brands · South Africa
Magelis HMI Tag Import: XVM Links and Interface Tests
Link Control Expert symbols to Magelis HMIs in Vijeo Designer, review XVM updates and data dictionaries, and test tag identity, types, scaling and changes.
Magelis HMI tag import from Control Expert is best understood as linking the PLC's symbol information to an HMI project and then verifying the resulting interface. A successful file operation does not prove that the screen reads the intended equipment, interprets the value correctly or preserves its alarm and logging configuration after a change.
This guide uses Schneider Electric's Vijeo Designer and Control Expert linking documentation, checked on 13 September 2026. Confirm the exact panel reference, engineering release and supported communication driver for your project. The worked manifest exercise is fictional and does not reproduce the native XVM file schema or claim a tested Magelis download.
We operate PLC Simulation Software and may benefit from product links. Its general HMI and program-testing resources do not imply native Vijeo Designer project support, Schneider certification or access to the vendor's engineering licences.
Identify the panel, driver and source application
Record the full HMI reference and the software used to maintain it. The word Magelis is not enough to establish a universal target, transfer method or project format. Keep the controller reference, firmware and Control Expert release with that record so another engineer can identify the intended combination.
For the data path, state which PLC the HMI must communicate with and which driver the project actually uses. Do not select a protocol solely from the age of the controller or the presence of a USB connector. A physical interface, an engineering connection and the HMI's application communication path are different facts.
Before exporting symbols, identify the source project and its relationship to the intended controller application. Give the export a clear name and keep it with a source-version note. An old export with familiar variable names can still describe an obsolete structure or address context.
The Control Expert project guide explains where data definitions fit within the application. The communication troubleshooting guide helps separate driver and connection checks from the later question of whether the values have the correct meaning.
Use Link Variables for the documented XVM workflow
Schneider's Control Expert and Vijeo Designer linking guide describes exporting from Variables & FB instances to an XVM file, then using Link Variables from the Vijeo Designer Variables node to select it. The guide also identifies the Control Expert Data dictionary setting under Project Settings and PLC embedded data, and describes updating the link after relevant changes.
Use the corresponding commands in the supported installed environment and inspect the variables offered by the link. Select the application interface deliberately. A name prefix can help organise the list, but it does not establish data type, units, quality behaviour or permission to write a value.
The Schneider symbol-import FAQ also describes an STU linking option when Control Expert is installed on the same computer as Vijeo Designer. Treat that as a distinct documented arrangement with dependencies. Do not assume that copying an STU file to any workstation makes it a universal interchange format.
Linking symbols and importing a variable-configuration export are different operations. Keep the purpose of each file clear in the project notes. A configuration export may help preserve HMI settings, while a refreshed symbol link establishes the connection to the appropriate PLC definitions.
Define the HMI interface before binding objects
For every displayed value, record the qualified source name, expected data type, units, scale, access direction and intended screen use. Distinguish a raw count from an engineering value. A numeric field displaying two decimal places changes presentation; it does not by itself convert a raw integer into the intended unit.
For a command or setting, document who accepts the request and what feedback confirms the result. A writable variable is not automatically an appropriate operator control. The worked exercise below uses read-only displays to keep the assessment focused on identity, interpretation and change detection.
Keep the same level of care for Boolean values. A field called Running could represent requested operation, confirmed feedback or a combined application status. The name alone cannot establish which meaning applies. Put the intended meaning in the interface record and verify it against the controller logic.
Map each screen object to the corresponding HMI variable and source equipment. Use deliberately different test values for different controllers or channels. If every source holds zero, a wrong binding can look correct throughout a demonstration.
The SCADA learning guide develops the wider distinction between source data, interpretation and presentation. Use that structure to make the tag review understandable to both the PLC programmer and the person assessing the HMI.
Data dictionary and structure changes need explicit review
Schneider's broken symbol-link FAQ identifies two relevant problems: same-named user data types with different members across PLCs, and deleted or renamed members in a previously linked structure. It also explains the role of Data dictionary when names and structures remain unchanged. That qualification matters: a dictionary is not proof that arbitrary structural changes need no review.
The same FAQ describes a recovery procedure that preserves HMI configuration and re-establishes links. Review the full procedure for the affected environment before using it. Deleting variables or types without preserving the dependent settings can create a second problem while investigating the first.
For a controlled training project, make a copy before testing a structural change. Record the original link, screen bindings and relevant alarm or logging settings. After updating, compare the resulting configuration and test the affected objects. A completed update dialog is evidence of one operation, not an acceptance test for the entire HMI.
Keep a list of dependants for each changed symbol. A variable can be used in more than one screen, a script, an alarm and a trend. Checking only the first numeric display leaves those other uses unverified.
Worked exercise: compare a display manifest
Define a fictional manifest with one row per HMI display input. Each row has a source identifier, symbol name, type, unit, scale and access mode. The source identifier is A or B, and the qualified key combines that identifier with the symbol name. The exercise treats spelling and case as significant.
The expected interface has three rows. A.Level is an unsigned integer expressed in tenths of a percent, so its display scale is 0.1 and its unit is percent. A.Running is Boolean, has no engineering unit and uses a scale of one. B.Level is a real-valued percentage with a scale of one. All three are read-only for this exercise.
| Qualified key | Type label | Unit | Display scale | Access |
|---|---|---|---|---|
| A.Level | UINT | percent | 0.1 | Read-only |
| A.Running | BOOL | none | 1 | Read-only |
| B.Level | REAL | percent | 1 | Read-only |
These rows are an educational contract, not literal XVM syntax. The scale, unit and access columns are review information; this article does not claim that the native symbol export automatically supplies or enforces them. Verify which information comes from the PLC, HMI configuration or separate application documentation.
The comparison has two stages. First validate that every supplied row has the required fields and that each qualified key appears only once. Then compare the complete key set and the attributes of matching rows with the expected manifest. A malformed or duplicate input is an invalid manifest; a well-formed but different interface produces a difference report.
Detect missing, unexpected and changed symbols
Start with an exact copy of the expected three rows in a different order. The comparison should pass because row ordering is not part of this fictional contract. It should not report a change merely because an export or review table has been sorted differently.
Now rename A.Level to A.TankLevel without changing the screen requirement. The report should show A.Level missing and A.TankLevel unexpected. The model does not guess that the two are the same signal. A deliberate rename needs an explicit mapping decision and a review of the dependent objects.
Next, change A.Level's type from UINT to INT while preserving its key. The report should identify a type change. Change its scale from 0.1 to one and it should identify a scale change. Change its access from Read-only to Read-write and it should identify an access change. Those differences have different implications and should not be combined into a vague “tag changed” message.
Add a second A.Level row and the manifest is invalid because the qualified identity is ambiguous. However, A.Level and B.Level are not duplicates in this model: their source identifiers differ. A checker that compares only the short symbol name would incorrectly reject the expected interface.
Remove B.Level and the report should identify it as missing. Add B.Pressure with its own valid attributes and the report should identify an unexpected symbol. An extra symbol may be intentional, but the expected contract must be reviewed rather than silently accepting any expanded interface.
The local reference checker verifies these cases. It establishes the behaviour of this fictional manifest comparison, not the native Vijeo importer's duplicate handling or file validation.
Use sentinel values to test interpretation
For A.Level, supply a fictional raw value of 375. Applying the specified scale of 0.1 gives 37.5 percent. For B.Level, supply a real value of 62.25, which remains 62.25 percent. Set A.Running to true. The expected screen values are therefore deliberately different and easy to identify.
If A.Level's scale is incorrectly changed to one, it displays 375 percent. If the two level bindings are crossed, the displays no longer match their expected source values. A common raw value of zero would not expose either error as clearly, which is why the assessment includes nonzero sentinel values.
Test zero and the upper boundary separately. Under the fictional A.Level contract, raw values zero and 1,000 display zero and 100 percent. Values outside that declared raw range should be treated according to an explicit invalid-data policy, not merely formatted as a plausible percentage.
A scale of 0.1 is not automatically correct because the controller tag is an integer. It is correct only because the application contract defines tenths of a percent. Another integer could represent pulses, a recipe identifier or a packed bit field. Verify meaning before applying a conversion.
The data conversion reference helps distinguish numeric type from engineering interpretation. Keep the original raw value visible in the test evidence so a reviewer can tell whether a wrong result came from the source, the scale or the binding.
Preserve the HMI configuration around a symbol update
Before a controlled update, record the relevant variable properties, screen bindings, alarm configuration, logging configuration and any scripts that reference the changed symbols. Decide which settings should remain identical and which changes are intentional. This gives the comparison a clear acceptance target.
After updating the symbol link, repeat the sentinel tests and review the dependent settings. A screen may show the correct current value while an alarm still uses an obsolete variable. Conversely, a variable may be correctly linked while a screen object points to an old HMI name.
If a rename is intentional, document the old and new qualified identities and the dependent objects that were changed. Do not rely on a guessed similarity between names. A.Level and A.TankLevel might refer to the same measurement, or they might represent different stages of a process.
Keep the original export and the updated export clearly separated. Record which one the HMI project links to. A successful test on one workstation can be difficult to reproduce if another workstation resolves a similarly named file from a different folder.
For an assessment, ask the learner to explain why a type change, a scale change and a source change need different test evidence. That answer demonstrates an understanding of the interface rather than familiarity with one import dialog.
Investigate errors at the correct boundary
If the symbol file cannot be opened, inspect the file identity, software dependencies and actual diagnostic message before changing PLC communications. Schneider's February 2026 symbol-file FAQ describes one documented cause involving SRCSDK component registration. It is a specific troubleshooting route, not proof that every symbol-file error has that cause.
If symbols link but values have poor quality, investigate the source application, structural changes, equipment assignment and applicable dictionary behaviour. If quality is good but the number is wrong, compare source value, type and scale. These are different failure boundaries and should not all be treated as a network fault.
For M580S projects, the Schneider process-variable import FAQ describes a memory-space issue and an automatically generated global XVM route for the documented situation. Check its applicability to the actual versions and project. Do not assume that every export node exposes every memory area in the same way.
Keep native safety-related configuration outside a general HMI tutorial's inferred scope. A display link does not establish authority to modify a protected or safety-related application. The practical learning objective here is to recognise the documented boundary and obtain the correct project-specific method.
Create a small, reproducible HMI assessment
Use a controlled project with the three fictional display inputs and a separate manifest. Ask the learner to establish the expected values, link the appropriate symbols and demonstrate each display. Then introduce one planned change: rename a symbol, change its scale or connect the wrong source.
Require a difference report and an explanation of the resulting display. The learner should identify the cause from evidence rather than simply restoring the last known-good project. Keep a record of the actual HMI configuration and which part of the exercise was tested only in the reference model.
Include a second pass after navigation or reopening the project. Confirm the link still resolves to the intended export and the relevant screen objects retain their bindings. A result that depends on an unrecorded local file or temporary setting is not yet reproducible.
For a portfolio, include the environment, manifest, expected sentinel values, observed results and one corrected defect. Clearly label fictional data and avoid presenting conceptual illustrations as screenshots of native Vijeo Designer. The images on this page are learning illustrations, not evidence from a customer installation.
Plan Magelis and Vijeo Designer learning in South Africa
When comparing courses, provide the exact panel reference and ask which software release and communication driver the practical sessions use. Request individual access to a suitable project and an assessment that includes symbol changes, data interpretation and dependent HMI settings. A general PLC course title does not confirm that depth of HMI work.
Use the Schneider training hub to connect this topic with controller learning. If the work also involves updating an old application, the Unity Pro to Control Expert migration guide explains how to separate project files, operating data and regression evidence.
For general preparation, inspect the product's HMI simulator learning page and PLC program testing resources. Check the current scope before buying. These links support transferable reasoning and do not promise native XVM import or Magelis project execution.
Training managers can use the training-centre evaluation guide to assess whether a learner can explain a changed interface and preserve relevant configuration. A useful result is a reproducible demonstration with clear limitations, not just a claim that the import succeeded.
Magelis tag-import questions
Is Link Variables the same as importing a variable CSV?
The documented symbol-link workflow uses Link Variables to connect the PLC symbol information. A variable-configuration export serves a different purpose, such as preserving HMI settings during a controlled recovery procedure. Identify what each file contains and why it is being used.
Does Data dictionary remove the need to review renamed members?
No. The checked guidance distinguishes unchanged names and structures from structural edits that can break address context. Review affected symbols and their dependants after a rename or deletion, and apply the matching vendor procedure where required.
Why does the worked manifest include the source identifier?
A.Level and B.Level are distinct signals even though their short names match. The source identifier prevents the review from treating them as duplicates or silently exchanging them. The native project still needs verified equipment and object bindings.
Does formatting a value to two decimals perform scaling?
No. Presentation precision and engineering conversion are separate decisions. The fictional raw value 375 becomes 37.5 percent because the contract explicitly applies a scale of 0.1, not because the screen has decimal places.
What proves that an updated link is correct?
Use the intended source file, compare changed symbols and dependent settings, then test distinct source values in the native project. The fictional manifest checker provides one form of expected evidence; it does not replace runtime verification of the actual HMI.
