Factorio Forums

For the last few weeks, I've been working on integrated productions units centered around roboports.

I was getting unsatisfied of overlapping / undercovered areas since I had to place roboports between production units.
I read about building around roboports instead of plopping roboports between assemblers.
So I started FIP

1 / all productions units are centered around one or two (or more) roboports. So a basic production unit takes 1x1 roboport area
2 / all productions units take only ore (smelting is done locally) and either crude oil (if LPG is the main need) or lubrifiant
3 / all productions units are fully balanced (that's the easy part ... hard part is getting all intermediate materials to the next step without getting out of the perimeter)
4 / minimal use of robots
5 / fed by a main bus of ores and oil products

Example :

One full belt of iron ore (coming from the left) feeds both iron smelting (bottom left) and steel smelting (top left) .
Specifics material (pipes and green circuits in this unit) are brought via robots from other production units. For pipes, I used to have an assembler in this unit, but the closest one has one too so it's cheaper to only have one.
Then final items production occurs on the top right : inserters, chemical factory, refinery, pumpjack. Repair tools happen to need the same materials

After that, it's just a matter of stacking multiple productions units. Right now, I've done all the products needed for a meta base (a base that builds items for other bases), from blue belt (150 belts a minute on a single roboport square) to circuits, weaponry and assemblers. Can share if interested.

I'm still working on modules level 3 units, which pumps 11,4 modules a minute on a 3x5 roboports area.

Next week, I'll plop the rocket parts production units, which are set up to 100 parts a minute and stackable without limits (previous game hit 2 rockets a minute)

Blueprint

Smelting locally causes a lot of trouble (space requirements and use amounts of materials stacked in the smelters (100 plates per smelter output cache)) while it doesn't really solve anything, since ore is never needed as ore (except for concrete), can only be used in a single recipe (ore -> plate), and ore is less compact to boot. So, of course feel free to take any challenge you want, but smelting locally doesn't make a lot of sense

For steel the argument is more subtle since making steel locally means one less type on the bus, at the cost of space requirements (e.g. you need 10 steel smelters to fully supply a single turret (yellow) assembly machine, ignoring modules on both ends) and a lot of iron in your steel output cache (100 plates per smelter).

(In terms of logistics, the easiest is to smelt at your mining outpost and never transport ores; in terms of building new outposts it is easiest to build a central smelting area at the "start" of your main bus.)

I used to do that
I know that conventional set up have a central smelting or outpost smelting.
I had some quite big ( > 5000 ) furnaces smelting area before.
However, once you start to scale (aka plop big blueprints - more than 200 x 100 ) you have to adapt your smelting area for each new production unit, which can become tedious. By running ore, you have just to plop some extra lanes of ore to get to the new production unit

Check again the basic accumulator unit :

Of course, feel free to design your factory anyway you want .

However, I still don't understand the advantage (and I would like to understand!). You never need to store ore, you can just smelt anything you mine. That means you need new smelters when you add more mines, but the # of smelters doesn't need to change if your add production capacity other than mines?

You might be right. I need a few days to test some variants, because I still believe that roboports centred units are the best way to build, but I am going to test them with plates