export Bool True as -1

    A Comfort Panel USB data log shows -1 for a true Boolean because the export serializes the all-ones bit pattern as a signed value. Assign the literals 1 and 0 conditionally into a separate numeric tag and log that tag; a direct copy into an Int or UInt keeps the -1. Verify the change in the raw CSV with a text editor.

    HMI and touchscreens

  • How-to

    Keep a PanelView Plus Trend Logging After Screen Changes

    A PanelView Plus Trend object only records while its screen is visible. Configure a Data Log model to start at power-up, show it through an XY trend with pen 2 as the X axis, and give the log a bounded record limit so it overwrites the oldest records instead of growing without bound.

    HMI and touchscreens

  • How-to

    Verifying HMI Historical Trend Retention Before Committing

    An 8-hour trend sampled every 5 minutes spans 96 intervals — 97 stored points if both endpoints are kept — so the panel needs a historical buffer of at least 97 samples per tag that repopulates the trend after screen changes. Verify the buffer depth, the interval setting, and restoration behavior on the bench before you build the screens.

    HMI and touchscreens