Sushi thread

[normal gulch]

Ohh i see where u saved a space, interesting

minse a lot easier to adjust imo

[hoary trail]

I recently updated it but this version doesn't have the update. I changed the middle belt to the "everything" signal rather than a specific signal to the particular item.

[normal gulch]

smart

[hoary trail]

So now all I have to do is paste it, and change the splitter setting, and that's it

One click and it's ready to go

[normal gulch]

Yeah can handle recycling better because ti runs a belt thru and that gets prioritized

I have a global wire so i put all requests on 1 CC

[hoary trail]

Well I guess two clicks since I need to set the rate too

[normal gulch]

this is cool.. if I want to redo mine I guess i can get rid of ac ombinator urs has 2 not counting constant

[hoary trail]

Mine handles recycling with the stack inserter, which is why it can't handle really high item rates, because it's limited to the speed of the stack inserter

[normal gulch]

It also cant h andle artilly shells that well

[hoary trail]

I've considered doing one that handles higher throughput better, but I can't see a use case for it. If I need that many items on one belt, I'm probably not going to be doing sushi

[normal gulch]

I do it for my wall, all adjustable I do 1K shells/minute sometimes

[hoary trail]

I mean I could throw arty shells on if I wanted to, but I don't really want to do that

[normal gulch]

and having it easily adjustable is amzing

yeah mines for defense and eventually anything i want

[hoary trail]

Mine is just meant to build my rails for now, and maybe a few other specialized things perhaps in the future

But I'm intending to use builder trains or spidertrons in conjunction with the sushi rails

Or I might just make these rails for fun 🤷

[hoary trail]

what does the extra combinator on yours do?

[normal gulch]

no idea lmao

calcium wizard did the smarts

ill send what they do in a minute

if i can remove a combinator, i can have near infinite capacity sushi ( so like 1K shells ) and in 3 tiles thisl lbe amazing

top combinat to bottom:
arth combinator: T + I output T
constant combinator:L Y 2.7K
decider combinator: T > ( or equal ) Y ouput Y input count
arthmetic combinator: Y * -1 output T

[hoary trail]

So you input Y, which is the timer reset number?

[normal gulch]

yes Y is 2.7K and thats belt speed

[hoary trail]

Oh so L is the number you change

Wait I'm confused why L only appears in one place

Oh, that an "i", not a 1. So yours adds a specific number to T rather than changing the timer reset number, which is what I do in mine

I wonder if that has better decimal accuracy for the item rate 🤔

[normal gulch]

im not sure as i didnt design it haha, if urs does work good I might have to switch for new projects

urs potentially can be my any item sushi limiter as well

[hoary trail]

What does that mean

[normal gulch]

something very complicated haha

[hoary trail]

I'm also not sure what you mean by infinite sushi either haha

[normal gulch]

so Ill put R in my list of items, and I have an outpost that needs items like belts, miners and itll put it on the RAS ( random access sushi ) limiters that have no buffer somehow to switch between item types

[hoary trail]

If there's no buffer, then that seems like it would work

[normal gulch]

yeah I was also thinking of just clocked inserters as welll

[hoary trail]

I didn't like clocked inserters

They were much less accurate

[normal gulch]

i mean it doesnt need to be 100% accurate

like put a ratio of random items to one limiter.. hm

[clear latch]

I didn't design the most important part of my kitchen either tbh. It's calcium's moving average counter, I just added an extra decider to disable an output belt (forcing a loopback) until it finished averaging out over 1 minute. Never needs resetting and can flow limit multiple belts over the same wire from 1 CC

[hoary trail]

I wanted to design my own, so I just currently use my own (worse) version haha

Mine has a separate rate combinator for each item, which is not as convenient, but it also allows me to separate my kitchen easily into distinct sections along my base, which I'm considering doing

[clear latch]

Yep I started with the exact same design goals as the last sentence, just with a manually reset throughput counter that I made myself. Had to toggle the CC every time I added a new item which felt clunky

[timid cosmos]

fresh prints, get yer fresh prints here! hot prints

think i can get that print?

[normal gulch]

no but theres a slightly bigger one haha

its 2 of them and it tiles

[timid cosmos]

active provider?

[normal gulch]

!bp 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

[hoary trail]

Seems to be working pretty well

[normal gulch]

u can get away with 1 requester chest probably

[normal gulch]

upgradable 2-4 lane and straigth -> 3 way -> 4 way?

[timid cosmos]

hot prints

i assume the provider chests are inputs

[hoary trail]

The straight, 3 way, and 4 way should be upgradable. I can't upgrade to 4 lanes though as this original train blueprint wasn't designed for 4 lanes

[timid cosmos]

i might use these little limiters for my sushi rail

[hoary trail]

I won't be playing my map long enough to need 4 lanes though

[normal gulch]

requesters are onse on the right

[hoary trail]

So I'll redesign a larger version if I ever want 4 lanes in the future

[timid cosmos]

err yeah requester

[normal gulch]

how big are they? 52 x 52?

[hoary trail]

Technically, they're 38x76

Each piece is a straight 38x38 plus a 38x38 intersection

Yes, I designed a weird rail system in the first place. But it does work with sushi if I'm careful with it

[normal gulch]

how many rails/minute in a straight line can it go?

[hoary trail]

I don't know how to measure that

Well the sushi line carries 300 rails/minute, and the whole 38x76 thing is 218 rails

So I can theoretically build one of these every minute

[normal gulch]

i mean just place one down and see how long it takes to build, and divide by distance

[clear latch]

I did separate 84x84 intersection and straight sections for rotational symmetry. Unfortunately they really aren't rotation safe because the roboports are offset around celtics, they connect segmented areas when rotated the other way, and the circuitry only makes sense in one direction

[timid cosmos]

Oh I see, fish becomes whatever input you're doing there. Whats with the ? shouldn't it just be an iron chest to hold potential backlog?

[hoary trail]

This is also the most expensive version, I'll probably mostly use a straight variant of the tracks and paste an plain rail intersection over top of it when I want to add a branch off the main line

But the sushi belt will need to have intersection capability already built in as you can't paste that on top

[normal gulch]

I had an issue when I changed amounts, eg walls from 420~ to 100 and too much getting in the chest

[clear latch]

I also experienced that in my kitchen and for some reason that item didn't want to leave the chest either until I manually removed the buildup

[timid cosmos]

I see. Surely a steel chest would be fine for my purposes though?

i don't want weird bot orders being sent out and canceled constantly

[normal gulch]

if ur not changing amounts on belts crazily, ur fine

im just a maniac with my sushi

[clear latch]

Inserters and loaders tend to empty the chests faster than the bots can fly over to steal from them unless there's a buildup, which there shouldn't be normally. I actually used that for bot kovarex once to bot out the excess into trains while the inserters took enough to sustain the recipe itself

[timid cosmos]

i think yeah i'm just gonna borrow this print for my sushi rail input. Seems way less headache and actually makes more sense than trying to implement the active sushi over time design throughout the whole system. I can tear up a bunch of circuitry and use those wires for something else, like managing priority stations.

[normal gulch]

tangled i figured out a use for a while but u probably dont need it

the amount of items u want in those at each outpost

[timid cosmos]

a use?

[normal gulch]

i change it from no solar to solar but again im a maniac with my sushi

[hoary trail]

I just timed it, and it took 90 seconds to build that intersection. So it's approximately 50 squares per minute. It laid 145 rails per minute. And this is with the maximum intersection (4-way crossing, ignoring 4-rail for now), with no surrounding bots to help it. When there's nearby belts, the cross-coverage can help it to build faster so I made sure to isolate it to measure it better

[timid cosmos]

i'm not sure i get you fireball what're you saying?

[normal gulch]

1 sec ill show u

[hoary trail]

I tried to use a single passive provider chest and change the inserter limits, but this ended up being insanely slow compared to my 10-chest method so I left the 10 chests. More expensive, but much faster

[clear latch]

You could also use 10 inserters and it would adapt with filters even beyond 10 items. Or with less than 10, each would pick up much faster

[normal gulch]

so it gets the items from the red wire

the green wire is a fish for artillery loading

[hoary trail]

But if I'm using ten inserters, I'll need ten chests, which is expensive (which was my point)

[clear latch]

Yeah I mean if you're having 10 chests and inserters anyway, you might as well set circuit instead of static filters so per item throughput isn't limited to 1 inserter and total item range isn't limited to 10. Like say you wanna flood the belt with those 1k artillery shells, now you can potentially have all 10 switch to picking up shells

[timid cosmos]

oh i see

[hoary trail]

Hmm, I see. But then I would need to add combinators to my sushi haha

[timid cosmos]

@normal gulch that circuit design fits exactly into my current system, as if they were made for each other 😲

@plain hare oughta see this that is just magic

[normal gulch]

the limiters are amazing

they also fit in train stations spacing

[timid cosmos]

it fits perfectly lol

this is pretty wide spacing for rails, specially chosen for this inside-the-rail sushi design

you and calc's design have revitalized my motivation to get this system up and running

i mean, it's done now, with some tweaking

[normal gulch]

awesome

how big is each block? idk if i asked u yet

[timid cosmos]

I just need to tear out all the old circuitry related to the active reading system

[timid cosmos]

i dont think i'm going to add any defense stuff

maybe localized, if i run into problems in a specific area

[normal gulch]

yeah this is pretty simple

i like it its simplistic

[timid cosmos]

closeup of the main piece, of course everything else is just this broken down

i can get rid of all those belts that are wired up

this was the original design which i liked the belts way more, but there was a fatal design flaw. That belt loop in the center is ONE TILE from being able to be build by another segmented roboport. So no loop, no movement, no bots in the new cell's roboport.

[normal gulch]

thats amazing

[timid cosmos]

it looked a lot better but was juuuuust barely not functional

[normal gulch]

i wish roboports were slightly smaller but this works well

[timid cosmos]

new design has the loop right up against the inner roboport, so at least the loop can be built

[normal gulch]

is it compiled into a set of BP's?

[timid cosmos]

The first version is, but it's non-functional

I need to work on this new one, tear out the old circuits, before I can break down the 4-way again into a new BP set.

[normal gulch]

Building straights and corners was hard enough for me with my rails + wall

[timid cosmos]

yeah this is significantly more spaghetti than V1

but

it's still rotationally symmetrical

which is key

[normal gulch]

yeah thats good

add priority to the splitters input and output so u dont have half sushi

[timid cosmos]

Can't have any priority in the center loop there

items gotta disperse

[normal gulch]

putin showed me a good image

this is how it goes thru

[timid cosmos]

here on each of the 4 arms of the intersection is where we see priority splitter. If the chest is not satisfied, that belt turns on and it forces all sushi flow to pass by that inserter

[normal gulch]

oh ok if that works then ok

[timid cosmos]

It works, but I'm not sure how it will do at scale.

[normal gulch]

u can do a squiggle there perhaps, im not sure. is there a roboport where ur charcter is?

[timid cosmos]

what i'm more concerned about is modularity of the system over sushi consistency

[timid cosmos]

squiggle? you can see the roboport in this pic

[normal gulch]

oh nevermind me

[timid cosmos]

man thats just great how that circuit slips right in. I'll have a working sushirail in no time

[normal gulch]

ill have to snag it when its done :D putin has some competition over the sushi rails now haha

[timid cosmos]

don't you have a design?

and Star is working on one too

[normal gulch]

its with walls and 100x100

[timid cosmos]

ah

i may regret not having walls

[normal gulch]

mine was designed as a wall, with rails and solar as an addon

[timid cosmos]

what if...

I put turrets in between the rails

and walls right up on the rails

[normal gulch]

[timid cosmos]

just 1 layer of wall, and enough turrets to kill expansion parties

well they cant go there, but somewhere

[normal gulch]

need inserters and then to put ammo on the belt this is looking very spagheti

[timid cosmos]

theres this design

i already worked on making it tileable

i could just slap those walls and turrets into the new setup

nothing has changed but the belt design and some circuitry

i'll have to think on it

this would have been awesome, but unfortunately theres no way to get the fluid to them and make it rotational

and of course V1 was flawed

[hoary trail]

The pipe itself wouldn't be midline, but I'd be curious whether it would have worked using barrels

[timid cosmos]

Cant fit an assembler in there, and if I could fit it somewhere else besides the center, can't get pipes to it

I have some modded pipes that would work, but i can't do flamers there with the new design anyway

[hoary trail]

Ah I see

I meant putting the assembler not inside the center and running an underground pipe to that spot right next to the flamer, but I can't tell on my phone if that's actually a valid spot or not. But I guess it's moot now

[plain hare]

Wow tangled

Those look great

They’ll be super effective I bet if you figure out the last few kinks

[hoary trail]

@plain hare what happens to your sushi loop if you were to connect two different ends of your track into a large loop? With my current priority settings, it would end up skipping the inside of the sushi belt. I'm planning to address this by just sticking in a rails-only blueprint for when I want to connect my tracks in such a way, but not connecting the belts so that the whole things still remains a continuous loop.

[plain hare]

yep thats what I have, if the sushi was gonna connect from 2 different directions it just doesnt

just rails and power only

[hoary trail]

OK, that makes sense. I have a current working sushi rails model now, but I'm going to try to optimize the provider chests a bit. I also need to decide what kind of mall I want to use to supply this whole thing. Sushi vs bot vs beacon spaghetti, I need to actually do some calculations and see whether I need beacons or not.

If I do need beacons, I'm leaning toward a bot mall since some of the things I need to make (bots and roboports) need space to route the inputs

I might still do a small bot mall (even if I don't need beacons) for my sushi rails because of the significantly smaller footprint.

[clear latch]

I just let all belts in all rails connect and wire a few belts to redirect flow if the items flow down the right-side belt (e.g. top to bottom instead of bottom to top) where I have no inserters if the CC/chests aren't satisfied. Possibly similar to how Tangled does it. Priority is set to the inside of the sushi belt but weaves around to force turns. So the intersection-capable straight tracks will prefer right turn if it exists, loop left and then straight ahead if right is cut off, loop down and then left if straight is cut off, and finally loop right and then back down if it's a dead end. I also have a "dumb" straight track without the weaving but the middle can be hollowed out if needed

[plain hare]

I believe star penguin was referring to a situation like this

https://media.discordapp.net/attachments/224238645127806977/925842112212254791/unknown.png

[hoary trail]

Mine uses the same logic in the right-preference, which works fine in almost all scenarios. I don't understand what yours does if you were to connect to separate ends to each other though

[plain hare]

Look at the loop of rails.

[hoary trail]

Yeah that's the situation I was talking about

[plain hare]

If you connected it all with belts you’d see funky flow directions

[hoary trail]

Also I feel like I'm the only person who's using non-square blueprints for this haha

All of yours are all nice and square

[clear latch]

Ahh I just let those back up at the ends until the flow normalizes again...maybe it's smarter to not have belts there if you can keep track of which is which, I don't like adjusting prints on the fly like that myself because I forget

[plain hare]

I don’t have to adjust anything, my print has the connecting belts hanging off the back so I just choose which one to connect to if this happens

(Another print with just rails)

[hoary trail]

It's the piece right here where I'm standing that would cause an issue if I were to include belts

[plain hare]

So I guess I just place 2 prints when that happens

[hoary trail]

But I can easily just make a plain print without the belts there

[clear latch]

How do you decide which belts to connect if you have a grid of many city blocks that share the same "sides"? Do you just try to build it slow and in consistent directions to keep the flow predictable?

[hoary trail]

Yes, that's my plan

[plain hare]

I don’t really do train grids, more tree trunk with branches.

[hoary trail]

But I'm not planning on doing a lot of city blocks

I may consider it in a few places if train throughput starts to become an issue, since my prints aren't upgradeable to 4 lanes

[clear latch]

I remember testing a block sushi print once and just pasting it in a 5x5 in four different directions at once, a couple blocks to the corner either never got built or took a long time for the belt to get to them

[hoary trail]

It does take a while sometimes with the yellow belt

It would be much faster with a red belt or a blue belt but that would be expensive and I'm not sure if my factory can afford it yet

[clear latch]

That was with modded 😂

[hoary trail]

lol

[plain hare]

Well, placing ghosts that can’t be built right away is maybe not ideal.

But also, I think most rails are headed out to ore patches so that’s a roughly linear pathway anyway

I’m gonna mostly extend mine in one direction, is my plan -> get the rich ore

[hoary trail]

currently mine takes about 90 seconds to build a 38x76 section

[clear latch]

Yep I was trying to see how far I could abuse Construction Queue / Dynamic Construction Queue, it was a stress test, in a real save I wouldn't have a reason to expand so many directions because I couldn't actually utilize them that fast

[hoary trail]

but I'm trying to optimize the provider chests better so it might get slower after that if I do a poor job

[plain hare]

I think, you can expand in multiple directions too if you wanted, I foresee myself keeping a straight line going north, and making west / east offshoots to grab ore

And then back home I might expand a bit for processing arrays

[hoary trail]

I thought you don't like going north, I thought you were an east/west explorer

[plain hare]

I might put in additional sushi kitchens to keep the sushi belt refreshed.

Well, going south kinda sucks with the toolbar in the way

So north with west / east seems legit

[hoary trail]

I see

I'm an east explorer

[plain hare]

Dunno if it matters too much ultimately xD

Right now I have one combined early game yellow belt sushi kitchen with belt based mall

My next sushi kitchen I’m gonna have a train delivering the items, and then I can easily make new malls / kitchens as needed.

Also let’s me make new independent sushi rail systems, all fed by trains .

Or just 2 kitchens on the same sushi belt.

The new style would work very well with a bit of robot and beacon abuse

[clear latch]

I haven't planned for multiple kitchens, but am considering decentralized trash. It's amazing how much wood, stone, and coal gets built over and I don't think my mall would be able to burn through it fast enough. So either the trash loop goes into AP => storage => trains or I somehow prevent those from getting into the main base and split them off to outposts that can consume them in bulk

[hoary trail]

This is satisfying to watch lol

[plain hare]

Super satisfying 😄

Even better when it comes back a belt full of trash from constructing, wood, stone, etc lol.

My belt has a lot of landmines on it at yellow belt tier.

[clear latch]

You know what I actually could make separate kitchens for some things...like modules, they build so slow and get built everywhere so it'll feel bad to train them to the mall and then put them on the belt, might as well put them on directly

[plain hare]

triple sushi belt intersection haha

yep im kinda doing that

the processing unit / robot frame / lds belt is my bus sushi belt

for the intermediates i want to centralize but dont want to buffer

ill probably end up with modules on that belt, and other expensive stuff too

probably science also

[hoary trail]

@plain hare how many provider chests do you use per intersection? I'm currently using four, which seems to be reasonably fast and isn't limited too much by the chests

It seems a bit expensive, but maybe it's really not that expensive compared to the iron cost of everything else

[plain hare]

My first set used 1 provider chest per unique item, sometimes more than one xD

Now my early game model uses very few. My late game ones I may change to use a lot more.

[hoary trail]

I have a separate chest for belts and a separate one for rails, as I need tons of those and I don't want those to be held up by inserters waiting for other items

[plain hare]

Yeah, it’s not that expensive in the scheme of things if you have full walls around each area, but if you’re doing like an early game one it can be expensive

[hoary trail]

But everything else I just put into two other chests

I'm kind of thinking to not use walls, as I'm not playing a deathworld

I'll probably use very small arty outposts to help keep lands cleared before I put up laser walls at chokepoints

[plain hare]

Ah. I just like having all that combined into the rails themselves. Very convenient to one click expand your artillery and rails.

Can just blueprint all the way to ore.

[hoary trail]

I don't think my rails would be able to withstand the counterattack if I attached an arty

[plain hare]

That’s the hard part xD

[hoary trail]

I'll probably have an arty outpost that snaps into my rail grid, that way I only need to put real defense at one section

My sushi rails are in a weird stage of the game, they're meant to be convenient, but they don't have to be dirt cheap as I'm post-rocket now

[plain hare]

only problem with that is theyll cross your rails to get to that artillery outpost.

[hoary trail]

oh, right, that's a problem

[plain hare]

yeah, thats the initial issue that made me design the sushi castle tbh

I was like.. what if the walls just protect the rails?

anyway aliens arent getting through that thats for sure

[hoary trail]

It seems too expensive to have to wall up and turret every single rail

[plain hare]

yea, it is, but my mk2 avoids this

I have landmine and turret sections

[hoary trail]

With landmines?

[plain hare]

so I can do landmine landmine landmine artillery

[hoary trail]

Do landmines one-shot even the behemoths?

[plain hare]

also, you can deconstruct old turrets and recycle with the sushi

they dont one shot behemoths but I dont expect to have any behemoths until im filthy rich anyway

[hoary trail]

That makes sense

[plain hare]

so many eff module shenanigans

[hoary trail]

I have behemoths now, so I can't use landmines anymore

I'm in a weird place haha

[plain hare]

perhaps somewhat induced by a premature module 3 transition tbh

theyre very expensive and polluting

[hoary trail]

Well, perhaps, but I've also launched 90 rockets so I don't feel like the biters have evolved that much in comparison

[plain hare]

but , I think you can probably throw turrets at them and be fine.

yeah, use some space science on damage upgrades, and youll be fine

[hoary trail]

Already did

My double wall lasers pretty much stop all expansion parties even when there's behemots in the mix

[plain hare]

ah, the other thing is lasers are just super expensive for their dps.

they have no upkeep but the dps is less than 10x what a uranium gun would be

so effectively they cost a lot more ore to setup, only paying off after repeated assaults where you dont spend bullets.

[hoary trail]

And my real walls occasionally take damage from a really large revenge party, but they've never really broken through

I use the double laser wall only for chokepoints, especially where I'm planning to eventually continue to expand

[plain hare]

I did a behemoth vs bot battle once with my old sushi castle

[hoary trail]

So they're not too expensive since it's a small area to have to wall off

[plain hare]

with items coming in constantly on sushi belt in the small network bots just replace everything

[hoary trail]

And I don't want to use other turrets since the wall is designed to be temporary

[plain hare]

its basically impenetrable given good supply since the repair time is tiny

[hoary trail]

I remember that

It was quite impressive

[plain hare]

yeah, lost 1200 bots in a few minutes but we won

fair point about the other turrets I think, if I have dedicated supply belts of uranium ammo my recycle time will be longer than build times...

oh well I think its fine I can do some things

[hoary trail]

I mean for the most part it's unlikely to be an issue

I just prefer it for convenience since I did some moving walls early in the game when I was establishing my perimeter and I really regretted the one that had flamers and gun turrets when I wanted to move it

[plain hare]

I would think flamers would be easy to move but the guns yeah

[hoary trail]

Well I didn't want to waste the ammo haha

[plain hare]

Oh. Fair enough. A bit of light oil might go down the drain

[hoary trail]

Also have you considered the expense of the buffer that this enormous sushi belt is once it fills up?

I assume you've calculated it out

[plain hare]

Yea, it’s some buffer for sure

Though it corresponds with the amount of items on the belt so it’s not the worst

[clear latch]

I use 2 passive provider chests, 2 stack filter inserters, and blue belts in straights, and double all those for intersections. It takes about a minute either way. Building over water is probably 2-3 minutes depending on how much landfill is added to the blueprints

[hoary trail]

Hmm let me retime mine again

87 seconds, not bad for going from 10 chests to 4

That's also in the slowest case scenario

[clear latch]

I use the resource scanner from recursive blueprints to detect nests and stop building to avoid drawing aggro. On second thought it's not really possible to detect enemies without also aggroing them in vanilla; circuits can't "know" whether an artillery turret is about to fire or where it's pointing. The most effective way I found to reduce buffer was to put an inserter counter in every segment and add its values to offset the CC. About 35% material savings compared to dumb offset but things like landfill and cliff explosives are variable and a counter can't anticipate that

[plain hare]

What I do is just count the number of robots in use and only let the artillery inserter load when robots in use are low

That indicates construction is done or near done

Before I do too much optimization of my own sushi rails I’m focusing on a sushi starter set to rush their automation.

Editor tests are only so good

[timid cosmos]

I think the BP is missing some wires? Or I don't know how to use it. Isn't the Y value the control?

[normal gulch]

i should explain

u have a green wire going across

the inputs of the arth combinator

at the top is a row of active providers for trash, + the 2 combinatosrs for however many items/m

the combinator connected to a speaker is to tell me if im going over the belt speed

the R in the constant combinator in that last image does nothing, its for something special I am planning on to add

[timid cosmos]

So I need to be inputting constant item signals somewhere? I thought the Y value and filter splitter was enough?

[normal gulch]

the Y stays at 2.7K lemme test something

[timid cosmos]

or 900 if I've downgrade to yellow, yeah?

[normal gulch]

yes

[plain hare]

There’s a green wire reaching across flow limiters and that’s where you input constant item signals representing item / minute desired

[normal gulch]

u can also do it wouthout a central CC bank, just put that I in each constant combinator for the items/m u want for each item

so like this is half a yellow belt

[timid cosmos]

i've plugged my constant signals in the green wire. the belt still never activates?

[normal gulch]

replace I with the item ur limiting

[timid cosmos]

last time i played with this circuit it switched on and off real fast

oh

[normal gulch]

so for the limiter u change the splitter and i signal and thats it

[timid cosmos]

oh hey its working

[timid cosmos]

Sushi Rail is a Go

la cocina

[normal gulch]

awesome :D

i wonder how fast in a straight line it can go

[timid cosmos]

i'm liking my design where it cuts off the flow when the chest is satisfied

[normal gulch]

thats really cool

[timid cosmos]

I need to set deadend splitter filters to avoid stuff getting stuck in there

[normal gulch]

yeah that gets expensive if u dont

[plain hare]

Feels bad when you see a roboport sitting on one for eternity

[timid cosmos]

well it's completely functional now

sushi build these straights

[normal gulch]

time it? see its speed

[timid cosmos]

kinda slow

[hoary trail]

Made a tiny bot-run sushi kitchen

It won't be viable for early game, so I'll have to design a much cheaper version if I want to use it in early game

[hoary trail]

Also I think mine has somewhat similar behavior to yours except that I placed the provider chests along the rail that connects between intersections, rather than at the intersection itself.

So there's always materials running by it to supply the inputs

The near belt will never have any items flowing on it once the final intersection is built, as the RHB (right-hand belting) will always force it to go the long ways counterclockwise.

The clockwise near belt is actually only there because the bots from the previous intersection can build the two near roboports as well as that belt that runs past them, which is enough to get it started and then the new roboports build the rest. The near belt makes sure that while the final loop is not yet complete, the sushi still has somewhere to flow.

[timid cosmos]

Yeah I ran that problem and had to redesign some stuff

[hoary trail]

Well, my pic isn't loading

Anyway, my point was, by moving the chest to not be along that near belt, I was able to implement it without any circuits at all, as the near belt doesn't need to be activated or deactivated after it's all built, since it'll naturally deactivate due to the splitter priorities. So perhaps you could get away with much fewer combinators?

Now the pic loaded

[plain hare]

I feel proud to see all these up and coming sushi rail designs lmao

For a bit there I thought I was going completely insane when I first had the idea

It’s absurd at its face xD

[hoary trail]

There were a lot of doubters

Mine is actually already ready for a test drive, I designed it to integrate seamlessly into the rail prints I already had made so I can just stick my sushi rails on the end and start building

I'm considering adding some defense though, I haven't decided yet.

[timid cosmos]

I'm going to design a belt-based mall that just makes the stuff for the rails

so i can deploy this right after chem science

[hoary trail]

I need to do that too, but for now I just made a bot-based mall instead for later in the game

[hoary trail]

@plain hare I've seen in your prints that you use yellow sushi belts. Do you plan to ever upgrade those? And if so, how do you plan to do that? I don't know of any way to paste a red belt print over a yellow belt one without mods, and it also sounds like pain to have to upgrade a whole long line of belts over a vast expanse of territory

My current idea is, when I get to the point when I want red belts instead of yellow, I'll just build a second mall and merge it with the existing sushi belt to double the items, and then split them back up when the sushi returns so that half goes back into the new mall and the other half returns along the yellow belts to the old mall

I can do the same for upgrading to blue belts, too

[timid cosmos]

oh no star

you don't know about the Upgrade Planner? 😲

the little green button next to your hotbar, click it, put it in your inventory, and right click it

your lifes about to change

oh i see what you're saying

how to return them to base

that'd be fairly easy, unfiltered storage chests in each cell, and a request chest, possibly with requests set by circuit, outputting back into the sushi

So You would do the the upgrade planner over the whole system, stuff goes to the storage chest in each cell. gets requested, and put back on the belt

this means you would need logistics bots in every cell as well as construction of course

Actually I'm way over thinking it

but that would be the most dynamic, potentially? An easy setup is just put a storage chest in each cell that outputs back into the sushi, maybe with priority.

Sort it out later

that would be simplest. I don't think the requester chest idea actually has any utility over just a simple storage chest, in this case.

just adds a middleman

[plain hare]

Actually the way to return stuff to the home mall is having the only unfiltered storage chest in the network feed the sushi belt. My rails already do this 🙂

And tangled - yeah, definitely a good idea for a belt based mall to feed the sushi immediately after Chem science. I have one of those too

The storage chest loads both sides of the belt, one side with inserter, one with side load. Either way that means priority is given to the items already on the belt, and the trash will only fill in gaps on the sushi belt.

As for the blue science tier mall - I have all the things needed to grow production available in it. I’ve calculated something like 100-300 items / min of production equipment to be added to the belt in addition to track building stuff, and this can replace an early game building train, allowing creation or something like 20 yellow belts of ore -> plate -> intermediates per hour.

What I’m working on now is the starter base to get through blue science and efficiently build the sushi mall to run the sushi rails.

I will include a small amount of purple / yellow science in the TM sushi starter set for getting beacons, coal liquefaction, assembler mk3, blue belts. Stuff that will be useful for large scale growth. From yellow science , likely military upgrades just to make my tank more powerful to clear room for the sushi rails.

Once I claim enough ore -> I’ll get spiders, and with even more ore, switch to full artillery based sushi castle.

With more ore, new sushi kitchens / malls that are faster, and builds that rely on higher tier modules. Eventually upgrading red belt / blue belt in the sushi rails.

it’s a plan. I’ve had it for a while but I’m slow xD

This is my speedrun 10k spm plan in 24 hours

I think it’s possible

[hoary trail]

I think I'm currently leaning against upgrading the belt tier tbh

Things like walls, turrets, higher belts and such aren't really necessary at the beginning of implementing the sushi

And it seems like a waste to upgrade to red belts over stretches of land that are already built and running just fine

And I can just feed the yellow belt sushi into the red belt sushi (filtering out things I don't want, like the yellow belt stuff)

[timid cosmos]

Maybe a good idea to install the system in your build, if you ever change your mind

[hoary trail]

Well, I'll have red belt versions

[timid cosmos]

I could see upgrading to red belts when you're rich, but blue belts...i dunno

[hoary trail]

But I'm just planning to leave the stuff that's already built with yellow belts alone, and the the new stuff after I'm ready to switch will be red belts

I'll design a special junction for the two sushi belts and feed a second mall into it to supply the rest of the red belt

I didn't really think about trash chests though, that's an oversight I need to fix before I implement it in my base

I'm not sure what I want to do with the trees, I don't really need any more wood

[hoary trail]

OK I've added a trash chest to slowly trickle the trash back to the sushi kitchen where I can dispose of them in a giant chest array.

(jk) I probably won't do a giant chest array for trash, but I do need to figure out what I want to do with it

[plain hare]

Remember upgrading belts in sushi rails is way better than other systems

Upgrading belts means you can keep the same item / minute rates but you have effectively less buffer per belt

So it’s very possible a belt upgrade could overall make a system cheaper

If you upgraded and kept the same items / minute you may see your sushi kitchen surge with items worth more than the belts themselves

[plain hare]

Send it home! Coal liquefaction!

[hoary trail]

Haha I could do that but that doesn't really save me much coal in the long run. I'm kinda thinking about making a small train in a tiny loop to just run in a circle and burn wood in a pollution free way. It would be completely unnecessary but kinda funny haha

[hoary trail]

Red belts are rather expensive though, and looking at the numbers I'm using far more belts than anything else. When the other items in the system are in the hundreds, the belts are in the thousands.

[foggy junco]

Haha, im using blue belts for sushi

Though no matter what belts you use, you should make a dedicated belt production just for sushi

Be it yellows, reds or blue

[hoary trail]

Yup, my sushi kitchen has its own small mall to run everything, including belts

[foggy junco]

Otherwise if one belt isnt enough you can add a second yellow belt

[hoary trail]

The rails are now implemented in my current game

[hallow sonnet]

LHD?

[hoary trail]

Yes

I'm partial to LHD as I prefer the intersection signaling over RHD

The sushi belts use RHD though (which shouldn't actually matter, I just didn't feel like flipping them after I'd realized I was going the wrong way)

[foggy junco]

That doesnt look that masterful to me though

[hallow sonnet]

sushi without algae?

\seaweed

[timid cosmos]

that is some pretty sad looking sushi

[timid cosmos]

It's nigiri. it's perfectly fine sushi, but that right there looks particularly sad

[hoary trail]

I haven't automated the actual outposts themselves yet, but the sushi rails are working as intended and I'm happily building away

[foggy junco]

Also looks like thats not the rice that is normally used

[hardy fable]

why 😢

[foggy junco]

the fish has a sad color

and i believe either that's the wrong rice or it has been prepared wrong

[hardy fable]

it's perfectly cooked japanese round rice with (somewhat) properly cut sashimi-quality red tuna 😭

also was the first time I made sushi

[foggy junco]

how did you prepare the rice

in a rice cooker?

[hardy fable]

rice cooker yeah

[foggy junco]

what kind of rice cooker makes the rice look like this

it looks wet

[hardy fable]

I have to admit it's missing the rice vinegar, but rice cooking is still pretty new to me

[hardy fable]

it is a bit, I mean you have to put a lil bit of moisture on your hands to make them

[foggy junco]

my brother makes sushi from time to time, and i believe he uses a different kind of rice

give me a minute

[hardy fable]

I use either akafuji rice or wei ming crystal rice

[foggy junco]

He uses mirokumai sushi rice

[hardy fable]

"akitakomachi" oh, I haven't heard of that one, I guess there are many

[foggy junco]

pretty much one of the highest quality ones out there iirc

[timid cosmos]

sorry i hated on your sushi, its probably better than i could do

[foggy junco]

i mean as long as you don't use jasmine rice or something i guess i can approve haha

[hardy fable]

basmati sushi lmao

[foggy junco]

please no

[hardy fable]

no prob, it was the first time I still have a lot to learn and experiment, I just shattered how proud I was of crafing my food instead of my usually pre-prepared meals XD

[foggy junco]

haha, i mean it takes a long time to master sushi

[hardy fable]

I also made some onigiri a few weeks ago but it's super hard to shape and give it the right width of rice

[timid cosmos]

I've never even made white rice, i always make spicy brown rice

[hardy fable]

a few years ago I tried to cook round rice with a pot, man how much I was not prepared how hard it is

but somehow a rice cooker manages to do that very hard thing very easily, it gets it perfect all the time

witchcraft

[foggy junco]

have you made maki sushi yet?

[hardy fable]

I'm not sure I'm yet ready for that

tho I have the sharp knife required

[foggy junco]

haha

[hardy fable]

but I sure will try someday when I'm more confident in the prep and handling of the rice

[foggy junco]

Uramaki is where the fun starts

[hardy fable]

heresy! it's not even japanese!

[foggy junco]

Haha

[hardy fable]

like frozen mochi lmao

[foggy junco]

They do serve it japan according to my brother though

When he was there

Anyways

With nigiri sushi the square ish kinda looks odd

You might wanna cut the fish into an oval-ish shape

Or a convex kinda-ish looking lens? Idk how you call that shape

[hardy fable]

it's eye-shaped if I start from a filet

[foggy junco]

Square fish on a round-ish rice ball looks off

you can make the rice squared and have square shaped fish on it, but mixing the two shapes just looks... off (?) i gotta say

[hardy fable]

ah, I can't find the name back, but for sashimi or nigiri sushi they cut the fish into a slab, thus resulting in square slices

[hardy fable]

aaah, I see!

[foggy junco]

lemme see if i can find some pictures

aaaand of course now my browser decided to think this is the perfect opportunity to freeze

[hardy fable]

there's probably too much rice anyway

[foggy junco]

like in this image

the rice is also a rectangle

[hardy fable]

and it is layed out flat

[foggy junco]

while here it's round-ish

[hardy fable]

kinda looks like a lobster from afar 😄

[foggy junco]

haha

you kinda don't want your fish to be bigger than the rice

[hoary trail]

@plain hare sushi rails are officially a success. I just built this outpost without ever having to be physically present at the site. The sushi rails built the rails and a builder train station, then a builder train comes in to supply the materials for the mining patch. After everything is built, the bots deconstruct the building train station and the outpost is ready to go.

[plain hare]

Glorious. Absolutely glorious.

Can’t wait until I have the same thing going on but sushi castle all the way in to the ore… 😄

[hoary trail]

It is really amazing haha. One of the tricky things I've learned is that because different items have different stack sizes, I have to have several station blueprints to accommodate different train sizes and also different stack sizes. I wish there were a way to change a combinator's settings remotely, or to change wiring without having to blueprint it in

Also do you remember when I was still struggling to understand basic active measurement sushi? I feel like I've learned so much since then

[plain hare]

yeah youve been making progress at an excellent pace for sure

youll be making megabases soon enough lol

[distant gorge]

You can put down a combinator near you, set it as needed, copy, then paste over the far one

[hoary trail]

that doesn't work for wiring, though, right?

I mean I guess I could still copy paste a wiring but I would need to delete all the old stuff first

[distant gorge]

Ideally the wiring isn’t going to change from station to station I would think, just the configuration

[distant gorge]

okay I think I have a set of rules for my Angel's Nuclear sushi

essentially only U 238 and Plutonium will be freely circulating, everything else gets sorted out immediately or goes to specific machines

So the primary plutonium and U 238 producers get gated behind thresholds on the circulating belt, the primary output (thorium cells) gets clocked to the desired rate, and everything else operates on backpressure or greedily consumes a byproduct

now to implement it >.>

[hoary trail]

Angel's sounds terrifying

[distant gorge]

You get used to it after a while

[plain hare]

Yesss more sushi. Goooood, gooood Anakin. cackles in palpatine

[distant gorge]

Looks like I'll have about 30 GW to play with if I need it... just going to burn it off for now I think

I think that's it for the actual building... time to set up the controls

I'm thinking of actually doing more of a meter sushi instead of an exact counter

do an exponential average with period equal to the round trip time of the belt, and use that to measure

[distant gorge]

It's working! taking a while to spin up though, and internet isn't stable enough to upload the image at the moment

I have the meter set to aim for 10 U 238 and 5 Plutonium per second

it'll take a while to get there

[distant gorge]

https://cdn.discordapp.com/attachments/231518135231053825/951340834069037076/unknown.png

First thorium fuel cells have been made

The meter, just an EMA with a negative offset on the output, then each endpoint producer/consumer just checks if its component is above or below 0

So I set the flow of 238 and plutonium I want by changing the offset, and the endpoints adjust to accomodate

[clear latch]

I made a compact steel sushi smelter with and without any . It keeps up to 77 steel and 65 coal in each of the middle belts, with the coal mooched off the sides and input prioritized at the bottom. The 77 is also used to disable the iron ore furnace inserters, got that idea from @plain hare 😄

[plain hare]

That’s a nice way to do it. Requires the output to get routed back to the input side, but it’s still a nice way to do it. Super cheap.

[hoary trail]

@plain hare we talked about how you wall in your rails of your sushi castle, but do you wall in your miners too?

[normal gulch]

why not? they are super polluters

[hoary trail]

Unless the rails go in front of the ore patch, then the biters will just attack the miners

Well it was just something I hadn't considered

I've always been one who expands out to a good chokepoint and then doesn't worry about having to wall in individual outposts

But if I'm going to wall in my sushi rails, then I would also need to wall in my outposts themselves, I guess

[normal gulch]

yeah thatd be what I do

[hoary trail]

That seems a whole lot more expensive

[normal gulch]

what else are u gonna do?

[hoary trail]

Just doing walls at chokepoints is a lot cheaper

[normal gulch]

Yeah, thats why i dont really understand this sushi rails idea

[hoary trail]

That's not the only reason to use sushi rails 🤔

[hoary trail]

So far, the system that I have is pretty much just as cheap as regular rails, and much cheaper and faster than if I wanted to do a roboport grid to build them for me.

[normal gulch]

Yeah but the walls confuse me i do walls very differently, at chokepoints

[hoary trail]

That's why I'm considering not bundling walls with my sushi rails

[normal gulch]

for my sushi, i have an idea for a compact way to get requests done

so items that arent usually on the belt can be

[plain hare]

Well, I’m kinda planning the sushi rails for worlds without choke points

From my point of view they’re nice, but they’re kind of a crutch for defense

If your defense relies on them you just can’t really do land maps

Im hoping with artillery and eff modules I mostly won’t need walls around outposts. But, I’m gonna design a thing that just attaches to my sushi rails and builds walls around an outpost if need be, either landmine or full

The other nice thing about the sushi castle is unidirectional expansion - you can mostly go in one direction and get really rich ore patches.

Oh the other advantage of the sushi castle is automated artillery emplacement clearing of aliens with low losses

It’s super super safe so you’re not gonna be getting rekt by big counter attacks

Especially cause you can always deconstruct the front castle and just scuttle your losses and fall back to the next line

[hoary trail]

@plain hare you don't do artillery range, do you?

[plain hare]

I’m not gonna do much but maybe 1 or 2 levels could be wise. Not super sure there, I’m not gonna do much

[hoary trail]

Have you calculated how much the rails cost with all your defenses added?

I'm concerned that it seems like it will cost quite a lot

[plain hare]

It does cost quite a lot yeah, but that’s why I have the early game landmine model, and low pollution stuffs, gonna claim tons of ore with low evolution aliens with tanks and spiders before I go for artillery, and then I’ll claim more… and more 😄

[hoary trail]

So I'm a bit late to the game with my sushi castle, as landmines aren't viable anymore

[normal gulch]

The 128k cost is where I stopped

[hoary trail]

I'm considering having occasional artillery outposts to clear out the biters in a large radius, and then just building regular rails out to the edge of the arty and building another arty outpost

If I do that, the arty outposts should protect against nest expansions, and I should need minimal defenses along the normal rails, maybe just a wall and a few lasers here and there

I would need some decent arty range to pull that strategy off, but I have a few upgrades already so I think it should work well enough now

[plain hare]

That’s basically the plan with the sushi castle but it’s just the more range you get the more ups hit you can expect

So there’s some middling range I think that would be good

[hoary trail]

Do you really need full defense then on the rails in between your arty locations?

As long as I don't put giant polluting factories out near my mines, I don't think pollution attacks should be much of an issue

[plain hare]

Well if you don’t protect the rails they’ll cross them to get to your artillery

And I don’t, my system is modular so I can have landmines between artillery sections

[hoary trail]

Hmm I guess I can't fully control which arty will shoot an expansion so I do need some kind of defense

Maybe some walls and some turret would be sufficient for regular rail defense

Are you just relying on landmines for between your artillery sections even when you get to behemoths?

Also do landmines trigger nearby landmines if put too close? I wonder how many upgrades I need for a landmine to kill a behemoth

[clear latch]

I've used train-supplied artillery outposts in a giant overlapping circle outside my pollution cloud - DeHackEd's base was the inspiration. The outposts went at the ends of long undefended spurs that branched off perimeter mainline tracks. No regular stops were scheduled outside the perimeter track. This provided a buffer between the artillery and mainline, but the spurs were rarely attacked anyway.
For a sushi supplied outpost you could just split off a long belt - like a sushi rail minus the rails and signals. The drawbacks are the massive area to cover around lategame pollution clouds and the expense of building out that far, but map view building should make that less tedious than when I did it.

Just tested landmines, they don't seem to explode each other and need 7 detonations to kill a behemoth biter or 4 for a spitter before infinite research

[plain hare]

They won’t blow each other up, and the plan is to keep evo extremely low until I claim lots of ore so I can just go full cost sushi castle and not really bat an eye, ideally all before behemoths for max deathworld

[hoary trail]

Well, I am no longer low evo haha

So I need to probably accept the cost

[plain hare]

Potentially, I think it can make back the cost pretty fast, it’s just an expensive initial investment

I think once you claim some super rich ore patches tho as you expand it will all become very worth it

[hoary trail]

Built my sushi belt upgrade kitchen

It merges two different sushi belts together and filters out the yellow belts

I'm sure there's better ways of doing it, but this seems to work and is pretty stable

And merging/unmerging sushi belts was a fun experiment

[timid cosmos]

You could make it dynamic and automatic when you change your constant sushi kitchen signals, using a bank of filter inserters that remove from the sushi belt. You would take your constant signals for your kitchen, and output EACH as a value of 1, inverse it * -1, plug into the inserters that takes stuff off the sushi belt. Then read the belt right in front of the inserter on "Hold", send the contents through a decider combinator, ANYTHING > 0, output EVERYTHING=1, and wire that to the inserters. Stuff that is supposed to be on the belt will count as a zero when it crosses, and anything not belonging in the kitchen outputs a 1, setting the filter inserter's filter.

[clear latch]

Also works with a negative CC or arithmetic with requests <= -2 and read pulse

[timid cosmos]

true

[timid cosmos]

you still want to use Hold though, pulse doesn't give the inserter time to grab it

[hoary trail]

That's what I'm already doing actually. I have two nonvariable inserters (although I'll be increasing to 3 for the Mark II rails), one for belts and one for rails. Mainly because those are the highest throughput items and I don't want them holding up the other inserters because it slowed down my building times during testing.

The other two inserters are dynamic and allow me to easily add a new item on the global sushi combinator if I should need a new item

I also do some regional dynamic stuff as well. I have logistic bots in the system, but most blocks don't need logistic bots. So if I do need logistic bots, I drop in a small addendum to the regular blueprint that allows it to also pick up logistic bots for that block.

I'm planning to add nuclear fuel to the sushi belt as well, so I'll need to do the same for the nuclear fuel.

[hoary trail]

The reason for that giant sushi belt merger is because it's designed for my transition from the Mark I to the Mark II sushi rails, which reuse a lot of the old components but add some new ones and use red belts instead of yellow. I didn't see any point in converting the old yellow belts to red, so this is designed as a merger between the two belts that filters things toward whichever sushi kitchen is in charge of them.

It also makes sure that the items that aren't needed for the Mark I system don't flow that direction, as the yellow belt would back up if I sent all that down that way.

[clear latch]

Nvm I just remembered filter inserters have a Blacklist mode, you don't need negative values at all

[timid cosmos]

In all my hours I've used blacklist like maybe twice, because I always forget about it.

[hoary trail]

I used negative numbers in my setup

[clear latch]

This is probably gonna be a bit of a pain to build before I actually have military science, but it seems to work so far. At least I don't have to worry about rushing the medium poles since all the furnaces are for military packs only

[foggy junco]

That seems super neat

[plain hare]

I like that. Maybe I’ll remake my starter bunker with 1 active sushi belt

[hoary trail]

A slightly odd approach to sushi lol

I'm sure there's a better, more compact approach using filter inserters, but this is simple and runs no risks or being limited by throughput

[foggy junco]

Eh

What is that

[hoary trail]

It's a sushi belt splitter. I merge two different sushi belts to form my red belt sushi.

So when the sushi returns, I need to split them back up again so I can send each to the proper kitchen.

[foggy junco]

So you split up the sushi into two belts?

Why not have it one sushi belt that rund through both of them

[hoary trail]

Because I wanted to recycle the materials I used for supply belt of the yellow belt sushi design

So when I am ready to upgrade to red belts, I build a second kitchen, and feed everything except for the yellow belts into the new red belt sushi. The second kitchen supplies the other materials that I didn't previously need.

[foggy junco]

hmm

so two different kitchens

[hoary trail]

Well, it was that, but I broke it with my trash chest system so I decoupled the two sushi systems and just made a new kitchen for the red belts.

I could have done it still, but I would have had to control the flow of the trash into the yellow belt system and I didn't want to bother with that

[foggy junco]

[hoary trail]

It's now just two completely separate systems.

[plain hare]

How do these sushi belts look https://media.discordapp.net/attachments/139677590393716737/954042367244124180/unknown.png

[clear latch]

Looks organized and easier to build than some of the MP blueprints I've seen, interesting side effect of using sushi instead of only spam. I also like the stone brick paving stacking on top of belt speed bonus in a clear path up or down, these all-in-one prints are usually a pain to wade through 😄

[hoary trail]

Looks very nice, it inspires me to make a better starter base haha

[foggy junco]

Nicee

[clear wedge]

Blueprint String

Vanilla Wirerate

[clear latch]

Simple calculator to convert from items per minute to items on belt of length X, for active control without combinators. It assumes items on belt/belt capacity is directly proportional to items/minute. Not hard math, I'm just lazy 🤷

[timid cosmos]

interesting, i could see that being useful

in the past i've done the easy math and then lowballed it

[glass narwhal]

Sushi thread is active again? 🇳 🇮 🇨 🇪

[clear latch]

So far the calculator numbers are working for red/green science, apparently I put too few products and too many plates before, no more guesswork. Production graph still looks horrible but I figure that's due to machines throttling when science packs are detected on the belt

[plain hare]

https://media.discordapp.net/attachments/334735604342325249/965086283019997224/blueprint.png

Working on my 45 spm sushi bunker. You can see it’s right up against flamethrowers, on the other side, short smelting lines.

Plan is this will let me conquer hard worlds where I can’t immediately start with my 90 spm set.

[foggy junco]

sushi bunker

nice

im taking notes

[glass narwhal]

so idk anything about sushi as i haven't watched a tutorial or anything but will this approach work which i just thought of. I know Circuit cables can see the contents of a belt and a belt can hold 8 items at a time. so if i am using the hold mode on the belt can i make a loop of belts and calculate the maximum number of items that can be in the loop before saturating it. so if there is a very short loop of 4 belts can i assign make inserters only become on when items are less then 4 or a number i have decided on, 32 being the max number of items on the loop so i can have 8 different items . btw limiting inserters to 1 item to avoid overshooting.

i know there will be a better way but i came up with this myself by just hearing about some sushi talk and nothing some bit of how to use circuits

[plain hare]

@glass narwhal what you describe is almost exactly what’s happening here

I’m measuring the contents of the sushi belt with hold and telling inserters to enable when contents < X

I’m also connecting some of the chests for end products, like cliff explosives, splitters, undergrounds - those don’t need to constantly be on the belt, so the inventory of the chest for those and the belt is combined

[glass narwhal]

oh so i guess sushi isnt as complicated as i thought it is

[plain hare]

Well, some methods are more complicated than others.

But this one is a great intro.

Less complex than others I’d say

[foggy junco]

i wouldn't recommend using memory cell sushi.

It's complicated to setup and manage, and there's a lot that can break memory cells

mostly i'd recommend for either active measurement (you measure the content of your belt, and just add stuff if the amount of the belt is too low), or go for flow-limiting sushi (you limit how much enters the belt)

one easy way to do flow limiting sushi is to have two yellow belts go into a splitter onto a red/blue belt

(and then have that loop back into the input)

[clear latch]

For dedicated inputs you can just read the belt by the first inserter to reduce items backing up. Active measurement/belt capacity is also directly proportional to flow rate. For example, 10 items in 10 segments at the start (10/80 belt capacity) works out to about 1/8 belt per minute. It's clumpy/spiky but that's ok for things like ammo or coal. You can even input prioritize by disabling the machine outbound belt/inserters as a secondary condition when items are read on the return belt.

[distant gorge]

I use active measurement for the one sushi belt I have, you just have to take care that it never fills up or jams in any way

[timid cosmos]

an active system should never jam. usually it's memory cell designs that get all screwed up

[clear latch]

It is possible to get an input sushi situation where the loop isn't 100% saturated, but the inserters or input belts wait too long for space to put stuff. That can result in unusually concentrated clumps that take too long to circulate to the right machines at the right time. I've avoided it by putting things in specific lanes with enough free space in the outserter lane and probably at least 1/3 space unused overall. Output sushi with only final products is usually a lot less crowded in my experience

Also too many undergrounds and splitters can hide items from the circuit network long enough to overproduce a lot. In one place where I had a red underground I ended up deliberately underordering

[distant gorge]

Angel’s probabilisticness throws a wrench in that sometimes

[timid cosmos]

I could see that if that's what you're using the sushi for

[distant gorge]

It’s the nuclear thing yeah

[timid cosmos]

I can think of a few ways to safeguard it

[foggy junco]

Using below a certain % of the belt should keep that from jamming

[plain hare]

So I’ve been running this design near capacity

Notice the inserter assembler, it’s inserters, it was dropping off farther back on the belt than it was grabbing

Normally you don’t think about that stuff cause you have separate belts

But by switching it I resolved the localized flow issues I was having around that assembler.

Lesson: when using sushi, input on the assembler then output (along belt flow direction) the input inserter will leave space for the output

Went from high uptime on green science assemblers to near full uptime

The belt was barely maintaining 10/50 requested inserters and changing the limit fixed it.

[plain hare]

thought you all might enjoy this size comparison

[distant gorge]

You’d think that, yes

When an item backs up onto the main belt that goes haywire though

[foggy junco]

You mean the output?

[plain hare]

cramming it all in...

[normal gulch]

lol arent they already compact enough :P

[plain hare]

No

I saved 1 tile over the non sushi design I would have gone with lol

I’m getting there https://images-ext-1.discordapp.net/external/zlL9htwnKzsYFcm0kKHL15fsdny2jbwz7j6-e-fTf4s/https/media.discordapp.net/attachments/139677590393716737/965439762217463809/unknown.png

Bit of an older image but you can see the space constraint im working with

[normal gulch]

theres also the fact of being easy to build

[plain hare]

I think these sushi designs will be pretty easy to build

For one thing it’s literally less to place

[normal gulch]

just dont make it like the one we did in MP

[plain hare]

Yeah no that was kinda butt

[normal gulch]

that was a lot of underground spaghetti

[plain hare]

I’ve been cutting down on that a lot

[normal gulch]

[plain hare]

I think this base is ultimately gonna be my bot rush base

It just seems right

I think I’ll extend the wall in the direction of the smelting lines to encompass the ore mines in a real world

Then rush flamer and get second ore patches

Oh and when I rebuilt the refinery I ensured the flamer fuel has to be heavy oil so enjoy that

But in the process of doing so each flamer during the switch from full crude pipe to low buffer heavy oil pipe will save me about 500 petro gas per flamer

[normal gulch]

I guess if it saves resources

but its not easy for first getting oil to use crude as defense

I guess just run a heavy oil pipe back

[plain hare]

I think you just run the crude pipe back and get 75 blue science then heavy comes out and you use that

Could shoot some crude out of flamers at the oil patch briefly if you had to also

[normal gulch]

is it pollution efficient to use heavy?

[plain hare]

I think so yeah since you’ll need petro gas to be pollution efficient

Refinery makes 660 gas / minute and 540 light oil / minute and 300 heavy oil / minute for 6 pollution / minute. And the Chem plant cracks 900 light oil / minute or 1200 heavy / minute. for 4 pollution/ minute. So combined crude -> gas (no heavy oil) is 1,020 gas / minute for 8.4 pollution / minute. Or 121.4 gas / pollution. When you add in the heavy cracking, it’s an additional 150 gas / minute and an additional 2 pollution / minute. Or 75 gas / pollution. Meaning the most pollution efficient route to petro gas is by not cracking heavy

Yep so every time you shoot heavy oil out of a flamer you’re ensuring more petroleum gas, and more pollution efficient petroleum gas

I did the math finally 😛

Another way to write that is if you do the whole chain of oil -> petro it’s 112.5 gas per pollution but if you skip heavy cracking it becomes 121.4 gas / pollution

Which is about 8% more gas per pollution

Wow I think I just officially turned into a heavy oil Stan for flamers

Final plus for heavy oil: it’s great once you get coal liq

[foggy junco]

what if heavy oil -> light oil -> solid fuel?

[plain hare]

Well, first of all, I don’t think solid fuel is all that valuable, but second I think heavy oil -> light oil is still a reduction in fluid so you’ll see the same kind of Math if you run the numbers

Just a little less pronounced

[normal gulch]

Lol well congrats lets do it next game

I’ll still always use light for the 10%

[plain hare]

I think light makes a lot of sense once you’re heavy cracking with more prod bonus

Cause then 100 heavy becomes 97.5 light with prod 3

[foggy junco]

fluid sushi let's go

kinda stole this idea from reddit haha

[pulsar lance]

I am surprised the pipes are clearing?

[foggy junco]

Yes, though when i make it twice as long for example ill have to put extra pumps to prevent the last refinery to get clogged

Lemme see if i can find the reddit post to the original one where everything is explained

https://www.reddit.com/r/factorio/comments/hd5cwb/fluid_mixing_oil_refinery_exploration/?utm_medium=android_app&utm_source=share

Wouldnt recommend using it with beacons like this guy does tho, if you watch you can see that refineries are inactive like 1/4th to 1/3rd of the time

Since its simply too fast for the fluids to empty

Works fine for early game refining though with low enough production rates

[plain hare]

Very interesting stuff. Fluid sushi 😛

[foggy junco]

I mean

Look at how compact it is

[plain hare]

Yes… though adding the cracking could be interesting

[foggy junco]

Eh, you filter the stuff into heavy, light and petgas and then do the same input like oil/water

[plain hare]

I just wonder if the space impact of that would make this take more space than the crackfinery

Especially with the slight downtime on refineries

The refinery stack gets compacted by 5 tiles with this approach but due to the pump / tank shenanigans you may lose out on space yet, you should see about making a perfect ratio model of this 20 refineries and see how the size comparison is

Though maybe it could be used just for an earlier game smaller ratio plant. Dunno.

[foggy junco]

there's none in the unbeaconed one

just got home from university, so rn i'm working on a full setup

[foggy junco]

seems like it'd be most efficienct with sushi input but kind of regular output

scaling it up creates some problems

or maybe

i'm not using loops correctly here

we need 1x1 pumps

those 0.0 fluid floats

are ruining everything

[pulsar lance]

Ask a modded to either implement 1*1 pumps or fix that 0.0 Problem

[distant gorge]

People always complain about the floats but I have never ran into it being an issue 🤔

It just baffles me

[clear latch]

Advanced Fluid Handling has 1x1 pumps but they require undergrounds on both ends

[foggy junco]

well, it does in fluid sushi haha

0.0 fluid in the pipe but it isn't empty

so other fluids can't pass through

[distant gorge]

Ah fair

[hardy fable]

you can get rid of 0.0 fluids in vanilla factorio tho

just don't put too many pipes between pumps and don't put pipes between tanks and pumps

like so:

well, to be clear the shape of the piping is more important than its length

like stubs are fine

but loops have a tendency to mess the emptying

[hardy fable]

so uh... I managed to fetch all fluids from a merge of all of a refinery outputs but I somehow broke the game 🤔

the tanks are not full, and the pumps that are not wired are not empty either

like what

ah shit it just jammed

wtf is happening

[foggy junco]

1. do the pumps have the correct fluid in them and 2) are the tanks they input from the main tank have at least 15k content at all times?

and is there any fluid in a pipe with "0.0" content that blocks the rest of the fluids maybe?

it's always those guys that break it for me

or these ones

[hardy fable]

you can literally see it in the picture 😄

the pump above is not running

although it contains the same fluid as the tank below

the theory was to fill the tanks to 25k and fill its upstream pump, then the downstream pump would be active only if there would be stuff in the common tank, so that the pump would never empty, thus would never pick another fluid

now in practice not only there is this weird behavior, but if there is between 0 and 1 fluid unit in a tank, the circuit network returns 0, which breaks everything (tho again in theory, the pump would be emptied enough to suck any remaining decimal)

I have no idea what kind of black magic is at work here

thinking at it again it gave me a hint, I managed to make it work :D

it's finally flawless

the issue was that the downstream port was waaay too overpowered, I limited it to like 200 fluid per second

broke again, whyyyyyyyy

it was running fine

ffs

[sly parcel]

Da fuck is sushi thread

[hardy fable]

a place where you make sushi :D
#930147588160782436 message

found a fix (the testing is in progress)
I just block the downstream pump if the dedicated fluid tank is not full

[foggy junco]

huh

[foggy junco]

in case you don't know what sushi is;

belts with more than 2 different items on them

[hardy fable]

it's still in a weird quantum state but it hasn't broke yet, looks promising

[normal gulch]

and that doesnt jam

[hardy fable]

ok the refinery is running entirely out of two pipes

next challenge: run a refinery with only one pipe

but that's gonna be for tomorrow

[foggy junco]

Any luck on that so far?

[hardy fable]

it did not work as I expected

stuff remained in the pipe

I don't think it's due to the pipes being too long because I've already worked with longer ones

it might be due to different things but it's a bit hard to pin point, I might try to mod the refinery to run much slower in order to properly investigate that issue

but since it doesn't really have a purpose, I'm fine with the fluid sorter already

[foggy junco]

Yeah, those annoying 0.0 floats

[hardy fable]

yup

[cold yarrow]

That's very broken

[hardy fable]

ikr 😆

so from what I can see here, pumping oil and water in/out is working flawlessly

but when the refinery output stuff then there's bits of junk remaining in the pipe system

wtf

[foggy junco]

yeah, same issue i had as well

[hardy fable]

at least I have improved my fluid filter

condition on the pumps is just anything=25k

I managed to get any arbitrary amount of different fluids out of a pipe, but I still don't understand what gives the output priority of a tank

hence why some of them are unloaded from below, and some others from above

[hardy fable]

I made it smaller but I still don't really understand the priorities

oh

it's defined by the build order

if I remove all vertical pumps then rebuild them it makes that layout work

[timid cosmos]

[foggy junco]

I use the condition fluid >15k, that still leaves fluid in the pumps and allows them to empty out the sushi pipe/tank faster

[hardy fable]

note that this is entirely a bug tho :D

[foggy junco]

Is it?

[hardy fable]

why would a pump not empty its buffer in a non-empty tank?

this is black magic

[foggy junco]

Why would it matter in the first place tho?

Id call it a feature

[hardy fable]

well, it's convenient for our purpose that's for sure

[foggy junco]

Doesnt hurt in single fluid systems and allows for filters in sushi

[hardy fable]

as long as the buffer is not empty, everything is alright

[clear latch]

Condensed my sushi request circuit from +2 deciders for the reroute splitter and each added inserter to +1. It used to output 1 from unsatisfied requests and again from each inserter/belt segment pair, then set filter for each = 2. But just each > 1 implied a matching item on the belt so I use that now. Same for rerouting; output 1 from before the splitter to the exit segment, share the unsatisfied requests output 1, enable exit if everything = 1.

I remember trying this before and it failed because I read "hold" near the inserters and they just picked up arbitrary items. Pulse doesn't let them react fast enough to do that but somehow works when compared to satisfaction

[foggy junco]

Could probably make a more efficient design in term of combinators required if you want same output for all inputs

Wait, why am i rwqding messages that far back in the chat

Just realized

Wtf

[clear latch]

You can use 1 circuit (like the moving average one with 4 arithmetics) to track and flow limit many different items. Just need a secondary loopback with a chest buffer that disables the merged output until the circuit finishes taking its first average (1 decider everything < small number). And it can still be somewhat clumpy, that's why I ended up reading belt segments instead of using 1 decider for requests

[distant gorge]

I added a smaller belt loop to the main belt to function as a mixer to fix the clumpiness problem

[timid cosmos]

I don't see any wagon imbalance potential with my design am i missing something?

[clear latch]

The imbalance I pictured was the first wagon picking up items along a belt first until full, then the second, etc. Also any kind of uneven unloading could leave some wagons constantly near full, and their inserters inactive

[timid cosmos]

@clear latch

oh

oh i see. they aren't picking up til full, just until a specified amount. what i'm having is uneven unloading * from the mall *

because the constant signals don't cycle until a wagon is satisfied

it's not uneven in any way, just slow

[clear latch]

Ahh so you're keeping track of individual wagons?

[timid cosmos]

yeah

[clear latch]

Kanos doesn't do that it just goes brrt everywhere

All inserters set filter at the same time

[timid cosmos]

i like to use a constant combinator to specify what goes in each wagon.

here's a fleet on logibot fed construction trains

my design allows for me to change the cargo contents easily, if i don't like how i set it up, and stuff will be removed automatically into a

[clear latch]

That looks a lot more robust than the 1 CC I like to use (for the whole train)

[timid cosmos]

1 CC per wagon

it also reads to make sure there is actually a train at the station before signals get passed around

so items aren't left in inserter arms

[clear latch]

I ended up using a rail signal and 34 tick timer for that, for some reason train ID read the ID before the train was truly stopped. I know it's crude, there was a better way with a memory cell

[timid cosmos]

i like that idea

i haven't had any problems though

[distant gorge]

Oh my

[grim crag]

it's even hooked up for 2 trains lol (bottom combinators)

the top row reads the belt and pulls accordingly, the middle reads the chest to pull from the train, and the bottom activates the station when the chest gets low

[pulsar lance]

I love the use of 4 yellow belts

[plain hare]

That’s like the an artery tier bloodbus

[grim crag]

the entire base is running off of sections of 4 yellow belts

i always build for 4 belts of throughput, because why not.

[hardy fable]

last night I figured out that using sushi belts for input items to make engines, eletric motors and robot frames is quite efficient

and the electric motor doesn't even require recycling

(I'm still trying to figure out if I can get rid of the recycling on the other ones using weird trickery)

[pulsar lance]

With recycling you mean looping?

[hardy fable]

yeah, fetching what remains in the belt and reusing it (with priority over new stuff)

[pulsar lance]

Have you tried clocking the output perfectly? So everything will get picked up? (That sounds horribly combinator intensive, as everything might need its own clock lol)

[hardy fable]

yeah I'm using perfect ratios

the issue is that the picking from inserters is pseudo random

thus after running for some time the last assembler will eventually get a long succession of an item while starving on the other

it doesn't happen with electric motors because there are only two items

so what I do is that I put circuit on the furthest lane, and a mix of engines and circuits on the other lane, and what happens is that inserters first pick all the circuit from the closest lane, and thus the last assembler gets one lane of circuit and one lane of engines, so it can't clog

what I am wondering and is that maybe changing the position of the items on the belt (i mean which item is mixed with which item on which lane) might introduce enough entropy to prevend the formation of a clog, I don't trust it too much but it's worth trying out

[pulsar lance]

Make sure you try out for long enough with all edge cases… I once had a design which seemed to work, but finally failed

[hardy fable]

yeah for sure

I hop into the editor and crank the timescale to x64

I'm usually at 3800ups or so

for very short runs of assemblers (1 or 2) it works all the time for sure

[wraith badger]

oh hey ive experimented w/ that. sushi w/o looping/recying. and from my experience assemblers starving each other does happen even w/ 2 items (including electric motors) if you run the for enough time. so each assembler set needs to have a memory cell to track how much they take so they dont starve each other. w/ the last assembler being biased to pick up less so that its more lenient in picking up any leftovers

item positioning in my experience does not work and breaks at a large time scale. but honestly outside of sandbox situations where you run x64 times its usable w/o issues

here's my single belt no loop sushi setup that i didn't get around finishing for 3 items. goal was single belt. so w/ only 3 inserters available to handle 4 inputs/outputs, handling cycling of items was more complicated

stack filter inseters are only capable of only one filtered item while non stack filter inserters are capable of filtering multiple item... which somebody on reddit made me aware of when i posted my older build..... which was extremely helpful but i donto think i was capable of actually utilizing it effectively and resorted back to stack filter inserter

idk if anybody wants to expirement but i was at a point where i didnt know if i should proceed w/ cycling singular inputs through a stack inserter (which can only handle one input) through circuits or pushing multiple inputs to a filter inserter for them to figure out themselves and handling inbalances through circuits

[plain hare]

That does seem like an awful lot of combinators for the production line

[wraith badger]

yea definitely

it gets signficantly more complicated when you only have 3 inserters w/ 1 reserved for outputs and 2 inserters for 3 inputs.

w/ only 2 inputs, inserters are fine to pick up extra cuz they just have to wait. but w/ 3 inputs and only 2 inserters. if they pick up extra, they clog up being able to produce a single product

[plain hare]

Would you say there is a significant advantage to building sushi belts in this style? Especially compared to other approaches ?

[wraith badger]

no lol. honestly looped sushi is the simplest and often times the most effective.

i experimented w/ non-looped sushi in hopes of being more compact because i like compact builds

[plain hare]

My plan if I want to not loop the sushi belt directly would be to throw a train station on the other side and loop it with the train network.

Also, I see what you’re going for with the compactness but doesn’t the combinator count kind of ruin any ability for this to be compact ?

[wraith badger]

yea that's definitely a big consideration

i think 2/ inputs you can try to compete as shown in my first vid

but w/ 3 inputs definitlely can not compete at all w/o compactness which is why i gave up

[plain hare]

Each of those has a combinator hanging off the side though in the first one?

[wraith badger]

but it might be because my approach is wrong.

[plain hare]

There’s a ton of ways to make sushi belts so I wouldn’t say there is a correct approach :D.

[wraith badger]

yes but you can tuck them in. i placed it outwards for visibility... i didn't calculate production/area but i woulnd't be suprised if it performed worse

[plain hare]

Okay, makes sense. Yeah I was able to tuck 1 combinator per machine for my sushi smelter I made

[wraith badger]

for the record, the most serious i was was for compact oil where i calculated production/area

but it ended up not being compact so its more for enjoyment at this point

[plain hare]

I have been making some new oil prints myself, not for that purpose exactly, but the goals align somewhat.

I found some very compact stuff

https://images-ext-1.discordapp.net/external/hT6CCtWo7fdVMud5bJ13r8gt1hIlNGk817VHT0UVCTU/https/media.discordapp.net/attachments/139677590393716737/975399268758204497/unknown.png

I found you can contain refinery pipes inside ‘low space usage’ lines - electric furnaces good for this

I’ll be taking these layouts and ideas into at some point.

[wraith badger]

oh i meant for sushi oil lol

[plain hare]

Ah lol yeah I haven’t messed with that yet

Suppose it could do better in width, though the reduced length might hurt I dunno.

[wraith badger]

i like the pic. wat stood out for me was handling of refinering to chem plant piping

but idk if its becuz of energy modules where the layout is feasible w/o throughput issues. but i think its a good reference

[plain hare]

Yeah the energy modules mean it’s really not going very fast which allows the size

Though the layout itself can be adapted, perhaps might be shorter and / or require 2 sided input though

And they’re pollution modules *** feelsbad. Xd

Rn that’s at like 900 heavy oil per second so it’s nearing pipe capacity for that

And I think it’s near water capacity also.

I’d be curious to see your compact sushi oil setups

[wraith badger]

oh uve posted it already lol

[hardy fable]

ah, you're talking about short item starving, yeah it may happen

I've managed to make that stuff work by belt weaving the input into itself, up to the midway point

so basically what I'm doing is that after half the assemblers, I put everything on a single lane and put the feedback on the now empty other lane

works for the engines at least

[hardy fable]

I finally have internet at home again 🎉

[foggy junco]

flow limiting?

[hardy fable]

my internet or the setup? I was sending text from work but my game and the screenshots are at home

optical fibers are weak af they broke for no apparent reason

but 900mbps sure is sweet

[foggy junco]

Sushi

Flow limiting sushi?

Or memory cell

[hardy fable]

I see, if by flow limiting you mean using splitters to mix the belts into one, yes I do

I don't use any combinator here

[foggy junco]

more specifically, using spliiters to limit the maximum imput of each material, to ensure that e.g. if you have no circuits but you're feeding the system gears or engines or whatever that it doesn't fill up, leaving no room for circuits

memory cell sushi is always a pain

[hardy fable]

I guess I'm using neither then

[foggy junco]

won't that one get clogged with one material if an input isn't saturated?

e.g. if you're running out of steel

[hardy fable]

absolutely

[foggy junco]

[hardy fable]

unless I add a wire to stop the output in case that happens

[foggy junco]

you can do the entire thing without circuits

with flow limiting only

as in, controlling the input rates

[hardy fable]

which is what I do in freeplay

[foggy junco]

i can send some example one second

[hardy fable]

yeah I know how to do that with 1:3 ratios

[foggy junco]

circuits

you can do any power of 2

[hardy fable]

using t1 on t3

[foggy junco]

with balancers you can pretty much do any ratio you want

[hardy fable]

yes I guess so

[foggy junco]

ok now

the real question is

apparently this one does 1/3rd

now how does it work

was just messing around

belt maths make my head hurt

[hardy fable]

it's 1/4 not 1/3

[foggy junco]

it's actually a perfect 1/3rd

you can let it run at any gamespeed and it'll perfectly saturate the yellow belt

[hardy fable]

I didn't got a full yellow when I did it myself

oh wait nvm

it did

[foggy junco]

perfect yellow belt

ok now i'm just messing around

[plain hare]

Splitter Flow limits are fun https://cdn.discordapp.com/attachments/753031298368536577/895935037835456512/unknown.png

[hardy fable]

I think I got it

if I'm right this is 1/5

[foggy junco]

lemme see

seems to check out, 1/5th

[hardy fable]

👌

[plain hare]

Flow limited sushi is the best for its breaking resistant properties

[foggy junco]

now this seems like some really wonky ratio

actually nevermind now it stabilised at a full half belt

[hardy fable]

now the real question is, can you get an irrational ratio?

[plain hare]

I think you can get basically any ratio you want with circuits

[hardy fable]

without circuits obviously

[foggy junco]

opinion on my beautiful 1:1 balancer?

i'll probably go use this just to trigger fireball

[normal gulch]

trigger? thats art

[hardy fable]

what

[foggy junco]

haha