IMDTouchKnowledge BaseShop main store

Troubleshooting

Why WinCC 8.0 Runtime Deactivates After a V7.5 Migration

HMI and touchscreens Published 7 min read

Short answer

The pop-up means an unhandled exception terminated a runtime component, so the whole project leaves runtime. Read the first exception in PDRTSh.log at the stop timestamp, then activate a blank project on the same machine: if it also stops, repair the installation or Windows build against the WinCC 8.0 Update 7 readme; if it runs, bisect the migrated content.

On this page
  1. Why does WinCC 8.0 runtime deactivate by itself after a V7.5 migration?
  2. Which entry in PDRTSh.log points to the cause?
  3. What do the deactivation timestamps reveal?
  4. How do I check the installation and the Windows build?
  5. How does a blank project split the fault?
  6. How do I bisect the migrated project?
  7. How do I repair the failing side?
  8. How do I verify the fix holds?
  9. Questions asked

WinCC 8.0 runtime that deactivates 3–4 times a day after a V7.5 migration is being terminated by an unhandled exception: an error raised inside a runtime component that no handler catches, so the hosting process dies and the project leaves runtime. Split the fault before you change anything. Read the first exception in PDRTSh.log at the exact deactivation timestamp, then activate a new, blank project on the same machine. If the blank project stops with the same pop-up, the installation or the Windows build is the fault, and you repair it against the compatibility statement in the WinCC 8.0 Update 7 readme. If the blank project runs clean, the migrated content is the fault, and you bisect it.

Why does WinCC 8.0 runtime deactivate by itself after a V7.5 migration? #

Because an unhandled exception terminates the process hosting the runtime, and WinCC then takes the entire project out of runtime — the pop-up is the death notice, not the cause.

The pop-up reads "An unhandled exception was raised. Look at PDRTSh.log for detail", and it carries no component name; the log does. Manual reactivation restarts the failed processes and clears the stopped state, but the defect that raised the exception is untouched, so the next occurrence of the trigger stops runtime again.

A V7.5 backup restored under 8.0 carries content built for the older version: global scripts, C and VBS actions, user archives, and any third-party controls or custom DLLs. Any one of them can raise under the 8.0 components. So can a damaged installation or an unsupported Windows build, which is why the blank-project test, not guesswork, picks the branch.

This is not the quieter migration fault where changes never reach the screen until a rebuild — why WinCC flexible edits do not show in runtime covers that one — a deactivation drops the whole station, screens and all.

Which entry in PDRTSh.log points to the cause? #

The first exception written at the deactivation timestamp, not the last line in the file — every entry after the first exception is shutdown cascade, and chasing the tail sends you after consequences.

  1. Copy PDRTSh.log off the machine before the next stop. Find it with a file search under the WinCC installation and the project directories; a restart or log rotation overwrites it.
  2. Note the wall-clock time of the deactivation pop-up.
  3. Open the log, go to that time, and read the first exception in the sequence.
  4. Record the module name, the exception type, and the call-stack fragment.
  5. Open the Windows Event Viewer Application log at the same timestamp; if the process crashed rather than shut down cleanly, it names a faulting module.

A module inside the WinCC installation points to the installation branch. A module belonging to your scripts, a third-party OCX, or a custom DLL points to the project branch. No entry at the deactivation time means you are reading the wrong file or the wrong node — repeat the search on the machine actually running runtime.

Graphics and VBS faults leave their own trace, and resolving unhandled exception entries in PDLRTSEH.LOG follows the same first-entry rule on that file. Script-side stops land in a different place, and resolving a WinCC V7 unhandled exception written to Script.log shows the same method against an older version.

What do the deactivation timestamps reveal? #

Spacing that varies by more than an hour between stops rules out a fixed-period job and points to an event trigger — the worked example below runs 5 h 38 min down to 4 h 25 min between stops.

Worked example, with illustrative stop times for a fault reported at 3–4 stops per day:

  • Stop times: 06:14, 11:52, 16:40, 21:05 — four stops.
  • 11:52 − 06:14 = 5 h 38 min
  • 16:40 − 11:52 = 4 h 48 min
  • 21:05 − 16:40 = 4 h 25 min
  • Fixed-period reference: 24 h ÷ 4 stops = 6 h 00 min; at 3 stops per day a timer would sit at 24 h ÷ 3 = 8 h 00 min.
  • Spread across the gaps: 5 h 38 min − 4 h 25 min = 1 h 13 min.

No measured gap sits on 6 h 00 min and the spread is over an hour, so no timer owns these stops. Find the log entry that precedes every stop: a scheduled archive or report run, a shift-change action, a specific operator command. If your measured gaps agree within a few minutes, chase the scheduler that owns that period instead.

How do I check the installation and the Windows build? #

Against the compatibility statement in the WinCC 8.0 Update 7 update readme — Siemens issues the 8.0 updates as cumulative and puts the version-compatibility list in that readme, so the V7.5 requirements do not carry over.

  1. Compare the installed Windows edition and build against the readme's supported list.
  2. Open Programs and Features: WinCC 8.0 present, Update 7 listed, and no leftover V7.5 components on the same machine.
  3. Confirm there is no pending reboot and no installer entry stuck in a failed state.

Any mismatch here is a finding on its own. Correct it before touching the project, because test results from an unsupported platform tell you nothing about the project.

How does a blank project split the fault? #

A new project with no tags, pictures, scripts, or archives exercises only the WinCC 8.0 runtime core on that exact machine, so whether it survives assigns the fault to one side.

Blank project result Conclusion Next action
Deactivates with the same pop-up and log signature Installation or platform fault Repair or reinstall, then repeat the test
Deactivates with a different signature Two faults, or a platform-level resource problem Resolve this signature first, then repeat
Runs clean Fault is in the migrated content Bisect the project

Take the station dark only when the process tolerates it: while runtime is down, that screen gives the operator no alarms and no control.

How do I bisect the migrated project? #

Start from the component the log entry named and disable the optional runtime components in halves — the component whose removal stops the stops, and whose return restarts them, is the root cause.

  1. Copy the migrated project and keep the V7.5 backup untouched.
  2. If the log names a script or function, open it and check for unguarded object references, uninitialized variables, and calls into libraries built for V7.5.
  3. If the log names no specific item, disable the optional startup components — alarm logging, tag logging, user archives, global scripts — in halves, reactivating after each halving.
  4. Recompile every C function and action in the 8.0 project, and clear each compile warning before running.
  5. Replace third-party controls and custom DLLs with versions released for WinCC 8.0.

The expected reading: the deactivation disappears when one component is disabled and returns when you re-enable it. A stop that persists with everything optional disabled sends you back to the installation branch.

How do I repair the failing side? #

On the installation branch, repair or reinstall WinCC 8.0, re-apply Update 7, reboot, and repeat the blank project; on the project branch, correct the identified component and restore everything you disabled — a disabled archive is a hidden outage, not a fix.

If the readme rules the installed Windows build out, the runtime moves to a machine on a supported build, and a station that sits at the machine itself is the job for industrial panel PCs. Repeat the blank-project test after any installation repair before you trust the migrated project on that machine again.

How do I verify the fix holds? #

Activate the full migrated project with every component restored and hold runtime up through one complete production cycle — a clean 24-hour run spans at least three of the stops you were seeing, so anything shorter proves little.

  1. Record the activation start time.
  2. Runtime state stays active with no operator intervention.
  3. PDRTSh.log holds zero exception entries after the start time.
  4. The Event Viewer Application log shows no faulting-module entries for WinCC processes in that window.
  5. Keep the run going through the trigger the timestamp analysis named: the scheduled archive, report, or shift-change event.

Nothing from the test runs carries over to production: the corrected project reaches operators only as a fresh activation, for the same reason why an HMI edit needs a new runtime download before anyone sees it.

Questions asked

Can I use automatic reactivation as a workaround while I investigate?

You can bridge with it, but know the cost: while runtime is down, alarms and tag logging stop, so the historian carries a hole for every gap and the operator works blind from that screen. Treat it as a bridge with a fixed repair date, never as the fix, because the exception trigger stays untouched.

Why did the same project run without stops under V7.5?

Because V7.5 never exercised the failing path the way 8.0 does. The migration recompiles scripts, archives, and controls into the 8.0 components, and content that the older runtime silently tolerated can raise an exception there. That is why the blank project and the log entry, not the project's V7.5 history, decide where the fault sits.

How long does the blank project have to run to be conclusive?

At least 24 hours. With stops arriving 3 to 4 times a day, the longest quiet gap you should ever see is about 8 hours, so a full day without a stop is a real result, not luck. A shorter run can land entirely inside one quiet gap and prove nothing about the runtime core.

Is applying a newer WinCC 8.0 update enough on its own?

Only if the blank project also stops, which puts you on the installation branch. The 8.0 updates are cumulative, so bringing the installation current belongs to that repair. A project-side fault survives every update until the named component is corrected, and a clean blank project tells you no update will fix it.

Does the migration damage the V7.5 project I restored from?

No, as long as you keep the backup untouched and work on the copy, the migration leaves the source alone. That also gives you the retreat path: if the project branch stalls, repair the platform first, then run the migration again from the same backup and bisect the fresh result.

Keep reading

  • How-to

    Migrate a PanelView Plus 6 project from ME 5.0 to 10.00

    Yes, the application can move from FactoryTalk View ME 5.0 to 10.00, but the runtime file itself is never version-converted. Confirm the terminal is a PanelView Plus 6 700-1500 or Compact 1000, restore an editable application from the .apa archive or a restorable .mer, convert it in FactoryTalk View Studio ME 10.00, build a new .mer at the 10.00 runtime version, and download it after flashing firmware 10.00-20180411. Part 11 compliance is a separate validation job.

    HMI and touchscreens

  • Explainer

    Start a published WebIQ HMI project without Runtime Manager

    No current WebIQ version starts a project automatically after publishing: start and stop live in Runtime Manager, and Designer only shows the running state. During development, point panel devices at the Designer preview server on port 10124 instead; publish and start only for tests that need the ctrlX reverse proxy.

    HMI and touchscreens

  • How-to

    How to migrate a WinCC flexible project to TIA Portal

    One TIA Portal installation replaces both STEP 7 Classic and WinCC flexible, but the two project types move differently: the STEP 7 V5.x project is imported, and the WinCC flexible 2008 project is migrated through an intermediate release because the newest TIA Portal versions no longer offer that migration. Verify the exact S7-300 CPU and panel against the device lists and license STEP 7 and WinCC Comfort separately.

    HMI and touchscreens

Migrating a WinCC 7.0 Project From Windows XP to Windows 7