Efficiency Checker - does not take in consideration belt splitters
So this is pretty simple to reproduce:
1. Get an input from miner or machine, for example Mk3 miner with 250% overclock in my case will get you 1200/min ore
2. Get a splitter right outside the machine
So now we have 2 belts and a simple splitter has (theoretically) exactly 50% output rate.
However ECM tool still shows it as 1200 per each belt.
I do acknowledge that simple conveyor splitter can have different output based on how input is being taken - depends on the further bottleneck, but I think the bottleneck should be calculated accordingly based on the connected machine\etc, right? I am just not sure how calculation goes right now 🙂
But now even if I connect a smelter to one of the conveyors, which is for example 30/min.. the result is somewhat better but still imho wrong - one belt is shown as 1200\min (the one connected to the smelter) and the other one is 1170/min (which is correct as the other belt is using only 30/min). Same goes if you connect belts into storage and then belt out of industrial storage - you will have each belt showing 1200/min regardless of real input after splitters.
Overall it may be frustrating, even I am confused on how it should work, as honestly predicting\calculating throughout versus input MAX versus real consumption is.. tough! But thanks for hearing me out 😄
P.S.
In my real-world scenario I've just started to build a factory and I have 30x refineries for water+copper ore receipe wich requires 15x ore per minute per machine and having 2 lines to the same input belt = 900 per minute raw ore. So I am now splitting pure nodes from 1200/min into two belts with 600/min (or four with 300/min) and then merging with normal and impure nodes to get exactly 900/min planned around though belts, which is easily achieved with "50% splitter" logic in mind. But I guess I will see real stats only when I will connect the machines(ore consumers)... hm.
11 Replies
@Marcio Neves just found as mod owner and summoning you to this purely insane & raw flow of thoughts 😄
It does not split even. And never will. It is consider what is being actually consumed on each side. It will never load ballance, it accounts for pure manifold only
hm, are you sure... I mean I can't agree as I built systems with splitters as simple and effective load balancers..
E.g. let's take a simple example:
1. raw ore node 600/min
2. 30/min required input smelter x 10 mahines x 2 lines (e.g. exact 600 consumption)
do you say this setup with splitters as load balancers will never distribute ore across machines evenly it will never run smooth on 100% efficiency?
P.S. if so, that would ruin my life and all factories I built >.<
Solution
I really can't visualize what you trying to explain 😓
no worries, I will see once factory setup, then it will fix naturally I guess based on machines consumtion 🙂
do you know the difference between Manifold and Load Balancer?
this is a manifold.
this is a load balancer.
The first example, a manifold, will overflow into the next machine every time. It will eventually run at 100% efficiency, once all the machines except the last one/two are full. Manifolds need time to fill up - but you can accelerate this process by prefilling machines.
This is because half the materials will go into the first machine and half will go onto all others. Splitters always evenly divide between outputs. Once the first machine is full, it will only take however much it needs for its cycles, and the rest will go on. That small branch to the machine will always be stalled except when the cycle finishes.
The second example, a load balancer, will evenly split materials into each machine. This is harder to accomplish, when you are doing not an even number of machines.
With simple load-balancers the checker still goes somewhat ok. It pitfalls if you start to have back-injection, like the 1/5 balancers, or re-mergers. For the two examples above, it would read fine the 4 "B" requirements if read before the first splitter. things start to getty confusing if you read after any of the splitters, specially in the load-ballancers
yea, it will be 50% for the first machine, then 25% for the second and etc for manifold. But when input >= outtake then it will work as constant feed across all machines pretty evenly.
my main point of the topic is that during construction it is impossible to check stats for such systems - regardless of balancer\manifold - unless you finish your build completely and connect all the betls. so in my example I am trying to do splitting to get even numbers (900/min) across belts, so I have to keep in mind "which belt is which" - e.g. have two 600/min and then four 300/min but they are from 1200 via splitter. and despite I know and it is definitely how it will work after belts connected, it will be shown incorrectly on ECM mod 🙂
But anyways, there is no way to make it work I guess, so it is just how it is...
🤷🏽♂️
Intake being consumers greater than what's being sent to the line?
It doesn't really balance out. The first one in the line will have all it needs, maybe a few more depending on how much a given machine needs. After that they all stall out.
And there is an animation startup \ shutdown time on machines, which means they actually are less effectient than the number of parts being sent to the manifold indicates it should be
But if your math is on point it will yes, that's how a manifold will work 🙂