Liquid valves with different behaviours

You don't have a loop there. Just confused pipes that sometime work and sometimes don't.

are you sure the liquid you see "splitting", comes from the valve and not from the overflow of the bridge?

1 hour ago, sakura_sk said:

are you sure the liquid you see "splitting", comes from the valve and not from the overflow of the bridge?

On the one working, they are all going in the right direction (from valve to bridge). 

In the one not working, should I force the overflow to fill up, by blocking the pipe after the bridge, the moment I restart the flow, the water will favour going down into the valve instead of up towards the bridge.

Edit: I managed to take quick videos and convert them into rough gifs to illustrate:

Edited February 27, 2022 by NeerieD20
Adding media

Never mind... I would love to hear a sane suggestion to what I'm seeing here:

same build, depending on the position behaves differently .

I've seen some (what I guess are) bugs in the sim in Spaced Out but nothing in the base game yet and certainly not in pipes (except from the steam deletion bug).

Sorry... I have no idea what I'm looking at. This build should work as you described or it shouldn't work. In either case it should be consistent at doing one or the other.

well.. which way *should* the flow go there? It either terminates at the white bridge input, or it terminates at the white valve input. If the game doesn't know for sure which one should be the terminus, it will make a decision... that decision is not always consistent.

It's best to make it absolutely clear which way flow should go, by adding another bridge or valve between the two white inputs.

Even if it were consistent in some way, what would it be consistent to? If you flip it upside down, should it do the same thing?

8 hours ago, Yobbo said:

well.. which way *should* the flow go there? It either terminates at the white bridge input, or it terminates at the white valve input. If the game doesn't know for sure which one should be the terminus, it will make a decision... that decision is not always consistent.

Is it not always consistent though? The way I see it, a bridge and a valve act as direction priorities, if the content (in this case liquid) is free to pass, it will prioritize that direction, and if it can't pass in full (valve flow control or exit pipe blocked), then the content will continue onto an overflow pipe if there's one.

This is after all one of the way how people build element filters for liquids and gases.

 

In this build, there's only one element in the pipes, so no element conflict. The valve is set to let half the content of a packet go through, the other half is free to go through the overflow to the bridge, which is what happens if I do not merge both pipes at the bridge, but only after (which I guess will be my temporary fix for the time being).

6 minutes ago, NeerieD20 said:

Is it not always consistent though? The way I see it, a bridge and a valve act as direction priorities, if the content (in this case liquid) is free to pass, it will prioritize that direction, and if it can't pass in full (valve flow control or exit pipe blocked), then the content will continue onto an overflow pipe if there's one.

You're assuming a "starting direction" - that all the liquid is initially coming from the aquatuner, and this is providing some sort of pressure which pushes liquid away. That's not how pipes work in the game (though it would be nice if it were).

As far as the game is concerned, the origin point of all liquid could in stead be the valve output, because all it sees is the green outputs, the white inputs, and the pipe connections. In that case the liquid should all push toward the input furthest away from the valve output (which happens to be the valve input), and that is what was happening in the non-working case.

A good rule of thumb is that you should always have sequences like "green green white white" and never "green white white green", because ultimately that's all the game can see. The first sequence has unambiguous direction everywhere, but the second sequence has no way of determining the correct direction in the middle white-white segment, without some extra notion of pressure. The fixed version you have now uses all the first type of sequence, and thus has no conflicts.

It does also require some care for loop filters. If you bridge on to them in the wrong place you get the same problem, which can lead to some confusing behaviour.

It seems the game "decides" the pipe direction depending on how long the pipe is. My question is: Should it do that?

There was another bug report concerning pipe priorities.

I was used to the "pipes are confused. You need to put a bridge there" advice but now it seems there is a "it can work sometimes but not work on other times. The game will decide"

From my experience length does matter and the game prioritizes the shorter length. I've frequently encountered it when I loopback a long section of pipe without a bridge into a conditional bridge build (the one where it will fill the primary line if it is empty). Without putting direction bridges in that long section, it will reverse flow instead instead of going the direction I want.