#How does belt throughput work?
1 messages · Page 1 of 1 (latest)
When the belt is entirely full and isn’t backed up, 30 items will go through a tile every second
15 per side
gotcha
is there a way for me to use that to calculate how much time it takes to move one item down a conveyor line of a certain length?
I mean, yes, but that’s not something that’s ever really useful to know for this game
You’re never working with just one item
I see what you're saying
I guess what I'm curious about is the roundtrip time for a set of items
Why?
just out of curiosity
I guess I want to understand how to use the belt throughput concept in a useful way
the issue I have right now at my base is that I have calculated rates that I have dedicated to particular items
so for example I'd dedicate a certain amount of iron plates per second to making circuits, another for making engines, etc.
since the rate of the thing I'm moving is usually less than the belt throughput, I just move it across
but things change when I start moving it by train
because then I have to control the rate by splitting the belt somehow
It’s useful for knowing how many belts you need to transport the rate of items you want, that’s pretty much it
Not really
The game gives you backpressure as a tool to never have to think about setting exact rates for splitting things
If the belt fills up, it stops
what is backpressure?
If items hit the end of the belt, they don’t spill off, they just sit thrrr
So items behind them have to wait on the belt
And it ends up acting like “pressure” all the way back to where you split off from
so you try to produce less of something if backpressure exists
You’re still thinking a bit too actively about it
You don’t have to do anything to slow down production if the belt backs up, that automatically happens
How much of the game have you played?
right and I get that. I guess what I want to do is try and not give too much of something rate-wise to something that doesn't need it
I'm at 80 hours in my solo world
and 40 into a coop world I believe
have yet to launch a rocket
Which doesn’t really happen, is what I’m saying
I'd like to explain a scenario in which it does
If it can’t eat it that fast, the belt backs up. If it can, eventually the output will back up and then the input backs up in response
Any scenario where it happens is temporary until the belts fill
I get that but think about this for example
let's say I bring iron plate out of a train at a rate of 30 per second, meaning I use the whole belt
and I want to dedicate some of that iron towards making circuits and engines
I don't want to dedicate the whole line just towards making circuits or engines
30 per second is too much for just circuits if there's other things that want to use that line
is splitter with priority towards circuits/engines not solving the case?
This is precisely the scenario I described though
I typically use splitters with priority for this sort of thing
.
but then the issue I get is that I'd have multiple splitters with priority across the belt and I have to make sure that things back up that far just for that to work
use several belts instead then. let's say mall+red and green sci consume 1 full iron belt (for example). then give them said belt and call it a day
sometimes I put stuff in passive provider chests
A single passive provider should be fine
Usually when people don’t “get” backpressure it’s because they’ve never seen it because they hoard things in chests for no reason
the reason I worry about this sort of thing is I don't want to pollute too much when it comes to producing ore
it is best to put passive providers on lowest priority belt branch
right I get that
but the fact you have to worry about priorities in the first place is what can get things out of hand
If you’re that worried about priorities then I think it’s a sign you don’t have enough production to adequately supply everything
Like, there’s something going on here if it’s that big of a concern
A screenshot would be helpful at this point
I want to make sure I don't pollute too much which is the thing
Yeah production is the issue
Then use efficiency modules
Or take the fight to them
If you’re not playing a deathworld then being that deathly afraid of pollution is just getting in the way
I agree
I guess what I want to do is just dedicate and exact rate to the things I supply
Micromanaging things that closely is not going to be easy
It can sorta be done with a circuit to rate limit the belt
But again, the game already does that for you
yeah, it's not really worth the effort
build bigger, massacre bugs and enjoy the game
sure
I see what you're saying
I'll give what you guys are recommending a try and see if it works out
thanks for the advice
I guess while I'm at it though, don't you guys at least try to figure out the exact ratios of things when made
Eh
i do - but it always measured in whole amount of fully working assemblers
Within a given production block, yes
so no micromanagement
right
It depends on the goal
but differences in these things do matter
If this is a “starter”/mall base, then I’m not trying to dial in exactly how much plate production I need because I’m going to need more lager anyway
no I know but sometimes you scale a little bit ahead if you know what I mean
Exactly
like advanced circuits are made at 4 times the rate it takes to make blue science
So you overbuild production
so you can overbuild advanced circuit assemblers and waste space if you don't do it right
For some context, this is a base where I was purposely trying to not overbuild like I’m mostly used to
Hold on, clipboard broke
nw
that looks beautiful
Much more forest but still
rate calculators is a big help for complicated bases - but before it some overproduction is really whatever
https://kirkmcdonald.github.io/calc.html#data=1-1-19&rate=s&items=automation-science-pack:r:1,logistic-science-pack:r:1
like here it shows 2.5 gears assemblers, but making it 3 is a ok
Yeah if it’s not a megabase, trying to micromanage production that hard isn’t helpful
right
I think what got me interested was when I was trying to do some math to figure out how much I need of everything down the line to make a certain amount of blue science a second
I realized that if I rounded at certain points and stopped the overproduction can be quite bad
again - building bigger IS the trick - take a look at the second link in comparison - beautiful round numbers everywhere
https://kirkmcdonald.github.io/calc.html#data=1-1-19&rate=s&items=automation-science-pack:r:2,logistic-science-pack:r:2
exactly 1 yellow belt of iron, no waste here too
You can do the math without rounding up at each step
yeah that's usually what I do
I try to leave the rounding to the end
or to where it'd be reasonable
thank you guys so much for bearing with me on this @mystic kindle @cedar wolf
and sorry for being pedantic
The items per second limit of belts comes from the length of an item on a transport lane and the speed at which the transport lane moves. A belt has 2 transport lanes and if it is straight these can hold exactly 8 items. This means each item has a length of exactly 1/4 tiles. If a belt moves 1 tile per second it can move 8 items across a tile per second, the maximum throughput of the belt.
As belt speed increases, fewer items need to be on the belt to achieve the same item per second count. This can be tricky to visualise since a compressed yellow belt with 8 items per tile (4 per lane) will turn into a red belt with 4 items per tile (2 per lane), half the density, while throughput remains the same.
As far as planning goes, you need to make sure that the sum of production going onto the belt or consumption trying to be removed from the belt does not exceed the belt item per second count. If this happens, then some production buildings will either not be able to unload their product onto the belt, or will not get the inputs needed to operate.
Splitters have a theoretical throughput of 2 belts of the same tier. The exception is if one of the outputs is blocked up in which case total thoughput will be lowered by how much that output is blocked up. This is especially a problem when dealing with spliters running in filter mode since if the same item appears on the same lane of both inputs at the same time, the output is blocked causing reduced throughput. For this reason using a splitter to filter 2 belts with two different item types mixed equally together the result will be significantly less than 2 belts of throughput. This can be an issue when dealing with some mods that have recipes output an assortment of different items that you may place on the same belt.
Did you read any of the above conversation at all?
Partly, and did not see people explaining that items have a length, which to me is the most confusing part about how items per second work.
It’s not needed to understand it at all though, and the conversation actually went toward “micromanaging belt throughputs means you’re probably doing something wrong”
You can certainly calculate an item length, but that’s not helpful whatsoever
This is an interesting read
in the scenario you're describing with two of the same input in to of the same side line, you should put one of the inputs on the other lane of the belt
instead of using a splitter
It’ll still only be half throughput if both input belts were full
right but then you wouldn't want to merge them
It's only typically an issue if you are dealing with mixed ore and are filtering it to put it on separate belts, in which case throughput isn't much of a problem because it's just one belt of ore and a mild temporary slowdown. For most anything else you should never have different items in the same lane of a belt
If you're doing some sort of sushi belt base, or use science belt sushi, maybe you'd run into it there too, but sushi bases are not exactly made for high efficiency
The problem mostly shows up in mods like pyanodons where a lot of recipes output 2+ different items. If you unload these onto the same belt and try to push belt throughput limits you can run into filter splitters becomming the bottleneck when trying to finally separate the product to be more useful. In such case I find the solution is to use a filter splitter per belt and then merge the resulting belts using additional splitters. The expanding and then reducing avoids any potential throughput bottlenecks from the filters on splitters.
I recommend the rate calculator mod to help made sense of things. You can select some machines or miners, whatever and it will show you the items/sec provided or needed. Basically if you want things operating at maximum efficiency, giving your machines enough items/sec or allowing them to output enough items/sec onto a belt is pretty important. https://mods.factorio.com/mod/RateCalculator
For example, if you want 30 items/sec of copper coil, and one machine outputs 15 copper coil/sec, you'll want to use 1 red belt or 2 yellow belts, one machine for each side of the belt for red belt or 1 machine per belt for yellow belts. maximum efficiency (the machines are always going to output their maximum if all of the copper coil is being used and the machines always have copper plates and an empty belt to place onto).
The other factor is how do you insert onto the close and far side of a belt? For that, you can either use bob's adjustable inserters mod or just realize that you can only output 7.5 items/sec (for yellow belts) since the inserters only put items onto one side of the belt. You'll have to combine your resources making one full belt to actually get 15 items/sec