Factorio! Build, Expand, Automate, Survive

[Morganite]

So here's something I'm currently prototyping.

Spoiler

I've never made much use of sushi belts before, and when I was looking at combinator-free ones I found myself eventually thinking "this takes up -how- much space and -how- many splitters?", but this particular arrangement seems to be quite compact.

And also quite low throughput. But being fast isn't necessary here, I just need to be able to grab a few of each of these machines whenever I want them, and this would get that done with a lot less faffing about than would otherwise be expected from multiple 7+ ingredient recipes. Does that make sense, or am I missing a better way?

-Morgan.

 

[Baughn]

Does that make sense, or am I missing a better way?

Click to shrink...

Depends on your goal. These look like single-purpose assemblers; if you're unfamiliar with circuit logic, you probably didn't consider the 'set recipe' option. It's entirely possible to build a set of circuits that:
- Figures out what items you're short on. (Reading from roboport or a set of chests.)
- Picks a random recipe from those.
- Sets the recipe for one (or all) assemblers to that recipe.
- Reads the ingredients from the assembler(s).
- Requests those to requestor chests... or puts it on a sushi belt, whichever.

It's pretty intensive on the robots, and it'll still be slow without them, but the benefit is you'll constantly have 6-10 assemblers running on whatever you're short on at that moment in time, and also you can add new items to the mall by three clicks on a constant combinator.

 

[Morganite]

if you're unfamiliar with circuit logic, you probably didn't consider the 'set recipe' option

Click to shrink...

This is a 1.1 game, so that's not available... but crafting combinator is, so maybe I could make that work. Though it's kind of fussy; sometimes feeding the item signal works and sometimes it doesn't and I'm not real clear on the pattern. And I already discovered this game that it's possible for both materials and finished products to end up inside its overflow chest, which is... potentially awkward. Though probably an issue that can be worked around given a large container...

...

Actually, this game has 6x6 AAI containers. I could probably dispense with the sushi belt entirely and have just have multiple assemblers next to this giant box pushing things in and out of it, both ingredients and final products ending up in the same (freaking huge!) container, with everything regulated by combinator logic. And if I'm willing to use a couple combinators per input item, I can get something that doesn't need manual entry of the quantities for everything.

(In theory it'd instead be possible to use the CC on a drone request depot and regulate things that way while only needing to use one depot, but for some reason I can't get it to regularly work on depots. Like I got it to fire once and have since been unable to repeat the feat.)

-Morgan.

 

[Yla]

I can tell you that it's simpler than the design I was using (pre-Space Age).

Overall it looks very similar, but I had an additional set of wires on the input belts. They impulse-counted newly added items, the red wires on the assembler inserters (they were literally also red) counted removed items; negate the latter and add both onto the memory cell. I then had a combinator with the target amounts; subtract and have that wire control a series of inserters that put them from the supply chest onto the input belts.
It seems to me like your design has each pulse from the inserters unstop the appropriate belt for a tick. As long as the supply chests are kept full and that always means release of an item, that should work and it's simpler than my design, working with two wires instead of three. I assume you just dropped the starting amount on the belt manually?

(I had about 50 assemblers, and instead of the sushi belt being a single loop, I used 2x2 AAI chests connected by loaders to belts going in each direction forming a (small) grid of multiple rows. Each Assembler had both a short and long inserter for both lanes. It thoroughly mixed everything and one greedy assembler at position 1 couldn't grab everything. Given enough time, that doesn't matter though.)

Adding edit: Big 6x6 AAI containers work as long as you have no more than 6 assemblers. If you need more, and then possibly different materials in each, then it gets either complicated or you just allow all materials into all containers. That inefficiency offended me so I didn't use it for the mall, but I made a lot of those for Space Exploration things. Long rows of factories for things with more ingredients than 2 belts & long inserters could have dealt with.

 

[Morganite]

It seems to me like your design has each pulse from the inserters unstop the appropriate belt for a tick. As long as the supply chests are kept full and that always means release of an item, that should work and it's simpler than my design, working with two wires instead of three. I assume you just dropped the starting amount on the belt manually?

Click to shrink...

Not my design as such. It's adapted from the one in this reddit post. The combinator logic anyway; since this is a SE game that beacon arrangement is obviously not going to fly. I don't *think* it's bothered by supply shortages, and the belt primes itself, but it's troublesome when the count in the memory cell is out of sync with what's actually on the belt.

Adding edit: Big 6x6 AAI containers work as long as you have no more than 6 assemblers.

Click to shrink...

Bob's Adjustible Inserters see your 6 and raise you 12. I was originally going to just have 8, but then I felt compelled to push it a bit, so...

Spoiler: My current prototype

Turns out I couldn't come up with a way to both regulate material feeding *and* not have to manually set filter conditions requires six combinators per feeder. But you know what, that's okay, it just means I need to set up automated production of them. (Something I haven't previously bothered with.) The grey belts are from a mod, and are half as fast as yellow belts, which help keep the incoming mixed, and there's less overfilling since the main box is being fed from both lanes of two belts instead of a single lane of one. Faster belts will help with that, of course, but that's still several science packs in the future right now.

One constant combinator for finished products, one for intermediates, so that the former can be offloaded to a depot. Most of the assemblers will randomly select a recipe from the needed items (with the random provided by that combinator with what looks like a 7 on it, which is another mod item), craft it once, then (possibly) select another one. Four do something else, so it should be really hard for it to stall out because a locally crafted intermediate is missing.

One catch is that the four corner machines can't always run as fast as the 'interior' ones. Since they can't dump any no-longer-needed ingredients back in the center box directly after a recipe swap, they have that wooden chest instead, and I've got the inserters feeding the corner machines set up to disable if there's anything in them, because otherwise they'll fill up quickly. That's more of an issue with some recipes than others, and should get better in general with inserter capacity research.

-Morgan.