Satisfactory: a splitter distributes to available outputs

Identify input and output roles

The splitter reference describes one input and up to three outputs. Begin by locating these roles in the current building view rather than assuming every side is equivalent. A belt connected to the wrong side does not become an input merely because it points toward the factory. For an introductory test, use one known item stream and two clearly distinguishable output branches. Keep the layout open enough to inspect each connection. The test should answer one question about distribution, not simultaneously introduce mixed items, several splitter variants and a complex network of mergers.

Separate division from multiplication

A splitter changes where items go; it does not increase the incoming supply. If an illustrative steady stream delivers 60 items per minute and two unrestricted outputs share it evenly, each receives 30 over time. Three equivalent outputs would share the same total rather than each receiving 60. Write a conservation check under the diagram: the combined outgoing flow cannot exceed the available incoming flow over a sustained period. This simple arithmetic prevents an attractive layout from promising more resource than the miner actually supplies. Always keep transport limits and actual source rate beside the hypothetical division.

Observe what a backed-up branch changes

The reference explains that the basic splitter skips unavailable full outputs and distributes among remaining available ones. A branch filling up therefore changes the distribution opportunity. Do not treat a temporary imbalance as proof that every basic splitter needs a programmable ratio setting. Inspect whether the consumer is taking items and whether its outgoing production is blocked. A full input buffer may be a downstream symptom rather than a failure at the splitter itself. Follow the item’s job through the consumer before adding more splitting equipment. The visible branch is part of a connected system, not an isolated pipe of guaranteed constant flow.

Run a controlled branch test

First observe two unrestricted branches, then deliberately let one destination fill in a simple practice setup. Watch how the other branch behaves, and restore the original condition afterward. Use ordinary materials and a setup you can easily explain. Do not claim exact rates from counting a few animated items; the purpose is understanding the change in availability. Once that behavior is clear, return to the real production line and ask which branch conditions differ. A small test gives the large layout a mechanical explanation, replacing the vague assumption that all outputs should always look equally busy regardless of their consumers.

Questions that come up