LAD generation

AI ladder logic for TIA Portal, written into your open project

Coilmind writes ladder logic straight into the project you have open in TIA Portal. Every block is compiled before it stays.

Start your 14-day free trial
The 70-second tour: LAD networks written into an open project, imported and compiled, then run on the simulator. Screen recording, no audio.

Why a chatbot cannot give you LAD

Ask a general chatbot for ladder logic and you get text back. LAD in TIA Portal is a graphical language: networks of contacts, coils and boxes stored inside the project. There is nothing in a chat window you can paste into a LAD network.

So the answer comes back in one of three forms. SCL, which you then have to rewrite as LAD by hand if the plant standard says LAD. A description of the rung in words, which you then draw yourself. Or ASCII art of a ladder that looks right and cannot be imported anywhere.

Either way you end up placing every contact and coil in the editor yourself, and the first compile happens after you have done the work.

Why AI tools for TIA Portal started with SCL

SCL is text, and a language model produces text, so SCL is where AI tools for TIA started and where many still stop. LAD needs a valid network topology: every contact in series or in parallel on the right branch, every box wired to the right power flow, every coil at the end of a rung. A model that guesses that layout as text gets it wrong often enough to be useless.

Coilmind does not ask the model to draw the network. The logic is built as an internal graph first, the graph is checked, and only then is the block written into the project through the TIA Openness API.

Your request in chat
Logic built as an internal graph
Dry-run check: topology, schema, semantics
Snapshot of the block
Import through Openness
Compile in TIA Portal

If the compile fails, Coilmind reads the errors and corrects the block. If it still does not build, the change is not applied and your block stays as it was.

What it writes in LAD

Standard LAD instructions for S7-1200 and S7-1500 blocks:

  • ›NO and NC contacts, positive and negative edge detection.
  • ›Set and reset coils alongside plain output coils.
  • ›Timers (TON, TOF, TP and the rest) and counters.
  • ›Comparisons, MOVE and math boxes.
  • ›FB calls with their instance DBs, and FC calls.
  • ›Network titles and comments, so the block explains itself in TIA.

What it does not write

It also writes SCL, so if a block is clearer as text, ask for SCL instead. Generated logic, LAD or SCL, has to be reviewed by a qualified engineer before it runs on a machine.

  • ›Fail-safe (F-) blocks. Safety programs stay with you and your safety tooling.
  • ›New STL. Coilmind reads STL and can rewrite an STL block as SCL or LAD, but it does not write fresh STL.

Example: from a task description to networks in TIA

Say you type: "In FB_Pump, add a start/stop for the pump. Start from the start button, stop from the stop button or a motor fault. If the flow switch hasn’t made 10 s after start, stop the pump and latch a dry-run alarm until reset."

What you would expect to land in FB_Pump is a few ordinary LAD networks. One with the start contact and a set coil on the run bit, and the stop contact, the fault contact and the dry-run alarm in parallel on its reset coil. One with a TON timer enabled by the run bit with the flow switch as an NC contact. One that sets the alarm bit from the timer output and resets it from the reset input. Each network has a title, the block compiles, and a snapshot of the previous version is there if you want it back.

A maintenance engineer opening that block sees contacts, coils and boxes in TIA like in any other network. There is no special viewer and no runtime from Coilmind in the PLC.

What you need

  • ›Windows and TIA Portal V17 to V21, with the project already open. Coilmind attaches to it through Openness.
  • ›I test day to day on V19 to V21. A few conveniences, block renaming among them, need V21.
  • ›PLCSIM Advanced V3 to V6 only if you also want to run tests on the generated blocks. Siemens sells it separately.

Common questions

Can AI write LAD for TIA Portal?

A chatbot on its own cannot, because it only emits text and LAD is graphical. Coilmind builds the block itself through the Openness API, checks the network topology before writing and imports it only if it compiles. What lands in the project is a normal LAD block.

Is the generated LAD readable for maintenance?

Yes. It is made of standard TIA networks with contacts, coils and boxes, and it can carry network titles and comments. Your maintenance staff read it in TIA Portal like any other block.

Which TIA Portal versions does it work with?

V17 to V21 on Windows. I test day to day on V19 to V21, and a few conveniences need V21, so on V17 and V18 expect the core to work and the edges to be thinner.

Which PLCs?

Coilmind works with whatever project is open in TIA Portal V17 to V21. LAD and SCL generation is aimed at S7-1200 and S7-1500 blocks.

Can it write safety (F-) blocks or STL?

No to both. It does not write fail-safe blocks. It reads STL and can rewrite it as SCL or LAD, but it does not write new STL.

What happens if the generated LAD does not compile?

Every block import is gated on compilation. Coilmind tries to fix the errors it gets back, and if the block still does not build, the change is not applied. A snapshot is taken before every block change, so you can also restore the previous version in one click.

Generate a LAD network in your own project

Open a project you know in TIA Portal, attach Coilmind and ask for a network you would normally draw by hand. Fourteen days, no card.

Start your 14-day free trial