Explainer
Why HMI simulation cannot talk to a simulated PLC
Short answer
EasyBuilder's PC simulation communicates with a real PLC, but SimuLab runs a LogicLab program without any communications service, so two PC simulators never exchange data. Integrated testing needs at least one physical CPU module; until then, test the screens and the controller logic separately.
On this page
- Can a simulated HMI exchange live data with a simulated PLC?
- Why do the two simulators stay in separate data spaces?
- What can you still test with no hardware at all?
- Do any vendors support simulator-to-simulator communication?
- How do you structure the test with one physical CPU?
- How do you verify the connection once both sides run?
- Is the fault in the tag or in the screen binding?
- Questions asked
No, and the reason is structural, not a setting you missed. EasyBuilder's PC simulation communicates with a real PLC, while SimuLab runs a LogicLab program without managing any communication, so an HMI simulator and a controller simulator on the same PC occupy two separate data spaces. The smallest setup that proves a command travelling from screen to controller and back is one physical CPU module. Below: what zero hardware can still test, why the simulators never meet, which vendors couple their simulators, and how to run and verify the one-module test.
Can a simulated HMI exchange live data with a simulated PLC? #
No. An EasyBuilder PC simulation communicates with a real PLC, while a LogicLab program running in SimuLab provides no communications service, so the two simulators never exchange a single value; the minimum hardware for an integrated test is one physical CPU module.
The split is documented tool behavior, not a missing setting. EasyBuilder's PC simulation is built to carry the real communication with a physical PLC through the project's selected driver. SimuLab exists to execute a LogicLab program on the PC, and it does not manage the PLC communication this connection needs. Running the two at once therefore joins nothing, and correcting tag addresses cannot conjure an endpoint the controller simulator does not implement. Zero hardware gives you two separate tests; one SlimLine CPU module gives you the integrated one.
Why do the two simulators stay in separate data spaces? #
Because an HMI driver talks a protocol to a controller endpoint, and SimuLab executes the LogicLab application without exposing that endpoint; sharing one Windows PC creates no protocol service, no address mapping and no request handling between two applications.
Every PLC driver is a protocol client: it frames requests, waits for answers and maps replies into tags. EasyBuilder Pro, Weintek's HMI engineering software, lists more than 400 PLC drivers, and every one of them targets a controller interface — not the memory of another Windows program. SimuLab executes the LogicLab application and stops there: it does not listen on the interface a driver calls, so from the HMI simulation's point of view the controller simply is not on the network. Windows process isolation does the rest; two programs sharing a desk is not a data path. The test that keeps this honest on any screen object: an HMI-local value animates by itself and proves only the runtime, while a PLC-linked value comes into existence only after a completed driver transaction.
What can you still test with no hardware at all? #
Screen navigation, object configuration and controller logic: open the screens in EasyBuilder's offline simulation with simulated values, and run the LogicLab program in SimuLab to watch the logic execute; what you cannot test, with zero hardware, is any transaction between them.
Offline simulation runs the HMI project on the PC so you can see how it looks and behaves while you develop it: screens open, navigation works, objects show their configuration, and HMI-local values animate. In parallel, SimuLab executes the LogicLab program, so you debug controller logic online with no terminal involved. Two cautions from practice. If the simulation window refuses to start and reports port 8000 already in use, a second simulation instance is holding the port — a local fault unrelated to PLC communication, handled in fixing the offline simulation port 8000 already in use error. And simulation is not the panel in other ways: functions such as user management can behave differently on the physical terminal, as when HMI user management works in simulation but not on the panel shows.
| Test objective | Where it runs | Hardware |
|---|---|---|
| Screen look, navigation, object configuration | EasyBuilder offline simulation on the PC | None |
| Controller logic behavior | LogicLab program in SimuLab | None |
| Operator commands with controller feedback | EasyBuilder PC simulation plus its PLC driver | One SlimLine CPU module |
Use the hardware-free branch for screen development and logic checkout, and label anything shown to a customer as demonstration values, not controller feedback.
Do any vendors support simulator-to-simulator communication? #
Yes, where one vendor builds both ends: Siemens documents a procedure for simulating HMI operator interfaces together with SIMATIC controllers in a coupled PC simulation; EasyBuilder and SimuLab, coming from different vendors, form no such pair.
Siemens publishes a support document on simulating HMI projects in connection with SIMATIC, covering the simulation of one or more HMI operator interfaces together with SIMATIC controllers. That works because one vendor supplies both ends and couples them inside its own engineering environment. EasyBuilder and SimuLab cross a vendor boundary, and nothing in either tool bridges it. The general rule: a simulated HMI talks to whatever answers its driver — normally a physical controller, and a controller simulator only when the HMI vendor documents the pairing. Other ecosystems show the same from the HMI side: simulating a Basic HMI against a real TM172 HVAC controller works precisely because the endpoint is a physical CPU.
How do you structure the test with one physical CPU? #
Run the operator interface in EasyBuilder's PC simulation and load the LogicLab application into the physical SlimLine CPU: the CPU becomes the communications endpoint, so the integrated test needs one module and no physical touchscreen.
- Keep both projects exactly as validated in the independent simulations, then load the intended LogicLab application into the SlimLine CPU.
- In the EasyBuilder device configuration, select the PLC device and driver that match the actual CPU interface, and take every connection value from the controller configuration rather than from simulation defaults.
- Start the EasyBuilder PC simulation and confirm communication is established before walking the full screen sequence.
One caution before the first write test: a real CPU answers with real output wiring, so isolate the outputs or put the machine in a safe state first.
The discipline is ecosystem-independent — the steps in testing HMI simulation against a live PLC are the reference procedure. With Siemens hardware, the same shape of test is walked through in connecting a simulated Simatic HMI to a physical S7-1200 PLC.
How do you verify the connection once both sides run? #
Read the driver's communication status in the running simulation first: a broad failure across many PLC-linked objects points at the driver and connection, while a single dead object with the rest updating points at that tag's binding; a responsive screen proves nothing about the link.
Worked example, with an illustrative variable (your names will differ): declare simTestCounter : INT := 0; in the LogicLab program, increment it once per cycle of a 100 ms task, and point a numeric display object at it in the simulated screen. After 30 s, the arithmetic is 30,000 ms ÷ 100 ms per increment = 300, so the display must read 300 ± 1 to account for scan timing. A display pinned at 0 means the driver transaction never happens; a value that reads correctly and then freezes means the connection or polling dropped; a steady 300 ± 1 proves the path end to end.
| Observation | Fault location | Next check |
|---|---|---|
| Many PLC-linked objects dead together | Shared driver and connection | Device selection, interface, connection settings |
| One object dead while the rest update | That object's binding | Device reference and tag address |
| Controller variable changes online, display does not follow | Mapping or data representation | Data type and word layout |
| Reads follow the controller, a command does nothing | Command path | Command destination and the logic consuming it |
Is the fault in the tag or in the screen binding? #
Separate them by changing one controller value you can predict and watching both ends: if the controller variable does not change, the fault is program conditions or the execution target; if it changes but the screen does not follow, the fault is the device reference, address mapping or data representation.
A tag-access fault concerns the data the screen asks the controller for — its address mapping or representation. A binding fault concerns which device and tag the object actually uses. The trap is an object left on an HMI-local value: it performs flawlessly and never sends a command. Commands add two more traps. A destination on a physical input cannot be written at all, as why an HMI toggle switch cannot write a physical PLC input explains. And a write that lands can still be overwritten by the next logic scan, which is why command and status belong in separate variables.
Decision: with zero hardware, run the two projects separately and call every number on the screen a demonstration value. With one SlimLine CPU module loaded and the EasyBuilder driver pointed at it, the same PC simulation becomes the integrated test, and the counter reading 300 ± 1 after 30 s is its pass criterion.
Questions asked
Would running the two simulators on separate PCs or in virtual machines connect them?
No. The barrier is not the machine boundary but the missing communications service: SimuLab does not expose the protocol endpoint an HMI driver needs, wherever it runs. A second PC, a virtual machine or a different network still leaves the driver with nothing to talk to. Only a physical CPU, or a controller simulator a vendor explicitly couples to the HMI tool, provides that endpoint.
My screens animate and navigate while the PLC values stay dead — is the link partly working?
No. Screen drawing, navigation, local scripts and most alarm presentation run inside the HMI runtime, so a fully interactive screen can sit on a completely failed PLC connection. Animation proves the runtime works, not the driver. Check the driver's communication status first, and trust only a PLC-linked object that updates as evidence of a live path.
Is this limitation specific to EasyBuilder and SimuLab?
The pairing is specific; the pattern is not. Most HMI simulators reach controllers through the vendor's driver list, and most controller simulators execute logic without a communications service. Simulator-to-simulator exchange exists only where one vendor documents a coupled simulation across both tools, as Siemens does for HMI operator interfaces and SIMATIC controllers; every other pairing needs a real controller on the driver's far end.
Do I need the physical touchscreen for the integrated test?
No. The EasyBuilder PC simulation keeps the operator interface on the PC, so one CPU module covers the controller and communication side of the integrated test. Bring the physical panel in later for touch behaviour and panel-specific checks; it is not part of proving the data path between screen and controller.
A value I write from the screen disappears immediately — did the transfer fail?
Probably not. When controller logic assigns that variable every scan, a successful write is overwritten within one cycle and looks like a rejection. Give the command and its feedback separate variables, then watch the status: a status that follows the command proves the transfer succeeded, while a silent command variable points at the logic consuming it.