System Diagram Editor — User Guide
Draw the system configuration of a machine (its devices and the networks between them) and, in the same act, produce data that serves as the container structure of the Memory Map.
1. Overview
The System Diagram Editor lets you place devices — PLCs, servo amplifiers, robots, remote I/O and so on — and connect them with communication links. What makes it different from an ordinary drawing package is that the picture you draw is structured data. A device's position is stored as x, y, a link is stored as the relation “these two devices are joined by this network”, and every “block” (module, functional unit) you create inside a device becomes a container in the Memory Map Editor.
This is not a general-purpose drawing tool for placing arbitrary shapes and text. All you can draw is devices, links and blocks — and because of that restriction the drawing can never drift apart from the specification (the variables and addresses). It runs entirely in the browser, and your data never leaves this machine.
2. Tutorial (your first diagram)
- Choose the kind “PLC / Controller” in the toolbar and press “Device”. Type a name and press Enter.
- Add “Servo amplifier”, “Robot”, “Remote I/O” and anything else the same way.
- Select a device and a ○ appears on its right edge. Drag it onto another device to draw a link.
- Click the link and choose the network type in the property pane (EtherCAT, CC-Link IE, RS-232C …).
- Select a device, press “Block”, and create the functional units inside it (command, status, parameters …).
- Ctrl+L arranges everything. To put the picture into a document use “Output → Export…” (Ctrl+E); to move on to the next step use “To Memory Map →”.
3. Devices and blocks
A device carries a kind (PLC, HMI, servo amplifier, inverter, robot, remote I/O, sensor/camera, network device, PC, other), a model, an IP address or station number, and notes. The kind drives the colour coding in the diagram and appears in the device heading when exporting to SVG or Draw.io.
A block is a functional unit inside a device. Splitting a device into “Command”, “Status” and “Parameters” lets the Memory Map assign addresses and copy to your PLC software one unit at a time. Blocks are optional — without them the Memory Map simply treats the variables as “unclassified”.
4. Drawing links
Select a device and a ○ (port) appears on its right edge; drag it onto another device to draw a link. Select the link and you can change its network type, its label and its direction (one-way or two-way) in the property pane. Fieldbuses such as EtherCAT and CC-Link IE are drawn with a solid line, serial links such as RS-232C with a dashed line, and digital I/O with a dotted line.
You can draw more than one link between the same pair of devices (for example EtherNet/IP alongside digital I/O). The label is a good place for notes such as “Req/Ack” or “station 1”.
Routes are worked out automatically from where the devices sit, but any line that does not come out the way you want can be tidied by hand. Drag the middle of a line and that part moves sideways. Moving it records waypoints on the line, and from then on the line is drawn through those waypoints instead of along the automatic route. Select a line and its waypoints appear as □ / ■, which you can drag. If a line ends up on the wrong device, select it and drag the □ (source) or ● (destination) at its end onto another device.
Double-click a waypoint, or drag the part you moved back onto the original straight line, and the waypoints disappear and the automatic route returns. Right-clicking the line and choosing “Back to the automatic route” does the same. Auto-arrange (Ctrl+L) puts every line back on its automatic route. Waypoints are saved in the JSON as waypoints inside links[].
Each link can also say which end is the master (select the link and use “Master” in the properties, or right-click). The master end is marked M and the slave end S. A cascaded setup, where a slave on one network is also the master of another (a PLC that is a slave of the line PC and the master of its remote I/O, for example), is expressed by setting it on each link. Auto-arrange puts a master that is nobody’s slave at the top.
5. Placement and auto-arrange
Devices can be dragged freely, and the position you give them is saved as coordinate data, so the layout is the same the next time you open the file. Auto-arrange (Ctrl+L) works out a hierarchy from the links — the PLC at the top, the devices it talks to below — and lays the diagram out again. Arranging first and then nudging the few devices that bother you is usually the quickest route.
Shift- or Ctrl-click, or drag a box around empty space, to select several devices. Selected devices move together when you drag them; Ctrl+arrow nudges them, and they can be copied, duplicated or deleted in one go. A line whose two ends are both selected keeps its hand-adjusted route as it moves.
6. Exporting the diagram and tables
Everything you can export lives in one place: “Output → Export…” (Ctrl+E). Each format says where it is meant to be pasted, so you pick one and press “Copy to clipboard” or “Save to a file”. The format you picked is remembered for next time.
- SVG: a white-background, black-line drawing for a specification. Save it to a file and bring it in with Insert → Picture in Word, Excel or PowerPoint, and the lines stay sharp however far you zoom.
- Image (PNG): copy it and paste it straight into Word, Excel, PowerPoint, an e-mail or a chat.
- Draw.io: shapes you can keep editing. Copy it and press Ctrl+V in the draw.io window. A drawing changed in draw.io cannot be loaded back into this editor, so change the devices and connections here and paste again.
- Markdown: a device table, a link table and the diagram in Mermaid form. GitHub and Notion render the picture directly.
- Mermaid: only the diagram, as Mermaid code (.mmd) — the same diagram as in the Markdown. Devices with the same location are framed together and safety links are drawn as thick lines.
- JSON: the format shared by all four tools — for feeding an AI or handing the data to the next tool.
If draw.io pastes it as plain XML text, add a new page with the + at the bottom left and choose Extras → Edit Diagram… on the menu bar (without the menu bar: + → Advanced on the toolbar). The box holds the XML of the current page: select it with Ctrl+A, paste, and press Apply. Apply replaces everything on the page, which is why a new page is used.
7. The design flow (four tools)
This editor is the entrance to the following flow.
- System Diagram Editor (this tool): decide the devices, blocks and networks
- Memory Map Editor: assign variables and addresses inside those containers
- Timing Chart Editor: decide how the variables move in time (the handshakes) and what causes what
- Flowchart Editor: turn those cause-and-effect relations into program flow
The “→” button at the right of each toolbar carries your data into the next tool, and the button beside it takes you back. Using any one of the tools on its own is perfectly fine.
8. Keyboard shortcuts
For the thinking behind this kind of drawing — what belongs on one sheet and the pitfalls seen on site — see the guide How to draw a system diagram (in Japanese). All four editors are listed on the FA Spec Tools home page.
Please send bug reports and feature requests to yk.builds.space@gmail.com. Steps to reproduce the problem — and, if you can, the saved JSON file — make it much quicker to look into.