#math-and-meta

1 messages · Page 22 of 1

oblique notch
#

Making some assumptions and telling it to stop if it its path to root has the recipe already

prime orchid
#

just change the game code EZ

wind spade
oblique notch
#

that does mean that it probably mises out on some possible chains that involve putting things back into each other a lot but it gets most of the common ones its good enough for me.

oblique notch
oblique hollow
#

the question is if the calc results are even readible / useful after that

#

quick lynk, write a recipe result decoder jace_smile

oblique notch
#

they arent really 😛

wind spade
oblique hollow
#

yeah but not all of them

oblique notch
#

basically it looks this: with each circle being a part and each square a recipe to produce that part. As you can imagine... it gets intense.

Every time it goes to add the children square recipe nodes, it checks back along the path to root to see if it has that recipe already in that singular path. If it finds it, it skips it. (preventing the infinite loops, but does loose out on some edge case scenarios).

However, it certainly is not efficient in terms of processing 🤣

image.png
#

but again, I wasnt going for a produciton planner like SFTools, but the ability to compare any two chain that I wanted.

In the end, while I still can do this, i ended up simplifying it and narrowing it down so that I provide the recipes I want to use for the chains and it will build just the one graph then I can compare them. Which... is pretty much what you did Greeny, except you have the optimization done through the linear algorithm but I wanted to manually pick and choose to see for myself.

#

which is ... very funny that we basically ended up at the same place with no communication and without me looking at your code at all

wind spade
#

yeah the "problem" with manual picking is that you can't guarantee that the recipe will be used... unless you build the whole tree yourself

oblique notch
# wind spade yeah the "problem" with manual picking is that you can't guarantee that the reci...

yah. Exactly. So while I could have used your tool to do exactly what I wanted as well - just by limiting what recipes were available - it would have been annoying to me to do and it doesn't include all the data i wanted either (total # of buildings, footprint of all buildings) or id have to do some calculations from what was outputed to get what i wanted. So I have a small front script that just outputs the recipes available for a given node and then asks you to select. This does mean if several leafs have the same item i could select different recipes for each branch if I wanted to (though I didnt really code it well enough to know what branch im currently on lol)

wind spade
#

I do have some plans for something that could be helpful for your case, but again, tons of work, not enough time 😄

oblique notch
#

🤣 know the feeling.

And I like programming, so It was fun to do it myself. I have a much better understanding at designing weighted graphs and binary trees now than I did before 😅

cinder silo
frozen kiln
#

Is there gonna be mk 3 for pipes?

odd spire
#

Doesn't seem.like it

wind spade
prime orchid
#

ok i actually think i got it to work

#

have i done anything wrong that you can see? i had to disable fertile uranium to force it to use more waste

frosty owl
#

You can maximize the TPR production

median heath
odd spire
frosty owl
#

No, that's SFTools.

tropic hawk
golden smelt
#

what would be a good amount of motors per minute, if I have mk2 drills, pure nodes, and mk3 conveyors

#

cause rn im only making 10 a min

restive heron
#

So I’m thinking of making a new oil extraction job hoping people with more knowledge than me can recommend focusing on byproduct or fuel and how one would go about that

versed violet
restive heron
#

Havnt looked at numbers of polymer resin but I’m guessing that’s probably best to rubber plastic production since it adds another layer of refinement?

versed violet
restive heron
#

Diluted packaged fuel feels a tad excessive but good to know that’s the optimal

#

Would regular diluted fuel not work as well?

versed violet
#

IF you have blenders, yes, easier to setup

restive heron
#

Gotcha

#

What about maximizing fuel you mentioned that one was good for rubber and plastic

#

I’m thinking I may want to go to each oil location and have it prioritize fuel and make a single byproduct

#

Stay off of Nukes as long as I can

versed violet
versed violet
wind spade
dense cave
#

So the first time I built my thing, I simply made a giant pad with a conveyor bus in the middle

#

then I can pull things from the bus build as many as I need or expand from there, then feed it back to the bus

#

its ugly, but it works wonderfully

versed violet
#

the bus method is not recommended in satis.

#

The common way is to build factories tree-like

dense cave
#

why would you say it isnt recommended?

restive heron
#

Cause if I make enough fuel I’ll probably be flooded with more plastic and rubber than I need anyways

versed violet
#

One pure node makes enough of all for all my needs.

restive heron
#

I’d rather have it in excess instead of having to go back to the drawing board

#

Power is the one thing I don’t mind going overboard on

wind spade
dense cave
#

not sure what you mean by bus, but this is what I mean. The idea of a bus is that you have multiple production lines with various inputs, some of which are repeated.

20220215203928_1.jpg 20220217170044_1.jpg 20220217165905_1.jpg
wind spade
#

bus is the belt stack in the middle with splitters and mergers

#

and it's not recommended

versed violet
#

too many belt = too many lag.
and bus implies centralized factory which also means bigger lag.

wind spade
#

and since you know input and output of your factory, you can just do miner -> factory -> storage

versed violet
#

Is there a way to tell from the ingame map, if radar tower is powered?

oblique notch
#

yeah, unpowered wont show all the special other icons

versed violet
#

and its not likely the tower would simply have no artifacts or fauna/flora in range?

#

are you sure its plugged in?
Yes, I'm sure-
... damn

image.png
#

I did find one with no singlas tho, even when powered. Its on very edge of map.

toxic latch
#

Never rly looked into the game before buying it but it seems i have got some way to go

maiden steeple
#

Yeah, there’s 4 space elevator phases and 8 tiers of stuff, and most people consider the time before unlocking everything the “tutorial”

tropic hawk
tropic hawk
tropic hawk
ionic bluff
tropic hawk
ionic bluff
#

I use load balancing when I have room or when I what a perfect split.

golden smelt
vapid gorge
# golden smelt 45-81 rule?

I believe it's a ratio rule to avoid repeating numbers in % production speed to get permanent even flow forever. If you have a repeating decimal every few thousand hours you'll get a stutter somewhere. I think. I never bothered with it because that level of control isn't interesting to me\

median heath
# golden smelt 45-81 rule?

So the premise of the 45-81 rule is getting production in "clean numbers" with respect to clock speed.
As you may or may not know, the game enforces a 4 decimal limit on clock speeds (it does not care about parts per minute, nor does anything in the game actually operate on parts per minute, EVERYTHING is done in individual cycles with respect to clock speed for cycle time)

Example: if you tell a Smelter to do 20/min instead of 30/min, the clock % changes to 66.6666%.
The UI will say 20 ppm, but this is a lie to make you feel better about yourself. The machine truncates anything after xx.xxxx%, so you do not actually get 20/min, you get 66.6666% of 30 (which is 19.99998).
Now, to MOST people, this doesn't matter. But I am a stickler for precision, so not actually getting 20 bothers me (even though you'd literally see a production difference only after a couple hundred thousand cycles).

So the math was done to find out how to make production lines "behave", and the first thing we solved was the 81 part of the rule. With respect to Rubber/Plastic - always do them in multiples of 81 and all machines in the recycled loop automatically fall within the 4 decimal limit. No repeating numbers, no imprecision.
The 45 part of the rule applies to non-oil products. Do your final product amount in multiples of 45 and everything in the production chain automatically behaves without you needing to think about it.
Exceptions: Caterium Circuit Board, Pure Iron, Fine Concrete
(CtCB and PureFe usually outweigh the rule because you'll be doing them in such large quantities that they fine a common multiple somewhere up the chain.
No one should ever use Fine Concrete.)

Note: 45-81 is the lowest whole number multiple, but you can cut them in halves to achieve closer to your actual goals.
I.E. 45/22.5/11.25/5.625/2.8125, or 81/40.5/20.25/10.125 (going lower than 10.125 on oil causes issues)

So if you needed 100 Plastic, I would instead say to make 81+20.25=101.25

digital crystal
#

To produce rotors t is better to use the normal rotor recipe with steel rod and screw alts than steel rotor or is their a better option

vapid gorge
digital crystal
#

i am making 1250 steel ingots

vapid gorge
#

Copper rotors are a favourite for reduced resources needs, but I like steel rotors since I’ll often be making pipes for other things and I like that sort of simplicity

#

Well do you need the steel ingots for something else nearby? If so maybe use copper rotor

digital crystal
#

how much copper does the copper rotor need since i am using a lot of copper for ai limiters

vapid gorge
#

A fair chunk. You can get more copper by using pure copper refineries and stretching it out that way?

But if you aren’t using the steel for other things steel rotors is nice. And iron wire is a good mix too

digital crystal
#

ok

#

i will use steel rotors then since it would be more simple

vapid gorge
# digital crystal i will use steel rotors then since it would be more simple

In general any particular recipe is good depending on your situation and location.

So something like iron wire might mean you can build a factory somewhere with no copper even if you need wire or cable.

Some people like using Caterium circuit boards and caterium computers together because you need feeder resource types

#

I like silica circuits and cat comps because I can use more resources to make more computers 😄

fierce cypress
#

crystal comps 😛

digital crystal
#

thanks for the help i also like silica circuits since you don't need plastic

vapid gorge
fierce cypress
#

steamed copper sheet 1:1 with copper rotor 👀

#

interesting find

#

but then again - screws

prime orchid
vapid gorge
#

Basically on my map anything that can avoid using copper I go that way XD

digital crystal
#

Automated Miner 🤡

fierce cypress
fierce cypress
vapid gorge
#

xD

digital crystal
#

has anyone ever used charcoal or biocoal

median heath
fierce cypress
digital crystal
#

yeah i can't think of anything good for them

fierce cypress
#

apparently its cheaper in terms of coupons if you buy biofuel and use biocoal instead of buying coal jacelul

#

but thats a pretty sad use case

digital crystal
#

maybe if you just started phase 1 and need power but have no coal nearby

fierce cypress
#

its completely useless imo

prime orchid
# median heath So the premise of the 45-81 rule is getting production in "clean numbers" with r...

Now, to MOST people, this doesn't matter. But I am a stickler for precision, so not actually getting 20 bothers me (even though you'd literally see a production difference only after a couple hundred thousand cycles).

you're chocking this up to basically no big deal, and rounded to 66.6667% is a smaller difference and one that doesn't cause bottlenecking. and when it shuts down it's not consuming power. that's not a problem imo. there's bigger issues like mk.5 merger manifolds trying to handle 780/min

median heath
#

when it shuts down it's not consuming power

Technically false. Machines on standby consume 0.1 MW

you're chocking this up to basically no big deal

I do not deal with degrees of severity. Something either is a deal, or is not. There is no "small" or "big" deal.
Imprecision bothers me. 66.6667 not being exact is a yes/no.
I can avoid it, so I do.
That's the entire premise of the rule - avoiding imprecision.

prime orchid
#

yes and i agree i am the same way.. however in a game that has other bigger issues that make it imperfect, i'm willing to see a difference of .00004 over-clocked perfect enough to not bother with the headache of trying to force things into certain multiples

vapid gorge
#

I think they are weird but they are allowed to be so

median heath
#

i am the same way.. however

I don't have a "however." So I would say we are not "the same way."

vapid gorge
median heath
#

I'm not memeing?

fierce cypress
#

Sev is Sev'ing simon_smile

vapid gorge
#

The ‘we are not the same’ meme image 🙂

prime orchid
#

snapping is another thing i'm probably going to have to give up on when trying to build giant factories. it takes way to much time trying to force things to snap when they aren't doing it automatically

vapid gorge
#

Snap?

#

What things aren’t snapping?

prime orchid
#

plenty

#

the biggest one i have issues with is junctions not getting snap points from floorholes

#

or even actual pipes for that matter

vapid gorge
#

Oh, I don’t think junctions are supposed to snap to those

prime orchid
#

or supports

#

they don't seem to snap to anything except other junctions and inputs/outputs of machines but most of the time that doesn't even work

vapid gorge
#

And you should avoid floor holes anyway and clip pipes through. You can put decorative holes after though

fierce cypress
#

i do that with belts too because im extra paranoid simon_smile

prime orchid
#

reason?

prime orchid
fierce cypress
vapid gorge
fierce cypress
vapid gorge
#

Also having very tiny sections of pipes can cause flow issues

prime orchid
#

i haven't had a single issue

#

i use them on every fluid machine

vapid gorge
#

@prime orchid if you want to snap junctions vertically like that you can put a pipe support stretched up from the ground with the pipe bit angled down. The clip the pipe up and put a junction on it

prime orchid
vapid gorge
#

Plus if you clip the pipe through and have a decorative hole it’s fewer world objects

prime orchid
vapid gorge
prime orchid
#

well still mentally i'm like that is wrong

vapid gorge
#

At least from U4

prime orchid
#

it feels like cheating

vapid gorge
#

Ah XD , part of my clean pipe system methodology is as few junctions and pipe segments that you can possibly get away with for smooth flow so clipping it through with a decorative hole is standard for me

vapid gorge
toxic latch
#

XD

fierce cypress
#

||the golden cup||

cinder silo
#

Merger test fun.

cinder silo
# cinder silo Merger test fun.

Sorry about the above, I've never created a thread before, the above is specifically about issues surrounding conveyor mergers and throughput issues specific to them.

supple spire
#

how long will it take to get 500 if im making 10 a min

#

i forgot how to solve this

cinder silo
#

Divide by 10.

supple spire
#

ok

#

will that work for 12.5?

cinder silo
#

Point is you're dividing the total by the number per minute, which gets you how many minutes it'll take.

maiden steeple
toxic latch
#

Im prob gonna reset then

#

Trying to get it done faster and better with new knowledge

maiden steeple
#

I personally am going to start with coal power researched from now on

cinder silo
#

Coal is quick enough to research at least.

maiden steeple
#

Yeah, I just don’t enjoy the bio burner phase much

#

I’ll still use them a bit ofc, I just want to be able to set up auto power once my area runs out of foliage

toxic latch
#

The thing im strugling with is expending after phase 1 of spece elevator

strong saffron
#

I just rushed phase 3 for trains so I can transport my infrastructure from grass plains to desert

#

My disappointment when I remembered oil was needed first

median heath
#

I'm usually knocking on the door of nuclear before I build a single train.

fierce ruin
#

Hello, its better to use trucks or trains?

wind spade
#

depends on a situation

tropic hawk
fierce ruin
tropic hawk
fierce ruin
median heath
#

Trucks can handle up to and including 1560/min throughput.
Anything above that is train territory.

deft lichen
median heath
cinder silo
#

I still haven't tried trucks again.

deft lichen
#

for some reason it just never occurred to me that truck lines actually have infinite throughput because it's possible to keep adding more trucks and more truck stations

#

not saying that it's a good idea, just that it's totally doable

cinder silo
#

I find myself imagining a pile up and a line of trucks after trying something like that.

#

Still having trucks and drones around would make the world look more alive.

barren elm
deft lichen
#

possible, I'm just theorizing, it might not actually be feasible to use

median heath
#

Do it right and you can unload into 2 stations at once 🙂

golden smelt
median heath
golden smelt
#

I fused them lol

#

I ment 81

median heath
#

81 is for Plastic/Rubber when using the Recycled Loop.

#

45 is for non-oil products.

golden smelt
#

So wait is 81 is not for all of oil, just a recycled loop ( I don't know what that is)

#

@median heath what exactly are the cycles

median heath
#

Recycled Loop

golden smelt
# median heath

So you keep sending plastic and rubber back and forth to make more rubber?

cinder silo
#

That's the power of recycled, you can make shit tons.

median heath
#

Yes.
You could probably google "Satisfactory Recycled Loop" if you want a more thorough breakdown of it.

golden smelt
#

And that's the only time the 81 rule is in place?

median heath
#

Yes..

golden smelt
#

Ok so I think it get it

#

You want non oil products to be multiples of 45 to maximize production with the exception of you are making a huge amount of something. While 81 is for a recycled loop to maximize production with rubber and plastic

median heath
#

It's not about maximizing production.
It's about eliminating imprecision.

golden smelt
#

Ok

wind spade
#

fyi it's also so little imprecision, that it in 99% of cases doesn't matter

#

the game is most likely less precise than the imprecision sev is trying to remove

median heath
#

If you choose to see things in degrees of severity, then the 45-81 rule is not for you.

wind spade
#

more severe is imo the (most likely) u7 bug where mk5 belts only reach 740-750 😛

#

that gives a lot more imprecision

golden smelt
wind spade
golden smelt
#

Damn

#

So the rule doesn't matter that much cause I'll never really see that imprecision

median heath
#

Thank you greeny.
For convincing people that hours of work doesn't matter at all 👍

wind spade
#

the imprecision is around 1 item per several hours (obviously depends on recipes and stuff, but it's definitely not something big)

#

I'm not saying it doesn't matter. I'm just saying that it's not the place where people want to spend time in most cases

#

spending a few extra hours just go get system to be 0.0001% more efficient is not really worth (in my book and in many others')

oblique notch
#

To you? You decide.

median heath
wind spade
#

the extra hours come from you having to figure out numbers that fit in the rule when planning

#

(you as the player)

golden smelt
#

Now that I know that my factory isn't producing as efficient as possible, it makes me kinda mad

wind spade
#

you can't just maximise production from nodes anymore

median heath
oblique notch
#

Or you just overclock/underclock everything to thr same number and walk away, 10 machines feeding into 10 machines feeding into 10 more :p

wind spade
wind spade
#

and we both now how many people solve from beginning or maximise nodes

median heath
#

This is genuinely coming across to me as a giant shitfest on the work that was done and how I choose to do the math.
I'm not sure if that's your intent or not so I'm just disengaging before I get any more depressed.
Bye.

wind spade
#

this is not aimed against you in any way. This is just aimed towards turtletrademark, to tell them that it's not as big deal as you make it sound (if they don't care about 0.00001% inefficiencies)

golden smelt
#

That ended pretty badly

wind spade
#

Sev being Sev, we're used to it

median heath
golden smelt
wind spade
#

for people that read your messages "inefficient factory" sounds like a big deal, "inefficient by 0.00001%" suddenly doesn't sound like a big deal. since I do know you, I'm very much aware of what you mean, but new people aren't.

wind spade
golden smelt
#

Makes sense

#

I've definitely learned a lot today

#

I do have one more question but I'll save that for later

tawdry sail
wind spade
#

I think if you scroll up, Sev just described the rule

tawdry sail
#

Oh. big text

wind spade
#

if it's too much text I can give you tl;dr

tawdry sail
#

Nono is fine

#

Thanks

wind spade
#

basically since you can only clock in increments of 0.0001%, some final production numbers are not reachable. So if you produce everything in multiplies of 45 (or 81 for oil products), you'll hit the "correct" numbers that you can hit with clock speed. With other numbers you can be off by 0.0000something per minute and for those that care about this, the 45-81 rule is great

tawdry sail
#

anyone ever made some calculation of how much this 0.0001% affect the production in a simple scenario? (in, idk, 10 hours of playing)

wind spade
#

it obviously depends exactly on a situation

wind spade
# tawdry sail anyone ever made some calculation of how much this 0.0001% affect the production...

but let's do some example... let's say Heavy Encased Frame (alt) at 2/min production, you need 0.711111111... machines. you have two choices of a clock speed: 71.1111% or 71.1112%. So, you'll either get 1.9999996875 per minute or 2.0000025 per minute. Over 10 hours that equals to 0.0001875 less or 0.0015 more modular frames.

The example was just randomly picked, I have no idea if it's smaller or bigger than average, but I'm gonna say there won't be cases where the difference is orders of magnitude bigger

#

(and also keep in mind that if you have exactly the correct amount of resources, picking the higher clock speed will result of the machine stopping for a few seconds every few hours, but you'll still get 2 plates per minute on average)

cinder silo
#

Throw in belts doing whatever they please and mergers doing the stutter step production can easily be knocked off kilter.

tawdry sail
#

Interesting how the decimal point affect that..
Honestly, i think that caring of this while playing is a little stupid.

wind spade
#

yeah this is level of precision is way more than we can currently get from the game

#

but again, the rule is mostly for people that want theoretical 100%

cinder silo
#

A couple of us have been running some serious experiments (mostly because we have nothing better to do 🤣 ) and uncovered some major flaws.

wind spade
cinder silo
#

The mk5 thing seems to be exacerbating the merger stutter step that was just a mild occasional blip in the past.

wind spade
#

well hopefully it's just a small bug that can be easily fixed 🙂

#

or it's a step towards halving the game

#

is that just mk5s btw?

cinder silo
#

Fingers crossed, seeing the experiments play out the way they have has been a huge eye opener.

#

High (750ish) throughput from a merger in to a mk5 seems to show the big stutters that cause backups and experiment rejects, I haven't seen it when testing lower mark belts yet but I haven't adjusted the miner based test yet.

wind spade
#

wait I thought it's also a case of miner -> mk5 -> splitter

cinder silo
#

This shown in the merger test fun thread shows the effects of spiky throughput with an average of only 311 parts per minute, the number of objects thrown should be 0, but there are loads after just four and a half hours.

#

I set this little experiment up on my live warehouse in feed from the north after seeing stalls on the merger.

#

Splitters have been pretty well behaved, there is a tiny lag when the merger decides what order the mixed feed will come out based on inputs, before the mk5 situation worsened it only really stood out when a line was oversaturated and backed up.

wind spade
#

wait, so miner -> single segment mk5 -> splitter can still do 780 (ish)?

#

it's just "don't merge over 720" and "don't do maxed belts over more than one segment" then?

cinder silo
#

The mk5 single section is debatable now, I posted that experiment under belt to belt, and that throws things too.

wind spade
#

well I guess it's the only place where you really don't have a choice 😄

cinder silo
#

I'm glad my bauxite & aluminium plants only do 720 and rubber/plastics are 600 per line because it would be a disaster.

wind spade
#

how severe is the single segment thing? is it up to 10 or more?

cinder silo
#

One moment please, I restarted my single section experiment a couple of hours ago, I'll grab its data.

#

Welp, in the space of two hours, the single section loop threw 21 pieces.

Screenshot_1049.png
wind spade
#

that's 0.00224% loss 🤔

#

or 779.825/min throughput

#

I'd say that's acceptable for "780ish" definition 🙂

cinder silo
#

Before update 7 though the single section experiment (running for months basically) threw 0

wind spade
#

yeah that's fair

cinder silo
wind spade
#

I know you're doing tons of those, but would be interesting to see them with mk1-4s so we know if it's the belt system or just mk5 belt being too fast

cinder silo
#

I'll need to expand the test area.

#

I do have a mk1 loop up, that hasn't thrown anything at all, the mk3 though threw 5 pieces but it took 36h.

wind spade
#

so I guess it's all belts, just relative to it's max speed

cinder silo
#

Gets way worse the faster the belt is.

#

It isn't a problem that can really be solved while overclocking is a thing, you have fractional outputs, like 4 decimal places that won't quite hit the mark and it rounds down, that tiny miss just accumulates in little errors that cause the throughput ripples that lead to the dreaded yellow lights.

past wagon
#

So I'm just looking into Nuclear; I've got a calculator for 28.8 Uranium Fuel Rods/m, which feeds 144 Reactors and making 1440 Waste/min..
And it looks like 1440 Uranium Waste/m can feed 10.8 Plutonium Rods/min (with default Non-Fissile) with 100% Waste consumption

Its not the most perfect outputs possible I know, but it has no waste and looks like 360,000 GW of power which should be excellent. Any issues with a plan like this for a first try? thinking_helmet

versed violet
cinder silo
#

10.8 plutonium rods from 1440 waste? I use default non fissile and get 5.25 plutonium rods from 1050 waste.

past wagon
versed violet
#

the second default recipe (plutonium pellets) also uses some waste to reduce processing.

past wagon
cinder silo
#

Incomplete there, the ratio of waste seems to show you're using the same reprocessing just with like 40% more waste which is why I'm not seeing how you can get double the plutonium rods.

versed violet
#

maybe alt recipe for plutonium rods is at play?

cinder silo
past wagon
#

The actual flowchart itself is very scary, really - but looks like 216 Encased Plutonium Cell goes up, and indeed goes to the Alt: Plutonium Fuel Unit

image.png image.png
cinder silo
versed violet
cinder silo
#

I only have a part of my old nuclear plan still since all the stored plans are on a pc that was stolen back in march 😦 , from back when I tried using SCIM as a production planner.

Stage_2_Nuclear.png
versed violet
cinder silo
past wagon
cinder silo
#

The goal of getting shot of waste is to burn as much as possible, the defaults are great at that.

versed violet
#

If your goal is to get rid of uranium waste, just use al default recipes to get to sinkable pluto rods with least headaches. You can also use them to fuel trucks, just make sure the truck loop is far away from where you want to walk.

past wagon
#

I've got a megaprint Train system ready to feed the area of the map I settle on, so I probably will just sink them after all

versed violet
#

sinking is good, unless you're a Titanic.

cinder silo
#

For really small nuclear loops you can get away with burning the plutonium since its waste accumulates slowly, for the most part unless you want a truly colossal amount of power, sinking the plut rods is generally the best way.

versed violet
#

if you plan on playing on said save for a long time, accumulating plutonium waste p[roblem is not worth it.

cinder silo
#

Unless you have 50-100 reactors burning plutonium though, you'd have to accumulate thousands of hours for the nuke tomb to become a problem.

#

That said I still sink it all, 100 reactors burning uranium fuel at 105% (for load balancer to work) is more than enough for me 😄

golden smelt
#

Is doing this stuff necessary for my factory run properly or am I waisting time.

IMG_20221128_101834704.jpg
#

this is me trying to take a 100 input and turn it into five outputs of 20

#

I know it isnt exact cause I didnt worry about the decimals

#

but would it be better to just make a conveyor line and split at certain points

past wagon
golden smelt
past wagon
# golden smelt yes

Then yeah you don't "have to" balance, a Manifold would work

Sure the first machine would take 60/min instead of 20/min, but once it fills up then all the overflow moves down the line, and so on and so on - eventually they all balance out and just take 20/min off the line each

image.png
crimson helm
golden smelt
past wagon
golden smelt
#

my mind cant handle the bus anymore

#

ima go back to manifold

tawdry sail
#

embrace manifold

dense cave
#

ya, manifold is the way in this game

#

I tried making balancers when I first started, its a giant waste of time

#

I mean the only time I do something even remotely like this is when I have 2 lines and I make it so if one is backed up it backfeeds into the otherone

vapid gorge
dense cave
#

well, its just when I have 2 production lines for say rods. More than a single belt can carry, so I have 2 outputs. I then use splitters and mergers in case one line isn't being used to the max, it overflows to the other line

vapid gorge
#

... just from words that sounds over engineered but I suspect it's not a complete enough description

dense cave
#

nah, otherwise I may have shortage on one line while the other sits idle

#

its a simple fix

vapid gorge
#

why not just clock the machines to use what is being provided though?\

dense cave
vapid gorge
#

right so you aren't making dedicated worth paths, got it

dense cave
#

ya

#

especially as I built stuff as I was teching, so I knew my upstream would eventually have to be upgraded

vapid gorge
#

fair enough 🙂

dense cave
#

In fact when I built my lines off say a node, I didn't just build it to say consume 270 ore

#

I made it for twice that

#

So if I had one making plates, and another making rods, each consuming a separate mine. I made them each for twice as much, so if one was backed up the ore would overflow into the other line

vapid gorge
#

I don't sink much stuff early on so I just let items back up and pause production, or splits items willy nilly to places not caring if enoug his getting everywhere because eventually as they fill they'll shut down.

dense cave
#

yeah I don't sink stuff either, thats why some stuff will back up and others wont

#

just a different way of building

#

had one dude earlier telling me foo foo about the way I made it due to lag (centralized hub with a dual bus). But I never had any issues with lag, then again I got a fast pc

#

build as you want, everyone has a different way

main mango
#

Anyone in here think they may be able to solve my fuel gen problems? lol

vapid gorge
#

Make a thread in questions and post LOTS of pictures of hte set up. Can't trouble shoot pipes w/o pics

tawdry sail
#

That is called balancer. In @dense cave example, a 2 line balancer. Very useful.

vapid gorge
lament jolt
#

So I'm planning a factory dedicated to Assembly Director Systems...with the alt recipes I have unlocked on this play through, my target of 18.75 per minute is calling for over 2500 buildings....does that sound right?

vapid gorge
#

That sounds like a lot. are you using tools?

#

oh wow a lot of that is dedicated to constructors with cable

#

yeah I'm seeing 3500 with base recipes

lament jolt
#

Eep...

vapid gorge
#

It's a pretty involved system though. tons of wire, hmf, supcomps.

lament jolt
#

I just don't want to sink a ton of stuff to get tickets for all the sub parts and bypass the production chain. 4000 for the employee of the planet.....then the next step of 4000 magnet field...

vapid gorge
#

You could always just make a system taht does like... 2 pm and you'll get the number you need pretty quickly

lament jolt
#

2/minute would take 33.34 hours

vapid gorge
#

how long will it take you to build the rest of the phase 4 space parts?

#

and plan it?

lament jolt
#

That part is done.... I've played over 1700 hours on several saves combined....so I helped myself for those. But I want to legit grind for the employee of the planet

vapid gorge
#

A lot of people just don't bother with actually autoimating space parts. They just smash the end products together to grind it away

#

If that's what you want 🙂

prime orchid
#

should i drop this here i guess?

#

train spiral info

vapid gorge
#

why 1m x 8m ?

prime orchid
#

ramp slopes

wind spade
#

Spirals 🤢

vapid gorge
#

1:8 is extremely shallow. 1:4 is mostly good for just about any train I think

#

while 4:8 is probably too steep for most

prime orchid
#

this vertical distance is for each quarter section of the spiral. you can also use it for any kind of radial turn you want with a vertical climb

image.png
#

yup already tested and removed that section

prime orchid
#

was gearing up for a cheat sheet that included half helix turns but it seems like the game doesn't like them

#

trying to do incorrect shape. but uh oh that means the half helixes might not even be proper.. hm

image.png
fierce cypress
#

the radius just isn't large enough

wind spade
#

imo train spirals are only good if you really want a train spiral from visual perspective 😄

prime orchid
#

we might have to hope for different modes for rails. like one for shortest distance which includes tightest possible curves, and one that follows the full curve of the tracks it's connecting

frosty owl
#

(Or adapt to how the rails currently work rolljace)

prime orchid
#

because this should work, it's just fighting itself to pull in closer to the two connection points instead of going out further to make one big curve

fierce cypress
#

you can get 180 degree curves tho can't you?

#

maybe the height difference is just stuffing with it or something

prime orchid
prime orchid
frosty owl
prime orchid
fierce cypress
#

must be misremembering then, i thought you could, although thinking back i might've done multiple smaller curves like ven said

noble timber
#

Yeah you can only do a 180 with two 90’s (Or more of less angle)

prime orchid
#

so this is why so many connections that should work, aren't working. it needs to be smart enough to know how to only curve as much as it's allowed instead of just saying it's impossible

noble timber
#

Physically, that would be too tight a turn for a train to be able to complete

#

You need a bigger turn radius (hence the two 90 degree tracks)

fierce cypress
#

i think what they're saying is it would be nice if the rail snapping forced the radius

#

to make curves that always have a good curve

noble timber
#

That radius will always remain the same unless you move the rails further apart from each other

fierce cypress
noble timber
#

I don’t understand what you mean then

#

A 180 turn with a two foundation gap is impossible

fierce cypress
#

the radius is large enough, the rail just tries to 'cut the corner'

prime orchid
#

only because of these tight curves it's forcing itself to use instead of going out further

image.png
noble timber
#

Oh I see what you mean now

prime orchid
#

well hang on let me make sure

noble timber
#

Although I still think that turn wouldn’t be possible?

fierce cypress
#

that curve is possible with two pieces, but building it in one doesn't work

#

at least i think it does simon_smile

noble timber
#

It’s been so long since I’ve built a track that I’m now doubting myself 😂

fierce cypress
#

according to wiki this is the tightest possible 180 degree turn, so maybe it is too tight

unknown.png
noble timber
#

Yeah

frosty owl
#

Tbf, I don't think belts act too differently, even with their additional snapping options comlared to rails.
For tight turns, there are situations in which one needs to place 2 belt segments where one would have sufficed but the belt was like "too sharp turn" (it it makes the same turn if forced through 2 segments)

noble timber
#

So would you be able to build that in one piece I wonder?

fierce cypress
noble timber
#

I’ve actually never tried to build a 180 in one track. Just programmed to build 2 90’s

frosty owl
fierce cypress
noble timber
prime orchid
#

yeah even here it's doing the same thing

noble timber
#

Oh that is 6 😂

fierce cypress
#

shouldn't it be 4.25 minimum for 180?

prime orchid
fierce cypress
#

if the wiki is correct in saying 17m is the minimum 90 degree radius

prime orchid
#

works with quarters but look how much farter out it is than when it's trying to do 180

frosty owl
fierce cypress
#

yea, its trying to do 1.5 foundations, whilst it should be 3 foundations

fierce cypress
prime orchid
#

connecting to the same points that were anchored straight, you can see the problem

#

very obvious

#

and my guess is that even the quarter turns are having the same behavior, just as belts would

#

like this

#

just less obvious

fierce cypress
#

depending on build order maybe?

#

but 90 degree turns are pretty reliable

prime orchid
#

you know how we have to force belts to make 90 degree turns like this, maybe there's also an optimal distance to force rails into a "perfect 90 degree turn"

image.png
fierce cypress
prime orchid
#

oh?

#

is 1.125 the distance diagonally?

fierce cypress
#

the radius, and 2.125 - not 1.125

prime orchid
#

oh right ok

#

not working for me

image.png
fierce cypress
wind spade
prime orchid
#

yes but then your snap points aren't straight

fierce cypress
prime orchid
#

you can build it without the anchors but then you wind up with this

image.png
#

so doesn't seem like it

fierce cypress
prime orchid
#

maybe it's not a perfect 90 degree after all, which would seem obvious when trying to snap them to anchors in the first place

fierce cypress
#

it is a perfect 90 degree turn, its just got weird connection requirements

#

probably a bug(?)

prime orchid
#

nah i think it's just the same exact issue i was pointing out above. imperfect curves and not being able to snap ends rotationally

fierce cypress
#

its a build order issue

prime orchid
#

because of....

prime orchid
fierce cypress
#

as i said before, it doesn't work if you are trying to connect to to snap points

#

which is what you are doing

prime orchid
#

because of the issue i pointed out above. if that wasn't an issue then you wouldn't need a specific and imperfect build order

fierce cypress
#

how is this different to this? - they are the same thing

unknown.png unknown.png
prime orchid
#

and thus you would be able to complete the circle

fierce cypress
prime orchid
#

nope they actually aren't because i removed the first quarter of the circle that was snapped to the anchor in the first point. so none of the connections are straight

fierce cypress
prime orchid
#

i wonder if the reason is the one i have been mentioning

fierce cypress
#

again, it is a perfect 90 degree turn that just bugs when you try and build it with that build order

fierce cypress
prime orchid
#

which is still relevant to the quarter segments

prime orchid
fierce cypress
#

well no, because perfect 90 degree segments are possible, perfect 180 degree segments are not

prime orchid
#

the evidence is stacked against you

#

i disagree

fierce cypress
#

how can you disagree - you can go in game and prove it yourself, hell you just did

#

what is there to disagree with?

prime orchid
#

opposite

#

i just proved everything i've been saying idk how you're trying to argue against it when it's obvious

fierce cypress
#

the 180 problem is different to the circle problem

prime orchid
#

same problem

fierce cypress
#

they.... aren't

prime orchid
#

ok 🤷‍♂️

fierce cypress
#

because as i said, you can build perfect 90 degree segments

#

so i don't see what your 'problem' is

#

the fact that you can't make a full circle with them is due to a build order bug

#

not because the 90 degree segments aren't perfect

prime orchid
#

you think you can build perfect 90 degree segments but in reality it is slightly imperfect but it's ok if you want to believe that it's perfect at least i know it's not and that's all that matters tbh

fierce cypress
#

prove to me that they are imperfect in some other way than "you can't make a circle with them"

prime orchid
#

i already did i don't know how else to help you buddy

fierce cypress
prime orchid
#

go back and re-read through the last hour of this channel

#

i could also prove to you through belts if you want

fierce cypress
#

i have been actively part of the discussion, and i cannot see any other 'proof', so please, point it out to me

noble timber
fierce cypress
#

who would've thought

prime orchid
#

same snap points, one is rotationally locked to 90 degrees (such as what we do when we anchor with rails, by god, i compared them gasp)

image.png
fierce cypress
#

yet you can visually see this is not the case with rails

prime orchid
#

so you don't think it's at all possible that the difference is minor enough that it's not visually clear but still enough to pose a problem?

fierce cypress
#

no, but you can't seem to prove it

prime orchid
#

maybe when we get rotational snap points for rails then you might get the proof you need but i already have enough proof

fierce cypress
#

i already have enough proof

being what?

prime orchid
#

lol you need help dude

#

i'm done trying

fierce cypress
#

sorry for trying to help you understand

frosty owl
#

Reminder that tone is hard to transmit or understand properly through text

fierce cypress
#

if you want some further clarification - the wiki has some pretty self explanatory explanations of how the rail curves and snapping work

#

also i feel like if all the curves in the game were imprecise, someone would have noticed before now

#

or you could simply prove that they aren't in any matter of ways

#

ironically one being a train spiral, which allows you to 'make a circle' except with vertical distance to avoid the build order bug

#

and wow, would you look at that - it lines up perfectly

#

anyway simon_smile i should probably sleep

prime orchid
#

build order is only half the problem, when snapping between two connections it exhibits the issue i mentioned above which is the other half of the problem, and when building free hand it seems to curve properly. this isn't necessarily proof that the snap points at the ends of free hand builds are straight, but would lend itself to favoring that assumption. and not being able to finish a circle is because of the same issue, and not necessarily because of imperfect connections, still open to proof

image.png
#

however when trying to freehand a 180, the same issue arises as when trying to snap between two connections, the proper curving seems to stop vaguely around 140-145 degrees

thick plank
# prime orchid build order is only half the problem, when snapping between two connections it e...

I would wish to add to your argumentation that there is a difference netween two 45 degrees turns and a 90 degrees turn that both cover the exact same distance (aka you build half of the 90 degree turn and then directly the other half).

Sometimes you cant directly build the full 90 degrees and instead have to split it down in two equal parts. Eventhough i couldnt ever figuere out when thats the case.

thick plank
prime orchid
thick plank
# fierce cypress

You should use the second option, just split the 3 in two halves and build them. Thats better and more precise

thick plank
prime orchid
# thick plank WHAT?

sorry, not radius, what would it be called radial arc? no i don't think that's right either. idk anyway you get what i'm saying right

#

they all have the same center point

thick plank
prime orchid
#

so their curves should all align

#

if i free hand that middle one, they align

thick plank
prime orchid
#

yeah it seems to in the case of connecting two points causing the rail to path differently than free-handing it

thick plank
prime orchid
#

so the general thought process of always wanting to anchor each end is actually not always good

prime orchid
thick plank
#

You dont always turn at the same rate

#

Its always dependent on the raidius

#

*radius

prime orchid
#

it should only be dependent on the center point

#

same center point, same curve

image.png
cinder silo
#

I don't know if its perspective but aren't two of those on the top left a bit close?

prime orchid
#

no i think it's perspective or fov they are perfect when checking up close

#

i changed the outer one to a connection between two anchors and this is how bad it is

image.png
#

halfway through the rail next to it

#

it is still possible that the free hand curves aren't perfect but they're clearly better than anchoring both sides

wicked tinsel
#

rail builder had always had this bug

#

it interprets point-point and point-free build modes differently

prime orchid
#

so it's looking like even the free hand curves aren't perfect after all, let me triple check

wicked tinsel
#

i have no idea what kind of spaghetti they cooked for it to be broken like this

#

its the reason why you cant make 3x3 curve connection and its a reason why you often cant rebuild deconstructed rails

#

probably one of worst bugs in railway maker, dunno if it affects other stuff like belts and pipes

#

tho those behave in weird ways too, so its quite possible

prime orchid
#

yeah it looks like even the freehand curves are too shallow and not using a proper center point to curve around

wicked tinsel
#

if you build straight rail here, decon the curve and try to rebuild it, it will either claim that its too sharp or will rebuild differently shaped rail

obraz.png
#

as for free hand curves, no clue, its possible they are also wrong indeed

#

but given super low precision of everything here, its not super visible

prime orchid
#

yeah but that's only on the inner one because it is tighter than the others, but yeah it's due to the same issue i've been mentioning

#

no it's precise i triple checked my placements. no weird numbers or anything, only variable is curve

#

it's a huge difference

image.png
prime orchid
wicked tinsel
#

yeah, they seem to be wrong

frozen tapir
#

10 computers PM + 5 circuit boards PM leaves 290 HOR which makes just enough power to run the factory self sufficiently. has to use the battery storage when i use a hover pack lol

image.png
prime orchid
#

rails seem to be a pretty decent gripe among the community so i assume they will get fixed eventually

barren elm
#

It annoys the hell out of me that some lengths of rail have a little gap between the segments

snow dove
#

or when you build rails on ramps and it does the little wavy thing

prime orchid
barren elm
#

Always build from existing rails to a new point, never build from a new point to existing rail

#

Your rails will always be pixel perfect straight

#

(new point to new point is also fine)

snow dove
#

I do existing to new point and it still is wavy

barren elm
#

Your existing rail is wonky then

#

It would, however, be nice if rails just snapped to nice rotations with some settings or something though, shouldn't have to go through this nonsense

prime orchid
barren elm
#

No, I don't use trail spirals, I think they look silly

#

Just ramps

#

And ramps follow the same "no waviness if building from existing 'straight' rail" logic

prime orchid
#

Oh yeah it’s way less complicated when building on straight ramps as long as you don’t start with flat rail. Gotta put an anchor on the ramp first

oblique notch
# wicked tinsel i have no idea what kind of spaghetti they cooked for it to be broken like this

It's splines. It's... just weird math. It doesn't really matter how clean their code is. It's trying to fit curves of any degree and arc length onto a square grid. It doesn't line-up for every radius - it's always going to be slightly off because it's a math thing.

Yes your radius at both ends of your 90* arc is X meters. For it to br a perfect curve it also needs to be X m from center to any point on that curve, as you no doubt know

But the game world is built on a square grid. Even though it's a basic x,y,z coordinate system there is no way to make a perfect circle on a square grid - after all, pi is never ending and you need pi to discover the length of any part of the circle. You pick any spot other than exactly left/right or up/down and you are going to find the length of the radii is going tonalways need more decimal points to be "exact' to make a good looking curve.

So you have to make an approximation. You can get super fine detailed That the human eye has trouble picking out the flaws but the human brain is really good at finding them if you stare long enough.

#

And that's a spline in essence - a mathematical formula to describe in Cartesian points as much as possible a curve. But you have to make sacrifices - you can make it super fine detailed and as close to perfect a system that can only really deal with powers of 2 (binary) can do... but that takes more time and processing. So you make sacrifices in order to ensure good game play.

barren elm
#

But the game world is built on a square grid.
It isn't

#

The global grid is a lie, it's more of a "snap to nearest 8 meters"

#

Everything you place in the world has a float coordinate that can be very fine

oblique notch
#

But
... still is. It's built on a Cartesian points (x,y,z) which... effectively makes a square gird just of squares of .000001 sides

wicked tinsel
cinder silo
#

Still makes circles a swine, how @vapid gorge has the patience for his rings within rings I'll never know.

wicked tinsel
#

you got one end fixed by previous rail and new one is generated based on where the rail ends (where you point your mouse)

oblique notch
wicked tinsel
#

so it defines start x,y,z+rotation & destination x,y,z+rotation and then fits a spline between them

#

however, for some arcane reason

vapid gorge
cinder silo
wicked tinsel
#

the reuslt is different depending on whenever end is dynamically computed based on mouse position or if its attached to something pre-existing

#

which is completely baffling, as it would mean they have two separate spline generation algorithms

#

thus the spaghetti

oblique notch
wicked tinsel
#

no

#

its same, thats the problem

#

but the result is different somehow

#

lemme paint it really quickly

oblique notch
#

I doubt it's the same. Undoubtedly the snap point is slightly higher z or has less granularity than your mouse look ray to find a point and sonthey are starting with different end numbers resulting in different splines.

Add into it it really looks to me that it actually divides the splines in large about 5 m chunks that is going to exasperate any small difference in values for the calculation

(I say about 5m because of the way the top and bottom of ramps curve out, meaning it is probably looking every 5 m or so for a major change in z to smooth out the ramp)

wicked tinsel
#

its like this

obraz.png
#

the problem is, despite the fact that red and green rails have same begin/end points, they get different shape

#

despite the fact that logically, only coordinates of begin/end should be an input for the curve generation algo

#

in other words, if you build three rails in row, scrap middle and rebuild it, you will get different result

#

which is completely wrong

oblique notch
#

And my point is I'm fairly certain snapping and mouse look are not providing same end coordinates

wicked tinsel
#

its possible that its broken in that way, but

#

logically, both should have same coords

#

if game is bugged and it provides different coordinates for both cases, then yeah, it would explain it

wind spade
#

but in first case there's no end rail to provide it's rotation, 🤔

oblique notch
#

Combined with what is obviously larger than fine grained float coordinates (the chunks to smooth out height changes) and you will end up with any small deviance growing larger

wicked tinsel
#

so fundamentally, rotation is still same

#

tldr: build three rails in row, in a way that middle one is curved, scrap middle one and rebuild it

#

the result will be different and it should not be

#

whenever the coordinates are offset weirdly or some other issue occurs isnt really relevant, the bug is still there

oblique notch
#

Then there is probably some end point clamping or manipulation when snapping to that is different or doesn't occur when snapping from

wicked tinsel
#

sure, its possible

oblique notch
#

Ie when you remove that middle piece and rebuild it, when it snaps to the 3rd piece it must use slightly different coorda

wicked tinsel
#

but still, by design, it should do the clamping action on the virtual destination too

#

so that result is consistent

oblique notch
#

Yes that would be a bug😅

wicked tinsel
#

its possible they fucked up some quaternion, there is no normalization or some other issue occurs

#

but all those would be just a bug

oblique notch
#

I doubt it - it's probably the same spline generation abstracted and across the entire game, just weird interactions on the way the values are given to it

wicked tinsel
#

ultimately, it means that some stuff cant be build, or build order is super important, or scrapping something accidentally is catastrophic unrecoverable action

#

all of which are bad 😦

#

well, if you ask me, i would throw entire rail builder to trash can and start from scratch at this point

oblique notch
#

thats rather extreme without seeing any of the code.

wicked tinsel
#

at least stuff like being able to cut rail arbitrary should had been there

oblique notch
#

given the experience i have (external, small, and mostly with discussion with just one or two CSS devs) their code is rather abstracted and well encapsulated. So, i agree there is a bug - there is something in there that needs to be redone, but probably lots of it is just fine. But there is some place that is not lining up and not doing the right things.

wicked tinsel
#

but i doubt its something that can be added at this point without major rewrite

#

its missing some super basic features like guidelines too

oblique notch
#

because it already has segements, there just is not any exposure of it

wicked tinsel
#

i also wish it was possible to somewhat free-form the end without building&scrapping guidelines :\

#

especially since its virtually impossible to tell what is "too sharp" for engine

#

and that aside from the fact that "too sharp"'ness calcuiation seem to be bugged too

#

it could, for example, display ghost in last viable position when you go too-sharp

#

there is no feedback on rail steepness either, despite the fact that its critical feature for gameplay

#

😔

oblique notch
#

it would be nice for an implementation of a ctrl snapping for curves

wicked tinsel
oblique hollow
#

Nothing is ever really unrelated

#

Backlogs are backlogs and i dont doubt they know of rail issues

#

but if anything, they will do thing as always: in one go

#

possibly when some spline optimization happens

wicked tinsel
#

nah, its not this game experience at all

#

its more like #patch-notes message "we fixed bunch of random minor stuff"

and it somehow broke vehicles completely

oblique hollow
#

or just a revisit of build mechanics

oblique hollow
#

people reported stuff still getting delivered

wicked tinsel
#

wasnt it broken on the physical range and working on the teleportation range?

oblique hollow
#

"teleportation" is gone

#

its more like non-phys now

#

since they move along spline curves now

wicked tinsel
#

name is as you like, effect is same

#

teleportation/warping/flying along spline

#

still, its not the first time such issue had occured

#

its usually minor patch with few minor stuff and something major explodes

#

which certainly puts doubt in the "abstracted and well encapsulated" statement 😄

oblique hollow
#

their QA couldnt catch everything as Ben said

#

EX is there for them to throw stuff out and see what works

#

in effect, WE are the QA

#

we are just unpaid CSS employees at this point jace_smile

#

Frankly, i wouldn't pay that much attention to quality either in EX.

Get all the features, fix major bugs when they happen, polish it up, move to EA

oblique notch
wicked tinsel
#

it might be perfectly encapsulated but ultimately, it behaves like big ball of mud architecture 😅

oblique notch
#

So it may just be my standards for "well done" are very low because of the crap I fix at work 🙄

oblique notch
wicked tinsel
#

might explain why their simulation is so jank 😄

oblique notch
#

Of course. They just need to offload it to the cloud! Doesn't Ficsit know proper Tech Uplift?

prime orchid
# oblique notch Then there is probably some end point clamping or manipulation when snapping to ...

i don't think that's quite the issue. it seems clear that there are different rules for track shape when connecting two different connections and snapping to an open ended point. for example when trying to snap to an open ended point the rules might be, "try to make a nice big curve to wherever the new point is" and when connecting two points it's like "try to connect these from point a to b, but give it some decent curves so it doesn't look like horrible sharp angles"
i think part of the reason the first rule works that way is because it doesn't have any connection angle info. so it's just maintaining the one from the other end. but when connecting two points, they could be different angles so it needs to be able to have separate rules to adjust for that

oblique notch
#

i could see that.

#

that would still be related to what i posited, but its not just origin-> endpoint coordinates, but rather other info such angle of interaction, distance, ect. some of that info is null/defaulted when not snapping to another rail, and some of it is provided when doing so.

#

if i were the one writing that code, it would still be the same spline/output calculation, just that in snapping/not snapping some data would be defaulted

#

and that would of course mean a different output

prime orchid
#

yes and unfortunately from my testing even the better curves without connecting two points still produces imperfect curves. they are straighter than they should be. and i don't think the reason for this is any kind of game grid 1's and 0's nonsense. we've been making perfect circles on computers for a very long time. it's not hard if you have two points with 2 trajectories and maintaining and arc around the calculated center point

#

i think the code just doesn't do what it "should" by some of our standards

#

but judging by the general consensus from the community that rails need work, hopefully they will receive some tlc in the future

oblique notch
#

(and to the naked eye at a reasonable distance, they look perfect enough)

#

but regardless of that - the track im fairly certain is in chunks - so that it can do the round off on the hills.

#

and 5 or 8 m chunks will defintely cause weird points in any curve, especialyl if its doing something like generating the curve for each chunk rather than as whole

prime orchid
#

let me fiddle with that real quick

#

hm i actually don't think it's doing that. it doesn't seem to follow hills. it goes straight through them. however it can kind of mimic following hills a tiny bit just from gathering data from the slope of the 2 endpoints. it has the same behavior as it does with placing on ramps. the end points are forced into the slope of the ramp angles. the only other thing that changes is height and distance. when a short length is used over a hill, the curves produced by the end point slopes is pretty effective, but over longer distances they don't do much for the middle of the rail that is mostly straight. sort of the same reason we get rollercoaster spirals

#

i bet i can find a pretty good example in game to capture

#

both ends have mostly flat trajectories so it goes straight through this hill. it's a pretty long rail and doesn't seem to get any info in the middle from the hill it's going through.

image.png
oblique notch
#

that doesnt happen instantly, that happens over a distance - and thats the chunk distancte

prime orchid
#

heres a rail in the same hill but each end is on the sloped area. it's just trying to smooth out the differences of each slope and not actually trying to follow the hill

image.png
versed violet
#

it does not follow hill. It bends according to what start and endpoints are bent too. If you put them on foundation, it goes flat, if you put it on ramp, it goes up.

prime orchid
#

yea look at this , ||can't place on 4m ramps since slope is too steep but|| look how drastic it is, from two open end points on ramps

image.png
versed violet
#

try making a quarter circle with 3x3 foundation radius and check if it keeps the radius or deviates either way

#

regarding the 'what kind of curve rails follow'

oblique notch
#

fact that there are two different direction curves - one from level to slope, and another from slope to level - indicates there are "sections" to this rail (at least). Each section with its own start and end (as a single spline calculation cannot do multiple curves by itself, not in a non wave like way, these are individual splines fused together into one)

#

and if i'm correct, its these mini chunks that are fusing into a single path that cause the weird ness - wherever their endpoint is internally to the line, and the calculation is not smoothing across it

versed violet
#

so it basically behaves like a rubber baton?

oblique notch
#

so there are points where it just will bend somewhat unnaturally, where as the parts in between are fine to curve, when it reaches one of the segments it may be a bit unnatural

#

and given the texture on the rails, it wouldnt surprise me if they are 8 m maximum segements (1 foundation).

So it generates the curve for the first, then the next, and the next and the next and so on, using the end point of one to be the start point of the next. But that does not necessarily make a smooth curve - the curve radius would need to be adjusted slightly and i bet that adjustment needs... a bit adjusting :p

image.png
#

so, with that in mind. This was built with the typical method of curves - 4x4 foundation L, track at start and track at end.

image.png
#

and we can see, if my guess on size is correct based on the texture, 6 segements of curve

versed violet
#

can you put a diagonal founcation with the ctrl-method to see the deviation?

oblique notch
#

maybe. wish i had mm working lol

versed violet
#

don't need mm to measure that, just vanilla rotation

#

I always get impression the curve is a bit flatter than it should be.

oblique notch
#

it does seem to be. can only really easily get 4 in there:

image.png
#

which would indicate that whatever the spline calculation is it starts perhaps from an angle rather than a flat surface

#

you can almost see what im talking about here: the first bit (between red adn blue) doesnt seem to curve enough, then suddenly starts curving more

#

im guessing the first segment of a curve created stays going straight instead of starting the curve right away

#

which, when a human looks at it we want the curve to start from the red line, where we started the damn thing. But because it goes straight for a little bit more? its actually creating a curve from a slightly smaller point

versed violet
#

could you start the curve from middle of foundation, then put a diagonal row of foundations from the focal point, to compare how it deviates?

oblique notch
#

which would also explain why a track built Straight -> Curve -> Straight in that order can produce a different curve when the curve part is removed and replaced, because its also doing the same thing at the end: It creates a small straight path when it ends (either when snapping or when not, im not sure which)

oblique notch
versed violet
prime orchid
#

it could be that it is segmented and each segment's curve is generated from the segments in front and behind it, so near the middle the segment changes would be less drastic than near the connection points where it's forced to be a certain angle

#

idk if i'm on board with the segment theory though

oblique notch
#

in order to take into account the curves when it changes in height it doesnt have a choice though. You can't generate a single spline that curves in multiple directions that way - it is always curve to a point, then another curve generation the other direction ...

assuming i understand the math right. lol.

But im 90% certain that the math of an S curve (like the height change) would be basically (x[splinecurvemath]y) + (y[-splinecurvemath]z) where x = the starting point and z equals the ending point, and y is some arbitrary point in between the two that defines where the spline changes direction.

Since thats the math, it means that the fact we can generate an S curve of any [height within clamped bounds] and it dynamically is factoring in these features, it is at the very least creating two segments in the math. This would have an effect on straight horziontal curves as well.

versed violet
#

Maybe Mr. Bezier could explain it better?

oblique notch
#

im actually going to posit that it is 3 segments.

1 first curve upwards
2 straight at an angle
3 second curve back to horizontal.

where the start and ends are
1: start of placment, end is some small amount into the track - say 8 m
3: start is 8m from the end of the placement, end is the placement
and the last part is just whatever is left.

This would explain why curves are showing a slight weirdness.

the first 8 m is generated on that spline, then it starts a second spline for the remainder to -8m from the end, and then a 3rd spline. Those 2 internal points would cause some slight bends as the math restarts, not creating a perfect curve. Potentially when snapping to another track however, if the two tracks are the same Z coord it skips the 3rd segment.

vapid gorge
oblique notch
#

Why tracks are inconsistent in curves

vapid gorge
#

ahh ok I thought it might have been ways to use rail to make arches xD

oblique notch
#

oh... oh.

#

that... oh. i wonder... no no way the foundations change directions, but you could do it with beams.

vapid gorge
#

probably would in fact

#

cause that was my first read about what you were talking about xD a more fluid means for pillar arches

#

not a fan of using rail to make horizontal curves cause it's imprecise but more versatile arches? Sign me up

vapid gorge
oblique notch
vapid gorge
#

Going to try tonight, might make for more flex in making arches

vapid gorge
noble current
#

got a little math question:

For 15 Limiter/min, it'd equal 300 QW/min and 75 Copper Sheet/min

75 CS/min would equal to 150 Ingots aka 5 Smelters 1 overclocked Mk 1 miner, correct?

(QW irrelevant bc I got 600/min intake)

wind spade
noble current
#

pure copper node and standard recipes

fierce cypress
#

wait i thought you said pure copper, as in the recipe simon_smile

noble current
#

hahaa sorry, meant pure node, my bad

fierce cypress
noble current
#

normal copper nodes would be too far away than to make use of

image.png
#

(the circled in is the qw factory)

noble current
fierce cypress
#

clockspeeds

#

7 at 100%, 1 at 50%

#

8 at 93.75%

#

3 at 250%

#

whatever way that adds up to 7.5

wind spade
#

750%

noble current
#

oh yea no, I forgot that it takes 20 ingots for 10 sheets and forgot about that 150 can't be evenly split to even amount of constructors of 20 each

frosty owl
#

7500%0

noble current
#

3 at 250% actually sounds even better thinking about it, personally prefer even rounded numbers

wind spade
#

50% is nice rounded number 😛

noble current
#

nah

frosty owl
wind spade
noble current
frosty owl
#

In other words: changing clocks can make for the smallest load-balancers rolljace

wind spade
noble current
#

but it's something that keeps bothering me personally

noble current
frosty owl
wind spade
#

I'd rather do 8 machines at 93.75%

#

because then you do 1:2

#

oh nvm it's the old screenshot lol

noble current
#

I mean I'd had as many constructors as assemblers, which makes the lines evenly split

#

with 250% each

wind spade
#

if you do 5 constructors at 150% you can do nice 1:1 setup

noble current
#

could you eventually explain me how 5 constructors on 3 assemblers would be 1:1 and not 5:1 while 3 constructors on 3 assemblers aren't? I don't really get load balancing yet, sorry >.<

wind spade
#

I assume we're talking about copper sheet machines?

unknown.png
noble current
#

ye

wind spade
#

ingot -> sheet would be 1:1

noble current
#

ouh that you meant

#

I kept being stuck at sheets -> limiters sorry

frosty owl
#

Splitting in 5 is just "splitting in 6 with extra steps" btw. Literally.
Split in 6, merge 1 back before the splitter. That's a 5-ways split

wind spade
#

or just manifold

noble current
#

or 1 smelter at 250%, 3 constructor at 250%

#

1:3 split - ez

#

wait hold on, I feel like I miscalculated

#

it'd be a 2:3, my bad

#

still clock it to 166.6667% for 1:1 ratio

#

would mean 3 smelters, 3 constructors, 3 assemblers

wind spade
#

inb4 Sev comes and tells you it's not 100% efficient

noble current
#

how?

wind spade
#

you're losing 0.000022222222222222222% 😛

#

because it's 166.66666666...%

noble current
#

but I've manually did set it to 50/min tho

image.png
wind spade
#

yeah but it's not exactly 50

#

game rounds to two decimals

noble current
#

how can it be not exactly 50/min when I set it to 50/min? I changed the amount, not the %

wind spade
#

so it's actually 50.00001/min

wind spade
#

and clock speed is limited to 4 decimals

#

amount/min is never saved anywhere

noble current
#

oh

wind spade
#

it's just "easier" way to set clock speed for people that don't want to calculate it

noble current
#

I mean, don't you just need to type the equation into the % field anyway? I feel like the flat amount field seems useless then but idk

wind spade
#

no matter what you do, you can't get perfect 50/min

frosty owl
#

It's useless if your equation exceeds the precision aviable and you want that extra precision. This covers a small portion of users

wind spade
#

the closest clock speeds are 166.6666% (49.99998/min) and 166.6667% (50.00001/min)
I personally wouldn't care about these small amounts and I'd just set the higher speed (and machine will sometimes stop for a few seconds). But Sev is Sev and he cares about these inefficiencies (even though the game is most likely less accurate than that)

noble current
#

5:3 it is then, I guess

are there any on how to do such a 5:3 lb set-up?

#

I mean I am someone who cares about those inefficiencies as well.. sadlly.. ._.

wind spade
#

merge 5 belts, one splitter

noble current
#

but wouldn't that mean that the Mk3 splits into 90 each?

wind spade
#

I thought you want a 5:3 load balancer?

frosty owl
#

150/3 is not 90...?

noble current
#

no but..

#

doesn't it try to force a 90?

wind spade
#

why would it?

#

why 90?

noble current
#

because I kept thinking, that a splitter tries to fill a belts capacity before "choosing" other belts..

wind spade
#

no, it just goes round robin

frosty owl
wind spade
#

it actually doesn't care about belt speeds or item input/min at all. First item goes left, second middle, third right, repeat

noble current
#

oh

prime orchid
#

@noble currentif you set them to 166.6667% you're not "losing" anything or bottlenecking in any way. it's just not perfect timing. that machine will just be working a tinytinytiny bit faster than what you need it to do, but it's not working less than what you need it too, which in that case you would be losing out on production

wind spade
frosty owl
wind spade
prime orchid
#

wtf is this sushi y'all are referring to?

frosty owl
frosty owl
wind spade
#

they fixed splitter duplication?

noble current
#

well I've just seen by a friend, how he added a smart splitter at his factory and the 2nd output to "overflow", yet even the main belt wasn't completely full, it kept sending to the overflow belt every once in a while, which is what confused me about splitters

wind spade
#

smart splitter can have priority settings

frosty owl
frosty owl
noble current
#

he did set the main belt to [item] and the overflow belt to..well...overflow, yet even tho the main belt wasn't maxed out, it kept sending stuff to the overflow belt every now and then

#

@frosty owl it happened like, last night

frosty owl
#

Was the output the same MK as the input?
What was making you think the output belt was empty, exactly?

noble current
#

well, he transported 100/min on a mk 3 soo...pretty sure it wasn't maxed out - the main belt after splitter was an mk 2, so still not maxed out theoretically

frosty owl
#

MK3 moves items faster than a MK2. If you provide the 100/min in "bursts" (eg: mutlipe items coming back-to-back on the MK3 belt) the MK2 belt won't be able to keep up all the time and overflow will be achieved

wind spade
#

if the items arrived in batches, then it's perfectly understandable that some will go on the side belt

frosty owl
noble current
#

wait, so will it be affected by not-smart splitters aka normal ones as well?

#

so when instead I have a normal splitter instead of the smart splitter, will it still cause an uneven split?

#

sorry, just trying to understand the mechanics >.<

frosty owl
#

Splitters rarely split evenly with outputs of different MKs

#

Eg: splitter with MK3 in and MK2s out will split evenly. If just ONE output is a different MK, the split may not be even

noble current
#

so as long as I have all belts in the 5:3 as one type of belt, it'd stay even, basically?

frosty owl
#

Yes. You can use lower MKs for the outputs (if throughput allows and you prefer doing so), just make sure each splitter outputs on belts all of the same MK (different MKs can pull different amounts of items/min from the splitter, possibly leading to uneven spitting)

noble current
#

I assume similar to mergers? Faster belts would prioritised on being merged over slower ones (output belt would fit max capacity), right?

frosty owl
#

Eh, kinda. Mergers don't prioritize belts, but if a belt can push items in the merger faster than another one, it will
Thus leads to some possibly counter-intuitive results: if you merge 60/min (on a MK2) with a full MK4 belt, the MK4 belt will back up but NOT the MK2. The merger takes more 3 items from the MK4 for every 1 it takes from the MK2, but that's enough to keep the MK2 going at full capacity (60/min)
If the MK2 was a MK1 it would back up slightly.

noble current
#

huh, interesting

#

and yea, that's kinda what I meant to ask, sorry for bad wording

#

out of curiosity, is there a more compact way for a 5:3 setup?

Screenshot20221130-12074100000.png
wind spade
#

yeah, manifold merge the 5

fierce cypress
#

you can build it more compact yes, but in terms of splitter/mergers i think(?) thats the minimum, it is a bottlenecked design though and doesn't input balance

wind spade
# 
|  |  |  |  |
---M--M--M---
      |
frosty owl
fierce cypress
noble current
#

Oh I see now

frosty owl
fierce cypress
# 
|  |  |
---M---
   |      |  |
   -------M---
          |
#

even less input balanced though simon_smile

#

and still bottlenecked of course

frosty owl
#

Input balancing is unneeded unless trying to output specific amounts via smart-splitting later from the merged line (then eliminating/reducing the bursts of items can be useful)

#

Well, unneeded outside of visuals, ofc

fierce cypress
noble current
#

mhm...

Screenshot20221130-12185700000.png
#

I mean it's a good way to dodge foliage one can't remove, I guess

#

I know I only need 150 ingots but I am currently debating myself if I shouldn't just do a 600 ingot/min refinery factory anyway

fierce cypress
#

You can always expand later if you leave room

noble current
#

space management is my biggest issue in this game, unfortunately

fierce cypress
#

Have you considered some verticality? simon_smile

noble current
#

yes, but planning in 3D is my weakspot

noble current
#

plus whatever necessary for SE

fierce cypress
#

Redesigning is always an option, there is a 100% refund on deconstruction

noble current
#

ye

frosty owl
# noble current yes, but planning in 3D is my weakspot

Am easy way to start is piling up experience. Make a second floor whenever you feel like it (don't have enough space, wish to replicate the same as you did on another floor, want to separate productions... Any excuse is good enough), see what you like/dislike and want to change, do better for the next floor. Repeat

noble current
#

@frosty owl well, my biggest issue are multi-level production lines, like for example all the iron/copper/steel factory stuff

#

tried looking for some guides but it felt like it didn't really helped me either..

frosty owl
#

What do you mean? Like how to organize the factory when multiple production steps are involved?

noble current
#

ye

#

organizing the outputs for further products like modular frames on one hand but also storing stuff at the same time - while keeping production rates at the same level without underclocking the "final" product

frosty owl
#

It's nigh impossible making "perfect" layouts the first time. As this is your first playthrough, I suggest accepting that everything you build at the moment will not be perfect and use it all as experience to see what you dislike about it later as you use it
Whenever you may redo a similar production line you'll have a much clearer idea of how you'd like to design your production lines

noble current
#

like for this qw factory for example - pretty simple, caterium ore gets in at 100%, refineries producing ingots at 100%, constructor producing qw at 100% and I got enough to confidentally supply a 100% 15/min limiter factory

but I can't see how I could do it with the copper/steel/iron products tho

Screenshot20221130-12482100000.png Screenshot20221130-12502700000.png Screenshot20221130-12535300000.png Screenshot20221130-12541500000.png
#

like, I feel like I'm missing this "red line" of production chain to get the most out of stuff while being able to store products

frosty owl
#

Maybe using a combination of the codex and the annotations (right side of the screen, like the to-do-list) can help you planning it out thinking_helmet

snow dove
#

I love annotations, best feature added recently

noble current
#

because at some point it's just so many products and side products to keep in mind - iron alone has like 7 "main products", copper 3 and with steel you already need more iron and still need product-combinations out of copper/iron

snow dove
#

a good part of the game is figuring out how to stay organized

noble current
#

it's just like "yea screws, plates, sheets, rods, wires and cables and gets hit by the truck of the other stuff"

frosty owl
#

Eg: I plan for X HMFs/min. This needs Y steel Pipes constructors, so I just increase the constructors by 1 to add some for personal use and plan the steel usage accordingly. When planning later, if something else needs pipes I reiterate the calculations for Pipes and Ingots....

noble current
#

and I am already thinking how to transport rubber and plastic

frosty owl
#

Keep notes and divide the production in "blocks" you find easier to manage

noble current
fierce cypress
#

personally, i usually start with a starter base which i don't care that much about that produces all the basic stuff - say up to tier 4. all the production there just goes straight to the storage room or to other parts in that factory. then for higher tier stuff i will build satellite factories and outposts and build mostly from the ground up only importing some things such as higher tier items or items purposely made at other factories because it makes more sense - such as oil products.

frosty owl
#

So setup 1 would make just smart plating
Setup 2 Modular Frames
Setup 3 stuff for personal use (?)
(Just an example, of course)

noble current
#

ye, but you can't make 3 divided production lines from like 2 nodes, no?

like where I am at rn, the closest are 2 pure iron nodes - I don't really see how one can make for every "base" iron product an individual chain - lets say for each line a separate iron rod line for example - you get what I mean?

#

and since I haven't unlocked trains yet long distance traveling, just to collect a stack of plates whenever needed feels kind of redundant

noble timber
#

Use trucks

#

You don’t need trains for long distance transport

frosty owl
fierce cypress
#

you can make 3 production lines out of 2 nodes, but also word of advice - stop thinking of producing based off nodes, think in terms of items/min, for example if i was creating 15 ai limiters/min, it requires 150 iron/min, i can get this from a multitude of combinations of node purities, miner mks and overclocking.

noble current
#

@frosty owl that's kinda of my worry, that I either don't have enough to make more 1/min for the further products or, that I have wait too long to be able to build further

frosty owl
#

As Tomato said... First build the part of the factory that provides you with items for personal use (milestones/building). Basically just machines connected to starages and sinks if you want to keep them going.
Then you can build the rest of the production based on what ores you have left or whichever parameter you prefer

#

The priority should be having building materials aviable as everything else depends on that

#

(No materials, no factory-building - > you repeat the whole "basic part" issue everytime you move)

noble current
#

yea, but since I already at T5 with HMF, Modular Motors and computers and such, it just feels bit very overwhelming than "just base materials"...

frosty owl
#

It's kinda weird that you wouldn't have production lines for building materials yet thinking_helmet
Or is this project involving a rebuild of the factory you used for basic components too?

noble current
#

ye, I meant more in terms of rebuilding into something that works but also looks nice

#

not just a flat concrete sky

frosty owl
#

Do you have alt recipes aviable for all products?

wind spade
#

if you're rebuilding, don't destroy old factory before building new one

noble current
#

ahhh sorry for adding too many pictures

wind spade
#

and start with one product, build that product, repeat for all products you want

#

for each product, decide how much you're gonna need for building and milestone progression and build that many

noble current
#

evreything up to this point, I have alt recipes for @frosty owl

image.png
frosty owl
# noble current ye, I meant more in terms of rebuilding into something that works but also looks...

I think that first you should decide how to plan it out: wether by end-product (this floor/setup makes X RIPs/min or any other final output) or step by step (this floor/setup makes X iron ingots/min)

While making it all you tweak the numbers as you go (help yourself with notes or Tools not to lose track of stuff), just place some machines to see the recipes, consumption, spacing... and move them around until satisfied, repeating it for all production areas/floors

#

Alt recipes give you a great flexibility in achieving the desired output/min using as few ores as you can

wind spade
#

"simple" way to do it is to have separate production lines for every item

#

so if you produce e.g. iron plates, you only produce them to storage. If something else needs iron plates, you'll build those plates as part of that factory

frosty owl
#

When trying to keep things balanced, I think it tends to be easier keeping everything from the ingots up on separate setups

noble current
#

thinking about it - things like smart splitter also help with separating things easier, don't they?

wind spade
#

not really, we're talking about complete separation

#

even different place

noble current
#

oh

wind spade
#

but if not different place, at least separate setups

#

basically if you want to have e.g. central iron plate production, then you need to know in advance how many plates you need for rest of the game. Which is super hard to do if you're not familiar with endgame and don't have clear end goal in mind.

Much easier is to just make what you need right now, separate everything into their own productions, so that if you choose to rebuild one, others are not affected

noble current
#

out of curiosity - in which dimensions should I plan? I heard some people saying, that they already planning their mk 1 miner factories as mk 3 miner factories.. are things like these adviseable?

wind spade
#

that's exactly why I suggest keeping things separate

#

because then you only need to plan for now

noble current
#

so basically (considering the small space of my current position) 1 production factory/node? this my current "main" area (plus long beach for the 4 oil nodes)

image.png
frosty owl
wind spade
#

basically I mean this

image.png
#

another advantage of this approach is that you don't have to have "main base". Every factory can be placed wherever you want and just the product will be transported to storage (which can be called "main base", but w/e)

#

e.g. place things near nodes it uses

noble current
#

@wind spade out of curiosity, is it okay to split from ingots into, let's say, plates and rods? or is it a 1 node 1 fix product type of thing?

Sorry for being too slow

wind spade
#

the image I sent is just an example.
f.e. the reinforced plates also need ingot->rod->screw, which can be part of that building

noble current
#

or like similar with copper being "main product" being copper sheets and whatever I have left being wire or cables?

snow dove
#

splitting a node is fine imo

wind spade
#

you can have one factory that makes multiple products

#

point is that you never connect factory to another factory

#

always raw ingredients -> production -> storage

#

which gives you the advantage of not having to plan ahead ("how many plates will I need for T6 items?"), but just plan for now ("how many plates do I need for building stuff?").

noble current
#

but doesn't HMF or steel related products basically push you to connect multiple factories with each other?

wind spade
#

and you can always just copy the factory to make more if you need

wind spade
snow dove
#

it requires you to connect multiple nodes

wind spade
#

example of mixed factory

image.png
#

example of separate factory:

image.png
#

total amount of buildings is the same, total amount of products is the same, but they are logistically disconnected

noble current
#

oh THAT's what you meant by factory

wind spade
#

you can do both approaches

#

but the bottom one is easier imo

#

because you don't need to know how many modular frames you're gonna make at the time you're building reinforced plate factory

noble current
#

yea, if it'd be me, I'd do add a container on each byproduct of the upper picture additionally

wind spade
#

there's no byproduct though

#

both are desired products

cinder silo
noble current
wind spade
#

yeah that's where you build smart splitters and overflow things to sink

#

if your storage is full, it makes sense to sink the rest, so that machines never stop and you get some nice coupons from it

cinder silo
#

Those coupons add up big time, I have more than 2700 spare thanks to an overflowing warehouse (and ofc plutonium rods)

noble current
wind spade
cinder silo
#

You're going to have to terminate with a sink sooner or later otherwise you will have an epic logjam.

cinder silo
noble current
cinder silo
noble current
#

I just unlocked s.splitters around I've finished my caterium production factory

noble current
#

I just did built with what I had at that time

wind spade
#

storage should always be the main destination for your items

#

with secondary destination being sink

cinder silo
#

Leave cheating until later imho, it can be amusing in any game but if you do so early you rob only yourself, keep at least one game cheat free if you're experimenting imho 🙂

wind spade
#

overflowing into another factory means that said factory doesn't work 100% of time, which is weird

cinder silo
#

I have a setup where material will overflow from one factory to another, but never in to storage.

noble current
#

@cinder silo yea no I don't cheat, just saying if I would had

cinder silo
noble current
#

all good

cinder silo
#

The reason my save is at 2000h is because I personally don't cheat and have spent ages running experiments to stress the game out and find bugs.

snow dove
#

HOLY

noble current
#

so best case with this would be demolish all, take one iron node, one concrete node and a coal node, make a single factory just for s.beams and pipes and i.beams only, right?

#

as an example

cinder silo
#

At least you didn't need to demolish a 1.2km facility while spending two days doing it 🤣

noble current
#

what is 30^8 x 26^8?

#

240m x 208m

#

kinda a good size tho

cinder silo
#

Oops sorry I had a quick chat with ben about the conveyor holes break, I just wasn't paying attention, sorry.

noble current
#

all good

#

..maybe I should've turned the miners off before starting to get rid of everything

cinder silo
#

Yeah shutting down a site before squaring everything away is generally a good idea.