math-and-meta

[astral zephyr]

common grok w?

[fierce ruin]

grok da goat

[astral zephyr]

frfr

[unique cypress]

2500 divided between 5 assemblers is 500 each, not 625

625 would be divided between 4

this AI is so fucking stupid

It mentions several times that fractional items and cycles aren't possible but then does absolutely fucking nothing with that information

Which I guess is par for the course for spicy autocorrect that cannot think

[vapid gorge]

Why use a program that routinely hallucinates instead of several dedicated tools specifically for this task?

Even skimming through this it’s bad. ‘Satisfactory produces in whole units’?

So ? Almost every plan you make is on a per min basis abd whole numbers don’t matter

[brisk smelt]

jesus fuck has this generation forgotten critical thinking and a calculator

[fierce ruin]

Sorry?

[fierce ruin]

What program hallucinates?

[vapid gorge]

Any ai chat bot.

They regularly make up shit.

This is because they don’t know anything, they can just regurgitate things that look right

Like the screen caps also prove it doesn’t ‘know’ how things in this game is planned out

[fierce ruin]

ok, I just wanted the calculations

I tried in game but it was incorrect, so came here

[dusky dust]

If you want a factory solver, there are many out there. My personal go-to is https://www.satisfactorytools.com/1.0/production

Relying on LLMs for calculations is even worse than attempting to rely on them for other facts

[vapid gorge]

But it didn’t even give you a plan?

You want per minute set ups.

[fierce ruin]

nope

I wanted a calclulation to what I asked

[dusky dust]

To be fair, I expect they're aiming to do the thing where you just produce exactly as many SE parts as is needed to complete deliveries, and no more

[fierce ruin]

so never use grok for anything, got it

[dusky dust]

But relying on LLMs for calculations is still an F-tier idea

[dusky dust]

There's plenty of room for LLMs in the space that LLMs are good at, but relying on any LLM for facts is a recipe for disaster

[vapid gorge]

Eh cram stuff on a box and let it run, or overflow items

Or even just use a calculator

[fierce ruin]

I think hating on people for asking questions is F tier

[dusky dust]

LLMs don't know anything. If you try to use an LLM for facts of any sort, you're gonna have to spend longer fact-checking its output than it'd take to just do it right the first time

[pseudo smelt]

I have started to do math in this game it has started

[dusky dust]

Apologies, honestly wasn't meant as a personal attack, but I see how it could easily be interpreted that way

I've got strong opinions about LLMs and sometimes that might bleed through; honestly not intended that way, so I'm sorry 'bout that. :)

[vapid gorge]

They’re trying to put you on the defensive for pointing out facts

Just deflection

[fierce ruin]

all good, wish you guys kept that same energy when dude was making domestic violence jokes in general

[dusky dust]

Nah, it's a fair criticism; I should be attacking the LLM itself, not the decisions people make to use them. Too personal when you're criticising intent. :)

[vapid gorge]

There was no personal attack in any of your comments

[fierce ruin]

I needed to figure out how many, exact, steel beams and modular frames I needed, not per min

[dusky dust]

Didn't see that myself; I'd've absolutely have alerted a mod if I had

[vapid gorge]

You pointing out the problems with LLMs and then them feeling negative? Thats on them

[fierce ruin]

this on me too?

I prompted grok incorrectly, it didn;'t make a mistake

[dusky dust]

Eh, I'm gonna stop talking about it after this, but I do believe that "relying on LLMs for calculations is still an F-tier idea" (which I said) is about half a step from a personal attack (ie: that the person doing so is stupid for doing so), and could easily be taken that way. Tone is important, emotion-via-text-chat is hard, and I could've phrased that differently. I'm not bothered by the callout, and IMO it's not a deflection. :)

(but still: don't rely on LLMs for calculations. )

[fierce ruin]

this is where I got 4 assemblers from
so 4 would need to produce 625 each

[mossy ibex]

well, I got aluminum running, but it's spending a fair bit of time idling because the sloppy alumina is too full going into the elecrode scrap, which is because the scrap is too full of water. Maybe a water buffer would help smooth it out?

[dense apex]

so I decided to break down the fundamentals from this reddit post:
https://reddit.com/r/SatisfactoryGame/comments/1mwji7e/beware_of_the_vertical_junctions/
and found that you don't even need the weird swirly pipe setup he does for infinite headlift. I provided a clear screenshot here: #screenshots message

This was such a fascinating find from that reddit user that I'm still struggling to fully believe it despite my own recreation. It's such an odd bug you kinda have to see it to believe it.

@turbid otter since you reacted to the image, in case you wanted more info its here ^

yeah i think this made me understand VIPs more since, before this discovery on its own, I just thought VIPs were black magic

afaik no dev at css dares touch pipe code because they have no idea how it works /s

it's interesting that i've been playing this game since update 3 and only now learned of the seam difference but it does explain some build inconsistencies in the past

theyre good people

[dense apex]

a what

not currently but I'll ping you with a screenshot tomorrow

I turned the entire rocky desert into my factory and the entire Western Beaches "Gold Coast" into my power production

[gloomy gyro]

I don’t think I’ve been there before

[vapid gorge]

Well, vips are essentially exploits. Not sure why you’re surprised

[dense apex]

im not, really. I already use trickle tower exploit

[vapid gorge]

It really wouldn’t
You’re directly feeding waste into fresh I take it?

[brisk urchin]

lmao real

[swift walrus]

I've got a wire factory with trains lined up to load wire - wire - iron plates. I am making a insulated cable factory, but the rubber comes in on the first station. Can I get a train to only accept wire from the second car? Or is the only solution to build a second station at the destination?

[hoary oar]

no, if you have wire in 2 cars you can't only unload from one sadly
the easiest soloution would probably be to move iron plates in to slot 1 or 2 and then only use a 2 car train.
then if you need the 3rd slot for something else you can fill the first 2 cars with dummy items

[fallow siren]

just place empty station and then a freight station

you only add freight station on the freight car you want to load/unload

everything else can go to empty station

[swift walrus]

Need the first station for rubber sadly. But I think I have a solution. Second train, use a fluid car for the first station and then pick up the wire.

nope nvm. I thought fluids had their own train car. Back to needing a second station.

[plush glen]

I think it'll work if you go load some fluid into the car

It'll just Have Fluid In It, so regular freight platforms won't touch it

Or I might be misunderstanding, I'm sleepy lol

[swift walrus]

Thats a good idea, set up a dummy fluid loading station and then use those to act like empty platforms but on a per train basis. Going to test it anyways.

[hoary oar]

i guess maybe thinking about when a fluid car gets fully unloaded, and then goes to a station where items would be loaded in that slot?

[swift walrus]

nah, I would just never unload it. fill it with as small of a fluid as possible and just let it sit there.

[hoary oar]

oh, use the fluid car just as a dummy? i guess that works

just have to pre load it of course

[woeful zenith]

That reddit post is really complex, but your screenshot's simplified example is awesome. I just tried it myself in my world, and my jaw dropped, just watching the water go up. This method is absolutely insane. Thank you for making a simple version we can use!

[plush glen]

Could also potentially use filters on what items get loaded/unloaded, but I'm not sure if that would be useful here

[amber igloo]

is this the recommended load balancing site? https://icemoonmagic.github.io/Satisfactory-Splitter-Calculator/

for some reason when i click calculate it’s adding an extra 240 items / min input that i don’t have

[hoary oar]

it works fine for me, you sure you arent requesting a higher output than you have available on the input?

specifically with this here

[swift walrus]

Yeah filtering things out would need a sink incase it gets backed up. I don't want to put a sink on the unload beacuse that will waste materials.

[amber igloo]

oh yeah i forgot to delete one 😭

[vapid gorge]

It’s UI is terrible and confusing

Just do your own splitting. It’s easy when you clock and plan things

[vapid gorge]

@mint girder is this one pipe that's going up and down again between the floors for each section?

so you've got the pipe branched into multiple sections AND itscoming from below?
yeah you're going to have severe flow issues

[mint girder]

yeah i run 3 blenders into one pipe (600 fuel/min) and then i split that pipe across 12 gens. i just run it under my factory for looks but i use a pump before it goes back up

[vapid gorge]

I thought you said it was 1 pipe to 12 machines? a lot more than 12 there

[mint girder]

yeah theres 16 in each line, i only have 12 connected tho

the first 12 are connected and the last 4 arent

[vapid gorge]

so how many gens total, how many branches?

[wind spade]

generally we don't even do load balancing

manifolds usually work fine, especially in combination with clock speed and machine grouping

[mint girder]

56 gens, wdym branches? i have 3 blenders outputting 600 fuel/min and then that goes into 12 gens, using 600 fuel/min. i have 14 blenders

[vapid gorge]

which are the 12 gens here? cause it looks like they are all connected underneath ?

[mint girder]

the last 4 arent connected to anything yet

[vapid gorge]

so these are 2 groups using 2 completely separate pipes and aren't connected below at all?

[mint girder]

correct

[vapid gorge]

so why is the pipe connected underneath? a sort of loops?

[mint girder]

that dude in the other channel said to connect both sides of the manifold

[vapid gorge]

oh a loop is good but you're also feeding from below

bring up the feed pipe to the surface and feed it in a loop like this

[mint girder]

like this?

[vapid gorge]

under clock a few machines, once you flood that see if it stays stable. If not slap a powered pump here and re flood it

[mint girder]

i still have the pump there

[vapid gorge]

that might be ok

[mint girder]

underclock the gens?

[vapid gorge]

underclock a couple of them so that you flood the system

[mint girder]

🙏

looks like the gens are full/staying full but the pipes arent 100% full, is that fine

[vapid gorge]

depends if they were fuill when you reclocked thigns up

[mint girder]

i just paused production on it

[vapid gorge]

wdym?

[mint girder]

the gens

turned on standby

[vapid gorge]

did yo uwait until every gen, every pipe, and every blender was ful?

[mint girder]

no but ill give it a shot

[vapid gorge]

that's what flooding means.

and it's a good way to stabalise things and then find out if you have an issue.

for example if everything had been flooded you haveing half full pipes would definitely be a problem

cause that would mean you've still got a flow issue, or a math issue

[swift walrus]

Just got the fluid cars working as a train car spacer. Unpacking a bit of fuel at a station where you make a train then set its schedule and it cant load/unload from that car.

[mint girder]

ah ok i see

[swift walrus]

Might be the most useful thing you can do with a fluid car.

[wind spade]

Moving fluids is useful tho?

[swift walrus]

Not really, you a better off just packing them and sending them as normal freight. Way better

[vapid gorge]

you need a second car to bring things back

[wind spade]

No? You waste power to package and unpackage

[swift walrus]

Minimal power to get way more fluids packed per car

[vapid gorge]

and the need to bring back containers

[wind spade]

You use same amount of cars

Because half of them are going back with empty canisters

(Or you are making them and sinking them, at which point you're wasting resources)

[swift walrus]

You just do that on the same car though further down the station

[wind spade]

That reduces throughput drastically

Because you need 4 stations instead of two

And possibly another train

Which means you could've just sent another fluid train instead

[unique cypress]

It's at least 1.5x more efficient to package liquids

Yeah, you need to add 2 extra stations, which adds 1 minute to the RTT, but you get double the capacity

At longer distances, you get more like 1.7-1.8x

Plus, with items you get to use "depart when empty/full", which increases throughput even more

[wind spade]

Hardly "efficient" when you're wasting power

[unique cypress]

Train efficient

I don't care about power, I care about traffic

[wind spade]

Then specify what kind of efficiency you mean in advance 🙂

[unique cypress]

A fluid train will easily contribute more than double because it has to use default settings

[unique cypress]

I did immediately afterwards?

[vapid gorge]

forget power, tons of space, cost in throughput.
you're much better off adding another car to the same train that takes in the empties

[wind spade]

Only after I asked

[unique cypress]

If you use a second set of wagons or a second train, you're back to exactly the same throughput per wagon

[unique cypress]

Bruh I've been talking about nothing but throughput. If I meant power or space efficiency, don't you think I would've talked about those instead

The only way to get any benefits from packaging is to use a second station to load the empties back into the same wagons.

Otherwise you're just wasting space and power only to end up with the exact same throughput as a fluid wagon (exception: gasses)

Are the benefits worth the extra space and power? Up to you. But at least there are benefits you gain when using extra stations

[mint girder]

there was a pretty huge problem 😭 on the inputs for the blender they need water, and i misconnected some pipes and wasnt even making enough fuel in the first place

was trying to use one mk2 pipe to feed 6 blenders, 3 blenders needs 600 water/min

[vapid gorge]

when you fix a final step you'll often uncover previous issues 🙂

[mint girder]

should finally be fixed tho 🙏

[normal latch]

i unlocked plastic and rubber, so i made a plant that had 2 plastic + 1 rubber + 1 coke (and expanded that to 3x so 6 plastic + 3 rubber + 3 coke)

the idea was to have 40 heavy oil output that gets consumed by the coke, which seems to work

the heavy oil filled up too much though i think, so i made 15 coal power generators to consume all the coke output (5 coal power generators per set of refineries with accompanying water pumps)

is there a reason the heavy oil filled all the way up?

i figured that it would have a one to one ratio since 120 heavy oil should get consumed fully by 3 coke refineries

its my first time with refineries though so im not sure if my homebrew setup is missing something important

[sage shale]

Guys! how many Foundrys can i keep running with a pure iron vein running at 0.1 seconds?

MK.2 Miner

hmm ill just try it out ig

[unique cypress]

Both miner and foundry tell you how much ore per minute the produce/consume

[vapid gorge]

are your machines using teh coke starving?

if so you've got a flow / throughput issue

if they are full and running find? likely a math issue

[vapid gorge]

calculate things by output and inputs. All miners and machines can output various amounts especially with clocking 🙂

and very often times recipes won't fit into whole numbers of machines

[mossy ibex]

your math is off. coal gens burn 25 coke pm so you were trying to consume 120 coke pm with only 75 pm of gens

oh actually I see now that I'm wrong

you said 5

5 should eat 120 yeah. Are you short on water?

[sage shale]

coal gens take 15 pm

[normal latch]

its possible that the issue was due to my initial time to setup

meaning, there was lots of rebuilding etc

so i flushed the heavy oil and will test to see if that solves it

also, i realized something from an earlier "issue" i had with a different plant

with the coke, for instance, it comes out all at once in a big burst of 120 coke

[sage shale]

mb you guys are talking about coke

[normal latch]

so when it hits the merger (using mk3 conveyors + merger), it appears to "stall" as in the 120 + 120 lines of coke dont just instantly all go into the merger at once, but instead there is a delay due to merging

is this accurate to say? or am i still not understanding something?

p.s. i will update on if flushing the system now that its fully setup works as the solution

and also, are awesome sinks really 0 energy cost?

[mossy ibex]

sinks cost 30mw I think to run, but consuming things doesn't cost extra

[oblique hollow]

If the output belt is a mk 3 and the input belts are mk 2, the merger shouldn't stall anythinf really

[normal latch]

all the belts are mk3

should i be using mk2 input belts in this case?

i figured to always use my highest tier belt for everything, never considered purposefully using lower tier belts

[oblique hollow]

if you use mk 3 it stalls because it moves that little burst of coke at 270/min into the merger - which can only output 270/min either, so it stops both inputs half the time

After that, the belt stays empty for a while

And thats normal behaviour, its not wrong if this happens

[crimson moat]

Yeah, if you're using a belt that is faster than the supply (for example a 1200 belt on a miner supplying 600) then it won't appear as a smooth flow.

That's just because all of the ore clumps towards the front of the belt though - it spends 50% of the time moving at 1200/min, and 50% moving at 0/min. That starts and stops, but averages the correct 600.

with a slower belt you don't see it as much (or at all, if the belt is equal or slower to input speed) because the ore that is "waiting" at the end of the faster belt has not even arrived at the merger yet.

[unique cypress]

Well, in the case of a miner, the flow will be smooth, just spaced out

A miner produces 1 item per cycle

[crimson moat]

It is not smooth if the belt is faster than the miner. This issue is worst i think with a pair of 600 miners on a 1200 belt, but similarly seen with 2x60 on a 120 belt etc.

the belt appears backed up before the merger, and starts/stops constantly on the supply side (even though it's 600+600 = 1200, and the 1200 belt appears smooth).

But it's not actually clogged, as evidenced further down the belt

If you connect source before drain and the two rates are matched, there will be a slight excess of ore at the merger because of ore which was already put on the belt before you connected the drain.

It has no way to clear that excess because input and output are the same, so it persists indefinitely without intervention.

The input belt should be in this state for smooth flow:

"0101010101010101"

but actually it's in this state since it had nowhere to send the first ore that it got

"11111111101010101"

When you look at the front of this belt you just see the "11111111" i.e. ore in every slot, starting and stopping, and it looks clogged. But it's actually not, because further down the belt it's in the "01010101" state and the buffer of the miner is empty.

The input/output rates are exactly matched, so you get this weird equilibrium where like the front 30% of the belt looks clogged and the back 70% doesn't, but functionally it is not clogged at all. It's just a harmless artifact of feeding a merger with multiple belts that are each faster than their inputs while also adding up exactly to the output belt speed.

[main thicket]

are drone port's transfer rates accurate when you have multiple drones incoming in one port?

[crimson moat]

they're not super accurate in general, i think so though in this case

[main thicket]

yeah thought so

thx

[unique cypress]

they're hella inaccurate even with one drone

[meager shadow]

unlike factory carts with truck stops which give me totally believable numbers like #screenshots message

[normal latch]

thanks for explaining the belt "issue", glad to know that my math was right and its just how the belts operate, and that the operation does in fact make sense and the results are expected (e.g. appearance of "clogged" belt that isn't really clogged)

[random creek]

Tbh my drones so far have read accurately

I needed 480 a minute. Drone said I was getting 240 a minute. I popped another. It also gave 240 and got me the 480 I need

But thats also the only setup where i cared enough to look so

[dense apex]

I'm glad it helped

[amber igloo]

does this work to even out items going through the manifold?

[unique cypress]

what are you even trying to do?

why the loopback?

[amber igloo]

wait what thats true

i was on autopiolot

mb 😭

in addition, if 3 machines after the splitters dont use all the ingots from my smelters, will the splitters automatically even them out to the next set of machines that may require more? therefor if i have these overflows on each section, they will only take what they need?

[unique cypress]

machines won't take more than they can use regardless

[amber igloo]

so thats how overflows work correct?

[topaz hedge]

No? If you use all of what the smelters produce they're will be no overflow

[brisk smelt]

think of it like this. you take x/min and you supply x/min. as long as you don't hit belt limit it doesn't matter how much is supplied to each individual machine because each individual one can only take so much

you just need a belt going to each machine in whatever way you like

same idea with pipes

[visual ocean]

i need some help with My rubber setup. Its the basic turn 300 crude into 900 rubber
I prefilled all the pipes thinking I was going to have issues starting up.

My problem is that I'm not making enough Resin to make the rubber to make the recycled plastic to then make the recycled rubber.
And i cant do that because i have all the pipes full with HOR and fuel.

I have tried to clean the refineries and also to flush the pipes (not the full thing tho)
In last resource I will flush the entire pipe system.

However i want some ideas first on how to get this going

EDIT: this screenshot has belts not doing things. They are all connected and with sinks

[amber igloo]

so yeah if i oversupply i can just feed that to the next factory segmant that uses the same resource

[brisk smelt]

you need to seperate the recycled rubber refineries

one of them feeds the loop and one is the output group

all the numbers are below 720, there shouldn't be throughput issues with that method single belted

except the final

[brisk smelt]

yeah exactly

[amber igloo]

thought so! thanks!

[visual ocean]

The rubber is separated. One is the input for plastic and the other is just factory output

the last rubber is even double belted to account for the 900min

I tried to clear the full pipe system of HOR

it worked, for 5 mins or less

So now Im waiting for everything to be saturated liquid wise to do the same for all the fuel (which will help the HOR to flow longer)

And i assume, with that fluid dump my rubber production wont stop on those 3 refiniries since the HOR will be produced non stop

[unique cypress]

I always set up my recycling like this #math-and-meta message and it always works perfectly

[visual ocean]

That's kinda what I'm doing. I dont do 2 outputs tho

I OC the 10 refineries

[unique cypress]

If you split the refineries, then no, that's not what you're doing

[visual ocean]

The residual rubber is feeding the recycled plastic which then goes to recycled rubber.
I can take from the last rubber output and put it into the plastic but that's more confusing to manage i think

I cant really use that method you tagged because im only doing rubber

[unique cypress]

this works for literally any output ratio

[plucky tusk]

I keep avoiding building my recycled rubber and plastic lol

[plucky tusk]

Still running off my residuals from my first fuel power

[visual ocean]

Okay, I'm dumb. I was having a hard time understanding what do you mean.
I though it the end I was going to be doing less than 900. But i can just do an overflow splitter

Sorry for being stubborn, and thank you for the help

[wheat lake]

any idea why my aluminium setup keeps screwing up the water pumps and refinery''s are all isolated from each other
Water Extractor 80m3/m
Sloppy Alumina consumes 200 Water/m
Aluminum Scrap produces 120 Water/m

but they're still backing up with water every minute or so after purging the pipes

[unique cypress]

because the sloppy alumina doesn't consume 200/min unless it's all running at 100%

and if it doesn't, you fill up on water instantly

[agile junco]

Pretty happy with the capacity now. I just need to lower my max, and fully utilize things better.

[wheat lake]

100% effecientcy or overclock?

[unique cypress]

uptime

[wheat lake]

it should have 100% uptime im not running out of resources and i have the outputs hooked up to sinks

[magic island]

• Extractor-water fills the pipes on initial startup
• Byproduct-water cannot enter the pipes because they're already full of extractor-water
• Scrap refineries stall because their water has nowhere to go

You need to send the byproduct-water somewhere separate. Options include:

• A separate set of Alumina Refineries that are not hooked up to fresh extracted water
• An entirely different production recipe
• Coal generators using leftover coal from your scrap lines

[unique cypress]

or use a VIP to prioritise byproduct water so it never fills up

[wheat lake]

whats a VIP?

[unique cypress]

variable input priority junction

it's in the plumbing manual in the pins here

[wheat lake]

oh cool

[magic island]

it's a pipe contraption that allows you to combine fluid from multiple sources while prioritizing one. if set up right, you can let byproduct water flow freely without being blocked by fresh water

I've never been one to suggest it because I find it a bit janky, but others swear by it, and it does work if built right

[unique cypress]

afaik, the only way to break it is to build it wrong

it has never failed me, unless I failed to build it correctly

[oblique hollow]

valves on the byproduct line deprioritize it.
remove the valve from the byproduct line

[mossy ibex]

imagine a world. Pipes are simple and work as expected. There is a vanilla fluid priority junction.

[oblique hollow]

also, this design is just plain unstable unless you make sure it always has 100% uptime

[wind spade]

That's "pipe is just covered belt" world, so obviously not reasonable

[mossy ibex]

I wouldn't be mad about that honestly, but I get you

[wind spade]

at that point we can go back to having no pipes

[fierce ruin]

is turbo fuel worth it over normal fuel?

[unique cypress]

not really

saves oil and that's about it

it needs coal and sulfur (fuel doesn't)

and both take about the same amount of machines to make

[fierce ruin]

thought so

[deft lichen]

you can just make more diluted fuel rather than going through the effort to import sulfur for turbofuel

[fierce ruin]

might just use oil until next upgrade

[deft lichen]

rocket fuel is a different story though

[fierce ruin]

hmm

is there a limit on how many dimension depots you can use? I have 17 atm

[solemn pivot]

only limited by spheres

[outer vale]

and there are enough spheres to not be a practical issue

[mossy ibex]

ok weird question -- using infinite nudge you can set the size of the scaling operation in the options, so like a large scale can be set up to 500%, medium to 100%. If I, eg, do the medium scaling, that would double the size of my part, but then if I do the operation again does it double the currently doubled size, but does it multiply by the original size, which would make it 3x overall?

I think it may be the latter, but it's been confusing me

ah yeah, I just answered by own question: it is in fact the latter. So scaling isn't like size * 2 * 2 it's like size * 2 + size * 2

[versed violet]

WTF is going on with Compacted steel ingot? Am I seeing correctly its 5 iron + 2.5 compacted coal = 10 steel ingot? Does this recipe have a comma in wrong spot?

[magic island]

Add an extra step and a third resource, double your Steel yield. That's not out of step with other alts

Solid Steel gives you 1.5x the default yield just by turning the iron into ingots first.

[fierce ruin]

it must have been a lift issue, fixed it now 🙂

[solemn pivot]

lift?

[versed violet]

yeah, but this is basically trading 5 sulfur for 5 coal, with extra preprocessing step and where sulfur is much more valuable than coal. plus heavy machine cost. Other recipes give from 45 to 100, so 10 really makes it look like somebody misplaced the comma. 50 iron + 25 compacted coal = 100 steel ingots sounds reasonable. one tenth of that, not really.

[unique cypress]

this is what compacted steel ingot was before 1.0

[versed violet]

same proportions, much more reasonable per/min output

[unique cypress]

it's slightly cheaper now, but slow af

[magic island]

oh, you just mean the overall recipe rate

yeah the resource balance is fine but it's a slow one

[unique cypress]

no clue why they gave it such a slow-ass craft time

even 10-12s would've been slow

even more proof that 1.0 wasn't balanced properly kekw

[versed violet]

which makes it kinda weird. in later tiers, you have better things to do with sulfur and space. in low tiers, you are not going to have enough power to use this.

[plucky tusk]

Nah but nitro rocket make compacted as a residual you could turn that into steel

[fierce ruin]

Conveyor lift

[unique tree]

I connected a group of generators and two mk2 pumps to a seperate grid. I cant for the life of me figure out why consumption is 116MW? and how is it 100MW more than the max consumption?

[deft lichen]

take that thing off your back

and please don't post in multiple channels

[unique tree]

sorry wont do that again, thx for the help tho!

[normal latch]

what was the solution? im guessing they had that hoverpack thing that ive seen in videos (havent got that far yet myself) and that was drawing power?

[agile junco]

It bothers me that the hoverpack isn't added to the max cons.

[brisk smelt]

its so insignificant that might as well not be considered

anyhow, hoverpack is not part of a power grid

[agile junco]

I notice it when I'm dealing with emergency battery installs.

[agile junco]

It's still wrong to have your consumption able to be more than the "max"

[brisk smelt]

there are a lot of things in life that can go more than max

[agile junco]

Perhaps we have different notions of what max means.

Also, in this context, max doesn't mean safe limit like it often does, e.g. max weight for elevator. That's the Capacity. So it's pretty unambiguously the value that you shouldn't ever be able to draw more than... (except apparently hoverpacks).

[fierce ruin]

How do you figure out what's the best way to make a specific item when you have multiple alt recipies?

[unique cypress]

SFTools

[fierce ruin]

whatever it comes up with first?

[unique cypress]

You don't have alts enabled...

Also, why so many?

[fierce ruin]

oh i dont need that many

[unique cypress]

Whatever graph it generates with everything enabled is usually decent (unless it's using SAM for no reason), but I usually tweak it

Like I'm not gonna bring in 30 coal to save 100 iron when I'm already spending 1000 so I just disable coal and use 1100 iron instead

Same with oil. I'm not gonna bother bringing in oil specifically for coated or adhered plates, unless I already need oil for something else

There are also things I don't like using so I always disable them (default modular frame, default alumina, screws)

Sometimes I'll also tweak the plan to better match the resources I can find or I'm already using

[fierce ruin]

Unfortunately we couldn't calculate any result.
This can be due to many things: missing resource required for the production line, not enough resources for the requested amount, disabled recipes required for the product, etc.

[unique cypress]

Like if I'm making Crystal Oscillators and High-Speed Connectors in one factory, I'll often use the alt for the Connectors because it uses quartz (that I need anyway) and it saves caterium

[unique cypress]

The hell did you do that it gives you that error when making HMFs?

[fierce ruin]

the alt recipie

but its not working at all

[unique cypress]

Send the link

(Share button in the top right)

[fierce ruin]

ok i unticked/tciked/ it says encased frame

https://www.satisfactorytools.com/1.0-ficsmas/production?share=6uE2JytzQCwOsqb8RgXF

way too many alt recipies

tf

[agile junco]

yea

[unique cypress]

This is what I'd normally do for HMFs
https://www.satisfactorytools.com/1.0-ficsmas/production?share=MBhkbjfJa4hQtVUsJeeJ

Also, that isn't really a lot of recipes lol

[fierce ruin]

do I enter all my current alt recipes or just ones that might be needed

[unique cypress]

You can enable all recipes in tools, because it optimizes for lowest resource usage

It won't use the ones that are more expensive

[fierce ruin]

I don;t have too much

[unique cypress]

It's not always exactly what you want, but it's a good start for something better (I did disable a few things that made it easier but "more expensive", after all)

[fierce ruin]

Yeah seems overkill, but I do need a heavy frame factory and then a computer factory

[unique cypress]

And it's not like you have to have all the enabled recipes, you only need the ones you end up using

Which is why I think enabling only the ones you currently have is not a good idea

There might be one recipe you don't have yet that makes a huge difference

[fierce ruin]

yep I just need iron and limestone, the nodes are close toghether 🙂

thanks for the tips!
ignore the numbers, but this is the set up

[brisk smelt]

input/output station should both be buffered if all belts are running at max cap right, i forgor 😭

[unique cypress]

You won't get 2 full belts per platform, even with buffers

But yes, buffers need to be on both loading and unloading sides

The actual max throughput per platform is this

s is stack size (50, 100, 200 or 500), b is belt speed (in items/min) and tp is pause time in minutes (0.452)

[brisk smelt]

ah yeah i mean 1200/platform, whatever i just put them on lmao

[normal latch]

is there a tier list (within reason) for alt recipes?

[unique cypress]

Technically yes. But I wouldn't suggest using one

[normal latch]

i guess i will just use my own mind to think of which alts to use for my first playthrough

[unique cypress]

Use SFTools to compare recipe chains to decide which one you want to go with

[normal latch]

maybe, i've been enjoying trying to solve it myself

my next two projects will be 1) make all space parts for unlocking the next phase 2) making my first oil power

oh and 3) making my first simple train to port rubber/plastic from oil location to main base

and 4) rebuild stuff and use alt recipes finally (i have so many unpicked harddrives lol like over 10)

the list goes on, this game is great

[unique cypress]

The most efficient oil alts are kinda difficult to figure out without an automatic calculator

[granite solstice]

okay so what do you do when you're producing more than your belts can handle?

[unique cypress]

A belt balancer

[agile junco]

Recycled plastic/rubber is my current favorite, just because they feed each other and it's fun to build.

[agile junco]

Do you have designs for optimal balancers?

[granite solstice]

what is that?

[unique cypress]

Or at least something that functions like one

[unique cypress]

A splitter/merger setup that connects every belt to every other belt

Allows to kinda bypass belt speed limits

So if you have like 120, 200, 240 inputs and want 160, 180, 220 outputs, you just slap a 3:3 balancer on there and it'll distribute the items correctly

No matter what the inputs and outputs actually are

[unique cypress]

I know the theory of how to design any balancer

But optimizing them for space is another thing entirely

[agile junco]

Yea talking about space

[granite solstice]

okay got it i'm off to learn how to build and implement balancers

[unique cypress]

But instead of actual balancers, you can use something like this: #math-and-meta message

[agile junco]

I'm always interested in minimal representations of general constructs like balancers and arrays.

[unique cypress]

It won't work for loading/unloading trains because it's not input/output balanced, but every input is connected to every output and with a full belt's capacity

So it'll distribute items in a factory perfectly fine

Just more unevenly at first

But as long as none of the belts is connected to something that can consume more items than planned, it'll work

(and, tbf, a balancer has the exact same limitation)

[agile junco]

What do you mean if it can consumer more than planned? It only works at saturation?

[vapid gorge]

well the way to minimise the space use to absolute minimum is to clock groups of machines so you don't need to LB them.

zero space used that way

[unique cypress]

A balancer by default outputs equal amounts of items from each output. If you connect it to machines that consume different amounts of items, the ones consuming less than average will have to back up before the ones consuming more get enough items

[agile junco]

Might be better to OC fully and load balance in some cases I would think. Plus I'm just talking about the balancer, perhaps I want a tight representation of that in a space not otherwise optimized.

[vapid gorge]

define 'better' ? LBing doesn't provide any mechanical benefits

[unique cypress]

1. we're not talking about load balancing machines
2. underclocking to match ratios often requires additional machines. Belt balancing never does

[agile junco]

Ah yea I've done that. I wouldn't have called it a "balancer" it's just a splitting arrangement.

[vapid gorge]

a belt balancer then?

[unique cypress]

What do you think a balancer is then?

Because that's exactly how balancers in Factorio work

Here, they're the exact same thing, just built out of different parts

[agile junco]

I feel like I'm misunderstanding you. A balancer to me is something which outputs fixed ratios of the input. So in you have a machine that needs 3/5 it gets it right away, no need worry what the other machines are doing.

Yes if other machines under consume, then you might get more.

[unique cypress]

Yeah, that is a ratio splitter, not a balancer

And it doesn't help with dealing with multiple belts

[vapid gorge]

is your goal to feed machines evenly, or to just have a number of belts with various parts per minute, recombined to have different parts per min?

[agile junco]

I see

[unique cypress]

Balancers input and output equally, but can use over/underflow to distribute items in any ratio

A manifold does that just fine, but only up to 1 belt (and it doesn't do equal ratios even when unloaded)

Which is how that nightmare splitter/merger wall works

It's just a cross of manifolds

One splitting vertically, one merging horizontally

Ultimately you get a manifold for manifolds

[agile junco]

Makes sense.

[unique cypress]

Inside a factory, works just as well as a balancer, because you don't need the equal ratios of an actual balancer, and the most important parts are the over/underflow between all belts

[agile junco]

I've always calculated what my consumption needs and used ratios to achieve it.

[unique cypress]

For loading and unloading trains, you specifically need the equality, so an actual balancer is required

[agile junco]

What's different about trains? I don't fully understand this yet.

[unique cypress]

I usually want to put in minimal effort into boring things (primary school level maths), and instead spend time on more interesting things (designing balancers)

[unique cypress]

Well, first of all, you don't need this to make a train work and transport items

[agile junco]

I assume it takes up more space though. No?

[agile junco]

Yea I get that.

[unique cypress]

But because trains are kinda stupid, default settings do not allow you to get the absolute maximum throughput and efficiency

To get that, you need "depart when empty/full", which forces the train to only move full wagons at a time, never partially full ones

But because it won't leave unless all wagons are empty/full, if one one of them gets loaded/unloaded slower than the others, it'll delay the entire train, dropping the throughput of every other wagon to match itself, the slowest wagon

The only way to avoid it is to have all wagons be the slowest, i.e. equal

[agile junco]

Ooohhh you mean balancing across multiple docks.

Gotcha

[unique cypress]

So you have to load and unload all wagons perfectly equally to get max throughout

[unique cypress]

I mean I only need 1 balancer between each step of production. And the balancer matches the number of belts I have.
Relatively small production - 2 or 3 belts of items - small balancer
Bigger production - more belts - bigger balancer

The production still takes up the majority of the footprint

[agile junco]

I'm still not understanding something... if you have 600/min going in and it all goes to dock 1 before dock 2, then who cares it still needs to wait for dock 2 and the same max rate is being sent through, no?

As long as the belt never backs up, you're getting full throughput

[unique cypress]

Especially because thanks to balanced belts, I don't need to worry about how many items need to go where exactly. All machines making one item go in one big compact block

[unique cypress]

If you're loading 1 belt total or less, that's correct

You can manifold platforms and it'll work fine, even in the efficient setting

[agile junco]

Well two for each dock

Since they have two inputs.

[unique cypress]

But we're talking about belt balancers, which handle multiple belts

[agile junco]

Lol I just don't get what that means

[unique cypress]

How would you ensure all belts are constantly moving when loading 4 of them onto a single train

[agile junco]

I see.

[unique cypress]

The belts aren't necessarily equal, nor do you necessarily have 4 wagons on your train

[agile junco]

Splitters into mergers. Aka a balancer

[unique cypress]

With one belt total, it will not stop moving until all the wagons are full. So you don't need to balance anything

Even with 2, you can probably connect one of them to the top inputs, one to the bottom inputs and I think it'll still work fine

[agile junco]

Of course

[unique cypress]

But once you hit 3 belts, you cannot load all 3 at max speed into one platform if that's the only one that's still half empty because the other ones filled first

[agile junco]

I mean, three belts each at 1200 simply cannot go into two inputs faster than 2400, so I don't see the issue with just a single merger.

Unless you need each belt to be evenly utilized, which is what I thought this was about.

But that's why I don't understand what this has to do with loading a train.

Watching https://m.youtube.com/watch?v=4CHS5zDBZIA

[unique cypress]

That's the point. They can't.

If you have 3 belts to load onto a train, and you want maximum efficiency, you need to make sure all 3 belts are moving at all times. If they stopped, you're losing throughput.

If you just load them willy nilly, at the beginning, everything will be fine, because all platforms are empty. But if you aren't loading them evenly, you might end up in a situation where most of the platforms are already full, but the last one still isn't.

At that point, you have all except one platform full, no longer accepting items, and one platform still accepting items at full speed.

But that full speed is 2 belts. And you have 3. So now the belts are not moving at full speed and you're losing throughput

The point of a balancer is to prevent this situation from happening in the first place. It ensures all platforms get the exact same amount of items, always transferring items at max speed, and all become full at the same time. Meaning one platform being slow cannot delay the train, because all platforms are full and ready for departure at the same time

[agile junco]

Maybe I need to play with it in game because just don't see what this does that a set of mergers alone don't do.

[plucky tusk]

how much plutonium waste at 100% clock speed

[agile junco]

Or maybe I'm just building a balancer at that point, right?

[unique cypress]

Depends on the item rates on the input belts. Like if you have 3 600/min belts, all coming from miners, so they're consistent and equal, and you're loading them into 3 or 6 or 9 or ... wagons, you don't need a balancer.

Your goal is to ensure that all platforms are loaded equally. If the belts are already equal, then you don't need to equalise them. Just plug em straight in. Maybe splitting each one in half or in thirds if you have more wagons than belts.

But even if you have 3 equal belts and 4 wagons, you can still get away without a balancer. You can split off 1/4 of each belt, and merge those, leaving you with 4 belts, each carrying 3/4 of what the original ones did

But a balancer guarantees the belts will be equal. You can plug in 678.954 into one input, 127.369 into another and 300 into the 3rd. In the end you will have equal outputs.

Same if one gets disconnected or whatever happens to it and it stops delivering items as expected. Your train will be loaded slower, but still equally.

Oh, and also, the same applies to the unloading side as well. For max throughput, you need to ensure all wagons are unloaded equally. The same logic applies there as well. Except the train will by default provide equal belts, and you need to ensure that none of them backs up before the others

If one backs up, all need to back up exactly the same

And again, a properly built balancer ensures that, but again, is not strictly required

Anyway, I gotta go, so if you have any more questions, probably DM them to me instead of posting them here. I might miss them otherwise

[echo vale]

here's the outside

here's my water pipes and extractors. despite there being health flow, the water is not reaching my refinaries. also have not utilized the large lake in the distance

[fallow siren]

those pumps are very high up, you sure the pipe below has enough head lift to reach those?

and also do you providing enough water?

[echo vale]

I probably need to use most of the lake

[echo vale]

each of these oil extractors goes to 3 refinaries

[oblique hollow]

Does none of the water reach the refineries or do only SOME refineries not have enough

[echo vale]

none

[fallow siren]

did you connect the pipe to wrong input?

some ppl did confused refinery's input/output

[oblique hollow]

If all the pipes up to the refineries are full the something was not connected right with the refineries themselves.

If part of the pipe network is empty, then there is a connection issue with the pipes

[severe girder]

My maths i did in my head last night at 2am, please help me verify. Im moving 50,750 items a minute. 1 cart - 3.200 items and 4 carts - 1 loco. I can fill the 3,200 in 81 seconds or a minute and 21 seconds. But we need to load and drop of so this takes 2 mins 42 seconds. To take all 50,750 items this is 4 locos and 16 carts. If the rain moves at an average speed of 70/km a hour it moves the distance of 2km in 104 seconds. This makes a time of 4.5 minutes. 50,750 and a loop time of 4.5 minutes means my train needs to carry 228,375 items on a train with 72 carts and 18 locos, or I can half this ect if I want more than 1 train

[proper dragon]

Anyone got ideas why this interesction isn't working? I know it's not clean but I have path signs at every entrance and blocks at every exit

The error is 'Signal Loops into itself'

[vapid gorge]

Intersections hate not being flat

[proper dragon]

I NOTICED
i got it working for now, next time I'll level things out
Thanks!

[fallow siren]

if intersection, always make it flat

if you want to ramp them, at that point just go for 3d junctions

[unique cypress]

1. to move 50750 items/min of stack size 100, you need at least 29 platforms to even load them all onto one train fast enough. and that's with mk6 belts
2. 1 train with 29 wagons, though, will only be able to deliver that up to 1.8 minutes round trip time
3. to know the actual round trip time, you need to actually build the route and drive it and measure the time it takes. remember that each load/unload operation takes 30 seconds, so if you don't perform (un)loads when you time, you'll need to add half a minute to the time for each station
4. adding a second train to the route will double the time the throughput can be sustained to 3.6 minutes, third will triple to 5.4, etc
5. however, that is all assuming the train is set to depart only when empty/full and all wagons are loaded and unloaded perfectly equally, which often requires a balancer

[crimson moat]

A set of splitters/mergers which take input belts which may not be identical, and create a set of identical outputs

[unique cypress]

As I often say in the Factorio Reddit: just because outputs are equal doesn't mean it's a balancer

[crimson moat]

that depends entirely on a subjective definition of "balance" then

and at that point, why argue

[unique cypress]

well, I'm going by the factorio definition of balancer

[crimson moat]

does factorio have a definition ingame?

[unique cypress]

because they exist here and work the same, I don't think they should have a different definition here

[crimson moat]

i have seen or heard of no such definition

[unique cypress]

It's probably mostly created by raynquist

[crimson moat]

so no, there isn't one and it's entirely subjective

[unique cypress]

all definitons are 🤷‍♂️

[crimson moat]

Nah, games have lots of them in an official capacity

that just isn't one of them

[unique cypress]

and?

[crimson moat]

I would leave it there, but you are telling me to change my usage of the word based on a subjective opinion of somebody who plays a different game?

[unique cypress]

there are plenty of community created definitons

pretty sure nobody argues what a bus is

[crimson moat]

bus has a computer science definition, usage in factory games is piggybacking on that.

[unique cypress]

so does a balancer

and the factorio balancer and an irl balancer work similarly

so it's entirely possible that it's based on that as well

there's also battery balancing that does a similar thing...

[crimson moat]

there is load balancing in parallel computing, but the extra word changes the definition and usage of the term quite massively

to balance something, in english, is literally to make it equal

but load balancing in computer science is kind of the opposite, it refers to distributing load unequally in a way that allows it to be served in equal time: e.g. it can mean e.g. outputting 66/33 so that the 66 goes to the component which is twice as capable, or has twice as much empty schedule time

[unique cypress]

yeah, it makes the demand on each one equal

balancers in this game and in Factorio also make belts equal

but just because you've made belts equal doesn't mean you made a balancer

because merging all belts into one and then splitting them equally doesn't make a balancer, even though all belts are equal

[crimson moat]

I would say:

Balancer, yes

Load balancer, no

[main thicket]

anyone know how much mw would a fully overclocked particle accelerator consume at the peak? (making plutonium pellet)

[crimson moat]

they are both useful terms and not the same thing

[unique cypress]

as far as I'm concerned, there's no such thing as a load balancer. there's load balancing

[crimson moat]

there is passive load balancing but not active (excluding overflow)

[unique cypress]

you can load balance with an acutal balancer, but you can also use a ratio splitter, or a rate splitter, or a belt limiter

I wouldn't call each of them a "load balancer"

you're doing load balancing with them

[crimson moat]

don't know any resource which has the values for each recipe

however the peak power does not actually matter if you have power storage, because storage discharge rate is uncapped and peaks aren't sustained for long enough to drain storage.

What matters is:

1: That you have enough generation to cover the average (otherwise there'll be a net drain over time)

2: That you have enough power storage to absorb the unused power during dips where consumption goes below average. That's 10 power storages per gigawatt that you go below your average draw.

Fulfilling these criteria guarantees that you'll have enough energy in the bank to cover peaks, because they're equal and opposite to the dips on average.

The capacity of power storage is also much larger than neccesary for this job; they're likely to fluctuate between around 98-100% and shouldn't drain deeply unless you have a problem with above criteria. We have to build way more energy storage capacity than we need because the limited charge rate (100mw each) is a much more severe limitation than the energy capacity.

[unique cypress]

2518.3 MW

[main thicket]

thank you

got it

i think

[severe girder]

Te will all be unloaded equally, they all start to unload and load at the same time. ( I think that’s how the game works them, lmk otherwise ) and yes I would need another train, this is why I said 72 train carts for a train or half if I have 2 trains ect.

So the only thing I didnt do right assuming my round trip time is right would be i didnt consider the unload and load animation

[unique cypress]

When aiming for the absolute maximum throughput, you need to ensure all wagons finish loading/unloading at the same time and then the train immediately departs

[severe girder]

Isnt that how they would normally work?

They all load and unload at the same time?

[unique cypress]

if you load them at different speeds, no

[severe girder]

They will all have the same conveyers for unload and load

Mk6 ( 1200 )

[unique cypress]

a platform connected to a 600/min belt will fill half as fast as a platform connected to a 1200/min belt

[severe girder]

Yea, sorry I didn’t know that’s what you mean. My bad

[unique cypress]

and it's not about the mk of the belt connected, it's about how many items it'll actually carry

[severe girder]

They should all carry the same amount

Or at least all of them but maybe 1 end one

[unique cypress]

well, if you use "depart when empty/full" and one wagon is loaded slower than the others, all will be loaded at the speed of the slowest one (well, 2-3% faster than that)

[severe girder]

Can you not say depart when empty/full on 1 of the full carts, and since the end 1 had less in it will be full by the time the others are, I say full bc the end 1 isnt going Yk carry as much, just under half full

[unique cypress]

no, that setting is for the entire train. all wagons need to be either completely full or completely empty, or not docked at all for the train to leave

[severe girder]

ahh alright

[unique cypress]

(though for the purposes of "full", all slots filled with something count, so the last wagon to fill will have 31 stacks and a few items in the 32nd slot, vs the others having 32 completely full slots, so that's where the 2-3% difference comes from)

[severe girder]

I think I understand

So I should only for full efficiency let a full train go, and keep the extra maybe 15 stacks and transport that with another method

[unique cypress]

you could also use a balancer kekw

[agile junco]

@unique cypress I had to actually build one before I really understood it. I see why these are important in some cases for optimizing train loading. Luckily I wont always need it because I can split extration evenly most of the time.

[unique cypress]

thugh even the minimal size one (n:29 and 29:n) would be massive

[severe girder]

Huh?

[unique cypress]

to guarantee perfectly equal loading of all wagons, you can use a balancer

[agile junco]

The space grows quadratically in N+M if I understand correctly.

[unique cypress]

no, the original message was correct lol

it is n*m

it might even be (n*m)^2 or ^1.5 or something

[severe girder]

Can you send me a video or wiki or something about a loader I would be using.

[crimson moat]

It would probably be easier to scale up a bit to like 32 carts than to balance to 29

or use smaller trains, like 8 carts per train

[unique cypress]

32 sux cause it's just a power of 2. that's a lot of stages. even 36 would be better

[unique cypress]

https://satisfactory.wiki.gg/wiki/Balancer#Balancer_examples

[unique cypress]

and frankly, 29 isn't even that bad. it's one less than 30, and 30 is 6x5

depends how many inputs you have, too

[agile junco]

It's actually not a nice graph, since the ratio between N and M kinda matters, no?

[unique cypress]

in the general case, you're splitting into n*m belts

but the splits cannot be done in one stage, so it grows even faster than n*m in volume

the height and width are proportional to n and m (one to n, one to m, doesn't matter)

but the length scales with n and m too

so it's more than n*m

[agile junco]

Yea that's what I mean. And the number of those intermediate stages is somewhat dependent on the ratio.

[unique cypress]

in the general case, no it doesn't depend on the ratio

[agile junco]

Or maybe not actually.

[unique cypress]

except this is the general case 4:4

[agile junco]

I'd kinda like to see a formula for S, M = f(N*M)

[unique cypress]

and this is the equivalent compacted version

the number of splitters and mergers is the same, but the layout is much easier to build

[agile junco]

I just built a 3:4 and it is S = 9, M = 4.

sorry, 9

[unique cypress]

n*m is 12, S+M is 13. close enough lol

[agile junco]

I mean, I only need one layer of merges for this, so that's not surprising.

[unique cypress]

but you need n*m belts between the splitter and merger stages

that's what always scales exactly with n*m

idk how splitters and mergers scale tbh

[agile junco]

My guess would be log(n*m)

[unique cypress]

well, no. the 4:4 has 24 splitters + mergers

that's more than n*m

[agile junco]

sorry, I'm thinking of the number of stages of splitters and mergers.

[unique cypress]

ah, yeah, that'd probably be ln(n*m)

or maybe not base e, but it has to be between 2 and 3

[severe girder]

And what does this do?

Why would you put in 4 and take out 4? Couldn’t you make it straight?

[unique cypress]

because they're cross connected

items from 1 input go into all 4 outputs

[agile junco]

Belt balancing allows uneqal rates on the inputs belts to give equal rates on the output belts.

[unique cypress]

or unequal consumption on the outputs to turn into equal consumption on the inputs

[agile junco]

you could have belts of 10/min+20/min+30/min+40/min and end up with 4x25/min.

Fun (maybe obvious) fact, unless I'm wrong somehow. A 4:3 balancer is a 3:4 balancer where all the splitter and mergers are replaced by the other kind and the belts are all reversed.

IDK why last night when I was trying to see how this helped train loading I was having so much trouble. I must have been tired.

[amber umbra]

I do like using balancers for train stations usually as a way to increase max round trip times. Used 1:4 and 4:1 balancers a lot.

But come from Factorio trains where it’s quite common practice.

[severe girder]

Ahh Alr

[agile junco]

A 1:4 is just 1 extra splitter or merger over a manifold, so I can see that.

[agile junco]

I'm going to go insane trying to find compact balancers now using belt lifts.

[severe girder]

@unique cypress i dont think modler has trains

Bc I think it was you who told me to use it, but it does have trains

[agile junco]

I want this to be a thing... but I really don't think it's possible

[knotty hornet]

Man, when I was looking at balancers, I did not see them getting this big.

[agile junco]

This is a really cool tool. I hadn't seen it before.

Would be super cool if it did trains at least the math for RtT and TtF.

Modeling the pathing and contention with other trains would be insane.

[agile junco]

Interesting I broke the app

[unique tree]

i managed to gobble up this nuclear setup trying to make it as simple as possible. still looks like a baby threw up crayons on a piece of paper. is there really no way to make wasteless nuclear more simply?

[agile junco]

Ficsonium is beyond me at the moment. But I assumed I'd have a smaller power plant which was sustainable enough to power key subfactories which produce some of these parts separately.

[unique cypress]

this is what I would consider the simplest and cheapest wasteless nuclear: https://www.satisfactorytools.com/1.0/production?share=FWahcaKx1Xn2tsx96dEx
scale up as needed (remember to scale up the waste input as well since SFTools doesn't handle it automatically)

[unique cypress]

ficsonium is unfortunately not worth making unless you specifically want to make it and don't care about the resource and time costs

[agile junco]

In terms of power output to material prodction power consumption?

[unique cypress]

https://www.reddit.com/r/SatisfactoryGame/comments/1g8pk8u/the_costs_of_powering_100_nuclear_power_plants/

[lone jewel]

i strongly recommend utilizing outposts/blueprints in modler, makes things so much easier/more organized

[agile junco]

Does anyone know why this broke? I want to start using modler, but I feel like I don't understand this.

[unique cypress]

Give it a while. When it was just released, I thought I finally had a balancer testing app. Unfortunately, even a simple 4:4 took like half an hour to calculate. It has gotten much faster since then, but I don't think it can handle even a 6:6 or an 8:8 within reasonable time even now

[lone jewel]

as an example, here's my nuclear setup:

[agile junco]

Oh weird...

I'm not immediately seeing why this work make the logic the app needs to do complex.

[lone jewel]

the parts that have a lot of steps involved have their own outposts, such as for control rods:

[unique cypress]

Switching it to manual would make it much faster, but in exchange you'd have to define all inputs and outputs

[lone jewel]

i've found that to not actually be true, there's very little difference between automatic and manual mode

[agile junco]

But it doesn't consider belt saturation, so isn't the math just divide the inputs by the number of outputs for each splurger?

[lone jewel]

there are some cases where you need to set a limit, but much of it works the same

[unique cypress]

or alternatively, you are in manual mode already. that also gives ? on all numbers sometimes

[lone jewel]

only no calculation mode requires you fill everything in

[unique cypress]

I'm building a 6:6 to test how fast it is and it gives me ? in manual mode

[lone jewel]

the ? generally come from the math not being possible to solve, not form using manual mode

[lone jewel]

there is nothing being consumed here either

[unique cypress]

full mode would give me 60 on all outputs

[lone jewel]

but yes, that's a mch more complex calculation

not sure why you'd ever need to model something like that though tbh

[agile junco]

Why would the math in my example be impossible!?

[lone jewel]

i said generally

doesn't mean every case is the same

[agile junco]

You can see adding one edge to another splitter broke it.

Just really makes me confused what was code is doing lol

[unique cypress]

ok, 6:6 built, we'll see how long it takes to give me any answer in full mode mode

so far, it's a whole minute in

[agile junco]

I was going to try a 4:5 since that seems like it might be fun.

Or a 3:5 actually.

[lone jewel]

what exactly are you trying to figure out exactly?

[unique cypress]

how much it decides to pull from each miner

[lone jewel]

belt balancers in 1.1 have become extremely easy

[agile junco]

How did they change?

[lone jewel]

that can scale to any size

[unique cypress]

nothing in 1.1 made it any easier or harder

[lone jewel]

yes it did

[unique cypress]

to design, at least

[agile junco]

I don't see how.

[lone jewel]

vertical merging made it 100x easier

[unique cypress]

so far, I still haven't used vertical splitting in any of my desings

[lone jewel]

want a 2:5 balancer? build 2 1:5 balancers stacked, merge them vertically

[agile junco]

Oh sure. I haven't built one that uses that yet though.

[lone jewel]

want a 12:12 balancer? build 12 1:12 balancers stacked, merge them vertically

[unique cypress]

except you'd have to balance them 12:1 vertically

[lone jewel]

you can even build them in smaller modules and connect them horizontally and vertically

[unique cypress]

which is not nearly as easy

[lone jewel]

well, that depends on if you want the inputs to be balanced i guess

[agile junco]

Yea I'm confused by this

[unique cypress]

5 mins in and still nothing btw

[lone jewel]

if you only need balanced outputs, that makes it super easy

[lone jewel]

i'm not surprised tbh

[agile junco]

It also uses extra splitters/mergers than needed in some cases.

[lone jewel]

the full algorithum is extremely slow when you interconnect things like that

[unique cypress]

I kinda can't be bothered to keep it running lol

[agile junco]

Why?

[unique cypress]

switching it to manual just did this

[agile junco]

It should just be sums and ratios, no?

[lone jewel]

🤷 just is

[agile junco]

lolol ok

You made it sound like you knew why it was complicated.

[lone jewel]

i just have a lot of experience with it is all

it can be super fast 1 second, then you make a single connection and it can take several minutes to do the calculation

this type of calculation is one of the few that manual mode is not able to do

[unique cypress]

if you don't need input or output balance, you can do this kekw #math-and-meta message

actually works as good as a balancer inside a factory

doesn't work for trains sadly

[lone jewel]

not quite sure what that's doing tbh

vertical splitters into horizontal mergers?

[unique cypress]

horizontally stacked vertical splitter manifolds into vertically stacked horizontal merger manifolds

every input is connected to every output, like in a balancer

[lone jewel]

yeah but that wouldn't be balanced in any way

[unique cypress]

basically the same thing you suggested earlier, but neither input or output is balanced

which is fine if you use it inside a factory

[lone jewel]

well in my suggestion, the outputs are fully balanced

[agile junco]

It's a mixer.

[lone jewel]

just the inputs are taken in unevenly

[unique cypress]

but the inputs aren't. in this thing. neither inputs or outputs are balanced

[lone jewel]

inputs aren't balanced, but outputs are

[agile junco]

Your designs actually seem like a decent compromise for blueprint reusability and 1 sided balance.

[unique cypress]

but you don't need output balance for feeding machines

[lone jewel]

you'll always end up with the same amount on each belt if they are not backed up

i use them on the output side of my large factories

for feeding train stations

[agile junco]

Yea would work well for loading trains.

But that's why I want to build a set of custom N:4 balancers. Since I generally wont need more than that for trains.

N:2 would also be good to have.

[unique cypress]

yeah, for un/loading a train, you need balance

but not for feeding machines

[lone jewel]

i have never needed one that balanced both input and output, i generally simplify connections between machines as much as possible

[agile junco]

Only one sided though, technically.

[unique cypress]

hmm, yeah, technically you don't need input balance for loading a train

nor do you need ouptput balance for unloading one

[agile junco]

Right

[lone jewel]

yeah, just overflow to sink on input side

[unique cypress]

it only matters that the train is balanced. the factory doesn't care

[lone jewel]

yep

[agile junco]

Or just back up if you don't care.

[lone jewel]

right

[agile junco]

Actually allowing for overflow is something you should do in either loading station.

[lone jewel]

it can be used to balance output to a train, or just to allow output to goto any train, in my larger factories i have a lot of trains that each can be taking a different amount of items

[agile junco]

Backed up production still consumes the same power right?

I guess it depends on the machines.

[unique cypress]

machines not doing anything consume 0.1 MW

[lone jewel]

the balancer allows each platform to always have 1200/min unless the total throughput exceeded what was being made

[agile junco]

Right I forgot.

[agile junco]

Minus loading time

[lone jewel]

no, each platform still has it's own industrial container

[unique cypress]

well, no, with mk6 belts, you can have 1200/min/platform. you just can't get 2400

[lone jewel]

which is fed by a mk6 belt from the balancer

[echo vale]

here's how I've oriented my oil pipes, the problem being an uneven input to the refineries

[lone jewel]

2x1200 to the platform so 1200/min is easily achievable

[unique cypress]

with stack size 50, it's barely achievable

[agile junco]

You don't need balancers at all for that.

[unique cypress]

but I imagine it would be much harder to get 1200+/min of something stackable to 50

[lone jewel]

the balancer is for the factory output total

if all my platforms were sinks, i'd get less than 1200/min per sink

escentially, as long as i don't exeeed for example the 9600/min of my aluminum ingot factory, each platform that's in use can reach 1200/min

it allows the trains to be used a lot more flexibly

only the load side is setup that way

[agile junco]

Why not run each platform slightly below 2400 and buffer, then make more platforms than you need and run semi-loaded trains?

[lone jewel]

because i don't need to know what's going where

so long as i don't exceed 1200/min on a single platform, and don't exceed the 9600/min total of the factory, i know everything will work efficiently

and i can have any train take the aluminum anywhere it's needed

without needing to have 1:1 load/unload stations

[agile junco]

Isn't 1200 per platform easy to load, since you can push 2400 if needed from a buffer?

[agile junco]

Ah! I see

[unique cypress]

aluminium ingots stack to 100. the most you can load into a platform is 1793/min

[lone jewel]

the main purpose is for loading flexibility, not to maximize throughput

[agile junco]

Wait, why does stack size affect how fast you can load?

[unique cypress]

because of the loading pause

which happens once every 32 stacks (at least)

[lone jewel]

because you'll need more trains

[agile junco]

The smaller the stacks the more time it dominates in the loop. Got it.

[lone jewel]

which means more pauses

[agile junco]

Yea, thought that was it. Just wanted to make sure.

[unique cypress]

max throughput is 32s/(32s/2b + tp)

because you unload 32 stacks, and that happens once every x minutes

and x is the time to move 32 stacks with 2 belts plus the pause time

[lone jewel]

for things that stack to 100, i (generally) try to target 600/min/platform

if it's a simple 1:1 load/unload that is

but i do have 1 setup where i'm transferring 1200/min/platform at a short distance

but for example, on the input side of that aluminum factory, i'm bringing in 9600 bauxite per minute, with 600/min/platform

[unique tree]

how would you make it use 300 uranium ore? and also balance the uranium waste? and what do you do with the plutonium fuel rod, just sink it?

[unique cypress]

that's for 300: https://www.satisfactorytools.com/1.0/production?share=RWFu6KlLgsXdChAFnVdm

[lone jewel]

yes, just sink it in most cases

[unique cypress]

as for the rods, sinking them or using them to power drones is the only way to remain waste-free

burning them for more power will create plutonium waste, which is much harder to get rid of

[lone jewel]

my plan for nuclear involves changing some over to ficsoium, but only some, the rest will get sunk

[unique tree]

okay thanks, powering drones is a great idea if i find it needed

[lone jewel]

mostly just want to do that for the challenge of it, but the amount of SAM it requires means you don't want to / can't recycle it all

[lone jewel]

if you're curious what those large balancers look like on the side of my factory:

so each blueprint there contains a 9x3 balancer on the above floor, and 3x3 on the bottom floor

[agile junco]

Using blueprints to make factories like that would definitely be more efficient than my approach.

[lone jewel]

i connect them on the one floor all horizontally with splitters, and between floors vertically

basically allows you to scale it up easily to any size

[lone jewel]

building factories in full blueprints like this has greatly improved the scalability and reliability of my factories

there's still plenty of logistics which are connected outside of the blueprints, but overall it kinda changed the game for me when i started building like this

it does of course somewhat limit the design of things though

but I'm terrible at design so it works out 🤷‍♂️

[stoic gorge]

how can I make so I have kind of an "infinite storage" bc my byproduct is making energy on oil phase I if the normal product fills the industrial storage the product will stop the energy production

[unique cypress]

sink

[stoic gorge]

BROOO IM SO DUMB

thank you

[stoic gorge]

do I need a smart splitter for that or normal splitter will work as well?

[unique cypress]

depends. normal splitter will split in half/in thirds by default

if that's fine with you, then you don't need a smart

[stoic gorge]

makes sense

[unique cypress]

generally, if you have a sink, you usually want to set the overflow there (with a smart splitter), so you're nor sinking useful items. but it's not always necessary and you can just use a regular one

[unique tree]

ohh thats a sick design, do you mind sharing your design? i would love to try it and maybe scale down

[night crest]

Dude your designs are so clean

[lone jewel]

my model is actually already designed to be scaled, it's designed for 5x420/min uranium ore (for the full 2100 on the map), you could definitely modify the numbers though to work with 300/min, i considered doing that myself so it would be 7x300

i can DM you the modler file

[lone jewel]

thanks! i put a lot of time into it to make sure all the numbers are as clean as possible and can be logically built out easily

[unique tree]

yes that would be awesome

[lone jewel]

sent

i've learned that simplifying logistics as much as possible between steps makes things a lot easier to get running reliably, which is especially important for nuclear

so generally speaking, i keep all my machine counts the same or multiples of each other

as an example:

the latest modeler makes this extremely easy to do with the + - buttons for clock rate

[night crest]

ty for the inspiration, I just made my own setup for 90GW with some diff recipes

[night crest]

Me and this dude are trying to do something with 1 impure node only to get the nuclear experience

the numbers aren't perfect either

[lone jewel]

Yeah, it can take some time to make the numbers better, for example I've used multiple different alt recipes combined to make some of my numbers easier to work with

[stoic gorge]

what is train most used for? bring raw resources direct from the miner or bring the final product?

I know I can play the game the way I want but wich is more beneficial

I fell like its bringing the raw resource

[lone jewel]

both

[stoic gorge]

ok then

[lone jewel]

it depends on the specific resource and how much you need

iron for example is available everywhere so there's less need to carry the iron ore by train

[agile junco]

Anyone here know the best way to model stations in modeler?

[unique cypress]

containers, probably

[lone jewel]

i mean you can use a storage container

but personally i don't at all

[agile junco]

Ah yea.

What's the point of the filledness of the containers?

[lone jewel]

it affects how things are calculated

just click the ? button to get a description of them

[agile junco]

Oh I didn't even see that, thanks.

[ornate grail]

I’m doing a short video on storing nuclear waste. How long (in real time hours) does it take to fill up one industrial storage with uranium waste if you are producing 10 nuclear waste p/m?

[lone jewel]

🤔 you're aware that you can just recycle it into plutonium fuel rods and sink them, right?

500 x 48 = 24000 / 10 = 2400 minutes / 60 = 40 hours

but seriously, just recycle and sink them lol

[wind spade]

honestly storing is fine

[agile junco]

Oh wow do I wish there was a way to select a group of nodes and convert it to an outpost. I need to redo a lot of this now.

[lone jewel]

there is

[agile junco]

oh!

[lone jewel]

just select them, cut, go create outpost, go into it, paste

then connect the inputs/outputs

[agile junco]

Oh, that's good enough, though convert to outpost would be even better.

[lone jewel]

it's a little bit of work but not a lot

[ornate grail]

Yes absolutely

Thank you for the answer btw, want me to shout you out in my video?

[lone jewel]

not necessary 🙂

[ornate grail]

Sounds good, thanks again though

[agile junco]

Modeler is awesome, but this is still a rats nest lol

forever probably.

[unique cypress]

yeah, when you see tha for more than 2 seconds, you're doing something it really doesn't like

[normal latch]

djGrrr you got any advice for someone who is nearing unlocking tier 7+? gonna build my first fuel power plant first, but then will get the parts ready for next tier

[agile junco]

I'm just going to manual mode and entering all the numbers of buildings manually...

[normal latch]

i just got an alt recipe for modular frames which doesnt use screws, so i think im about to pop off

[unique cypress]

you don't need to put in all the numbers

[agile junco]

probably not, but it's easier that way.

[unique cypress]

it can figure things out if there are no other options

[agile junco]

Less room for it to do something weird.

[unique cypress]

it'll give you red nodes when there's something it can't figure out

[agile junco]

I don't see that

[unique cypress]

I guess only actual machines can be marked with an error then

that's a little less helpful than I thought it'd be

[agile junco]

This is happening too.

But also the numbers it's choosing are sometimes too high.

[unique cypress]

so it does show issues in outposts

what does the inside of that one look like?

[agile junco]

Not 100% clear

[unique cypress]

it thinks it needs 27 refineries for steamed sheets to meet some demand downstream

[agile junco]

idk why it thinks I need 27.16 refineries making sheets.

yea, but it won't once I'm done telling it things.

🤞

[unique cypress]

I think you generally need to give it numbers for machines before sinks, containers and disconnected outputs (which are unlimited)

[agile junco]

It says it still doesn't know some downstream, so it's honestly kinda inconsistent here.

@unique cypress how do I tell it I only want 10/min. This didn't work like I expected it to.

[unique cypress]

lmao what

I thinks it needs to make 110 motors

for some reason

[agile junco]

This app is great UX, but the math seems... iffy

[unique cypress]

what are you making them out of? any oil? anything else with byproducts?

[agile junco]

Oh interesting thought.

I don't think so.

I could send you the save, but it's a lot.

Interestingly, it still hasn't figured out how much rubber it thinks I want.

Let me try adding other output constraints.

dude, this makes no sense.

[agile junco]

man I really wish you could have tabs going in this thing.

I would have a panic attack.

Is this some kind of race through the milestones without blueprints?

I'm hoping to never start a new save and have to deal with leaves/wood again >.<

[wind spade]

you can in some other planners 🙂

[agile junco]

Dude if SFtools saved node positions I'd be SOOOO happy.

[wind spade]

it does until refresh

[agile junco]

Yea, but my computer sucks at going to sleep.

Also, being able to make "outposts" is a great feature.

[unique cypress]

put it to hibernate it instead

[agile junco]

Does windows even do S4 anymore?

[wind spade]

the problem is much more complicated, it doesn't save the production line at all, it calculates it again when you refresh 🙂 and there's tons of technical reasons why that is

in new tools all that would hopefully go away

[unique cypress]

I hibernate my PC every day

reboot it only like once every 2 weeks

[agile junco]

new tools sounds exciting.

[wind spade]

I sometimes reboot once in two months or something 😄

[agile junco]

I'll look into it.

I found it pretty hilarious when I tried to sleep windows to maintain the node layout and it woke up again in like 5 minutes.

[unique cypress]

pretty sure it does wear out the SSD much faster than normal

[agile junco]

I use linux a lot more than windows.

[unique cypress]

after all, hibernate dumps RAM onto the drive

[agile junco]

I only have 16GB, that's not a huge deal.

Though I was going to buy another 16 this weekend...

[unique cypress]

my previous boot SSD

(current boot drive is untouched but it's only a few months old)

[agile junco]

That's a lot of I/O

I think I figured out at least part of the issue... The first image is when I have both rubber productions sent to recycled plastic and the output node. The second image is when I use a splurger.

Why they are different, don't know.

[compact magnet]

can anyone tell me is this solt of how many item per minte or is this how many items like exple what in the ref what number do i put in the satisfactory modaler

[wind spade]

it says right there?

[compact magnet]

so that number i put on the black line

[wind spade]

no idea how modeller works

I don't really use it, since it's for factory planning, which I'd do ingame instead

[unique cypress]

isn't that box for machine count?

[compact magnet]

idk it ether items per minte or that

[unique cypress]

test it then?

[compact magnet]

ik i did but dont know what number is right and how many machines i need

[agile junco]

Here's a fun one... First is with the sink, second is without. It's breaking later things.

OK, I'm learning things. It's safest to put an awesome sink next to each resource input, so they don't force you to over produce.

Or just use storage containers and do the math for the number of extractors later.

I think the issue is it doesn't allow machines to run idle at all.

err, idk, it should just be a fractional number.

Ah! if the "limit" is set but it's an outpost and fed from an input, then it tries to consume the full amount it seems.

Adding a sink fixes it, though in reality I'll just let the machines idle.

[agile junco]

I'm wasting way too much time tying to hack this app into submission

[unique cypress]

It's not really a solution tho. Now your iron and coal consumption is inflated

I feel like it really shouldn't work that way

[agile junco]

yeaaa

[unique cypress]

But I've never used outposts so I have no clue

[agile junco]

The only reason I'm trying this thing is for the outposts.

[severe girder]

can you guys please suggest to me a recipe or if I should rescan? or keep so if its a bad 1 I don't get it again later? lol

[unique cypress]

cus like in this example everything is calculated properly from 1 set value

and it is set to manual

[vapid gorge]

recipes aren't bad or good. They have uses in different situations

if you decide a single recipe is the only good one of a group it'll be 'bad' then because you don't consider your options

I've used every single recipe in your images

[agile junco]

Maybe I'll try seeing how it works without outposts... maybe I'm still doing something wrong. It's just way too many connections for me to manage without them

[severe girder]

are there any ones that point out to you as much more time saving than others?

[vapid gorge]

I don't know what kind of factories you want to build or how

[severe girder]

or save resources, thats what i ean by better

[vapid gorge]

basically every single recipe that 'saves resources' costs a different resource

[vapid gorge]

the only kinda counter argument to that are the 'pure ingot' recipe because that's extra power, space and water. Which are kinda free?
but I also hate those recipes

[unique cypress]

Put a white dot next to the ones I'd want ↑

[vapid gorge]

kyo also makes a lot of subjective opinions without much backing

[severe girder]

thankx :D

[vapid gorge]

you may as well just pick recipes at random to use if you get a rando to just name stuff

[unique cypress]

Almost all of my choices are decided by "what saves me the most time"

[severe girder]

i think solid steel ingot i shall get bc steel is annoying xD

[vapid gorge]

that's what I'd call a generalist recipe

they are generally very low tier item recipes that cause more changes down the line

there's also specialist recipes that are more niche but change things dramatically for their section of production

[unique cypress]

You can also almost entirely get rid of the need for steel with the right recipe choices

It's gonna cost a lot of iron, but you can

solid steel was a much more useful and important recipe back before you could do that

[agile junco]

I think I understand outposts now.

You need to make sure to pass back out excess.

otherwise it tries to use it all.

This is my full factory... need to figure out a clever way to organize it.

[unique cypress]

idk how you're doing them because for me a machine in an outpost works exactly the same as a machine outside

[agile junco]

There's just lots of weird things I'm noticing that I cannot explain fully.

For example, trying to model passthrough outposts seems to break and I don't know what. That is an outpost which takes in and outputs it's input. Like why is it suddenly -10!?

[unique cypress]

it seems machine count limits have priority over sink limits

wait no

this errors out the sink

hmm but so does this

but this caps the smelter below the set value instead

even if there's one limit vs several higher ones

and this is perfectly capable of calculating a result even with just sink limits

[unique cypress]

My suggestion would still be to only set limits on sinks and make it possible for the calculator to figure everything out from that

[agile junco]

I'm going to make a new save and try that now.

The numbers are close, and some negative are to be expected becuase I'm using whole values. I hope your idea will just fix this.

Behold! A factory.

[wind spade]

usually the answer is "whichever recipe you like the most"

people have different preferences, goals, and situations, so it's usually best to decide based on your own situation rather than asking other people to basically roll a dice for you

[agile junco]

@unique cypress you know what... I bet if I made the outposts blueprints, it would allow for fractional amounts and it would have been easier the way I was doing things.

But this is good too. Since I was going to need to validate my numbers and adjust some overclocking anyway.

[agile junco]

This tool rocks!

Needs remarkably little imports.

[dense marten]

I need to get 24 mk2 pipes of water to the grass fields, what is the best way for me to do that (I already have a factory built on top of the waterfall so it would be difficult to get it from there)

[crimson moat]

1: build your stuff at the water instead. There's more water nearby.

2: autoconnecting 24-pipe blueprint

[dense marten]

Can't change where I'm building it as I'm too far into the project, and was going to auto connect the blueprint anyway, I meant where should I get it from

[orchid brook]

is this too much?

[topaz jetty]

Boy I remember a time when the modeler was used by almost no one

[rose ingot]

guys i'm trying to get some uranium but i swore this was pure right?

[thorn trail]

there are no pure uranium nodes

[rose ingot]

I swore it was pure though... i need more sleep..

[frosty owl]

Too much lack of zoom to be easily readable? Yes

[hoary oar]

there have never been pure uranium nodes

[rose ingot]

;-; well damn

[hoary oar]

early access had 3 normal and 1 impure uranium node
1.0+ has 2 normal and 3 impure uranium nodes

[low spruce]

these refineries in the circle produce 600/m water in biproduct and fill the 3 refineries that take 600/m water. will moving the line up above the 3 refineries and then down into each solve it or should i split into 2 lines of 300 on each side of the 3 refineries they currently do not stay stable

its all a straight line i hate moving 600/m fluids it should just work mane

[low spruce]

somehow this works

[vapid gorge]

direct feeding waste sometimes work yes

[low spruce]

i wish it would always work 🫠

[frosty owl]

Maybe having the input pipe feed the 3 refineries from the middle rather than the side can suffice, using one junction instead of more than one (it would also be convenient if you didn't cover up the pipework with arrows 😅)

Pipes do like simmetry, so if you can achieve some simmetry easily (eg: middle-feed Vs side-feed) it's almost always worth it

[low spruce]

yea true sorry 🤣 but I'll try that