#sandcatposting

the assembler seems to stop every so often. looks like it's not getting enough iron plates as input. maybe giving it two stack inserters will fix it! just offset the rows by a bit, and...!

um. it's still not saturating the belts.

the issue looks different now, since only the assemblers at the end are being input starved. there's not enough input belts!

looking back at the ratios, this current setup needs 9 belts of iron, which is 3 blue belts. 4 is close enough, right?

let's keep the current setup, but instead of 2x4, let's stack it 4x2 instead. this should let us get some better beacon efficiency too!

new problem! we're output constrained! it seems like each assembler with 6 beacons needs two inserters for output!

notice how the offset introduced in [2] makes the space between the assemblers 4 tiles instead of 3, which prevents us from using a single belt to get the output from both columns. we can add some long inserters to compensate, but it isn't enough. let's just add more lines of belts. how many could we need, anyway?

oh. that's a lot of belts.

theoretically this makes it pretty tileable to a certain degree. we can get a maximum of 3 blue belts (9 belts) of gear output from a fully sized array. it looks a bit inefficient, though, and needs a bit of processing at the end to consolidate the half full belts. let's go back to the drawing board.

more beacons have diminishing effects. if we go down to 3 beacons per assembler, we need 9 assemblers. let's put them in a 3x3 matrix! this has the bonus that every row can now be fed a single blue belt.

it works! we get full throughput! and at a max output of 6.2 belts, it's pretty close to the 6 belts of output that we have!

it's just. um. a little space inefficient. also it looks bad.

really, the most space efficient designs we had were the 2x4 ones. why don't we just... put the input on the outside instead of the inside?

this.. actually makes it much easier to deal with the input routing! and for the output, we can do a cool trick i learned from Nilaus and put the output directly onto a splitter!

also, since this design is a line, it means we can try and fit as many beacons as we can in between them! fun!

here's the final design! it takes in 3-ish blue belts (8.5 belts) of iron plates and puts out two blue belts (6 belts) of iron gears! yippee! also it looks way better now!

postmortem: turns out i needed to make one change to my original design. sigh! will i update my smelting array designs? maybe! power is one thing i have more than enough of. also please don't think this design is optimal. i am not that good at factorio.

and here's the use case!

is this almost certainly too much iron for a single mall? yes! ok theoretically it can use up to 27 belts of iron (9 blue belts) when running at max capacity but like. it's my belt/inserter mall. it's not going to be running at full capacity. yet. maybe. look i'm just planning for the future ok.

um yeah that's it. i just did this for fun. maybe this is like, educational or something.

tada, yellow science! hey, i thought it turned out pretty good.