[0.16.36] Is this a splitter bug or intended?

[zOldBulldog]

Splitters don't always seem to behave as I'd expect. Usually do, but please look at the screenshots.

- The splitters are set to Left Output Priority.
- When I use up to 4 splitters in this cascade configuration, product seems to consistently shift left as much as it can.
- When I use more than 4 (6 in this case) there seems to be a traffic jam condition in the splitter chain, causing the rightmost splitter to output a full belt even if other belts aren't outputting little or nothing.
- Using Blue splitters with Red belts helps a little, but it still can't handle the chain of 6 splitters.

I addressed the issue in my build by adding a second chain of splitters on the 4 belts after I draw off the first two. But it feels hacky.

So the key question is:

- Is the traffic jam a bug in the splitters? Or is failing to cascade past 4 an intended design constraint?

- If it is the intended design constraint there must be a good reason. I would love to understand it. If it is, can someone enlighten me?

[posila]

If I understand it corretly, the right most (the bottom most) splitter doesn't behave as you'd expect.

On its input, it seems to have 2 compressed copper belts, on so it will push as much copper as possible to the priority output and rest will go to the other output. So if both inputs are saturated, both outputs will be saturated too.

splitters2.png (102.3 KiB) Viewed 2530 times

[zOldBulldog]

If I understand it corretly, the right most (the bottom most) splitter doesn't behave as you'd expect.

On its input, it seems to have 2 compressed copper belts, on so it will push as much copper as possible to the priority output and rest will go to the other output. So if both inputs are saturated, both outputs will be saturated too.

Good catch. Not a bug. If I had started feeding the 6 lanes from the left I would not have even noticed the issue. And no wonder the second set of splitters did the trick, that one is fed by only one belt. LOL, price of building one Smeltery at a time

[The Eriksonn]

I agree on the fact that some compact splitter systems can be hard to debug since it is hard to see what is going on between them, so i usually do as posila and leave a gap of belts so that you can see what is going on. It helped alot when i tried splitter sorting before it became a feature.

[Raish0]

Can someone explain me the issue here in different words? I read it over and over but can't quite get what is happening.

[zOldBulldog]

Can someone explain me the issue here in different words? I read it over and over but can't quite get what is happening.

I was expecting no output on the right lane under the conditions shown in the screenshot.

Posila figured out the cause, very obvious once you know what to look for.

[dood]

As a sidenote, if you just run the belts into greens without going for a bus, there is no need to split them at all.
You just need to balance the input belts and that's that.

Here's an 8 belt balancer. Just slap that at the input of the copper belts and call it a day:

blueprint

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