[2.0.15] Electromagnetic plant electrolyte disappears with no output

[SamsaraRazor]

While trying to fix a fluid mixing issue, I came across what I think is a seperate bug.

Use attached save file to reproduce.
1. Deconstruct electrolyte plant
2. Using either ctrl+z or manual build, replace electrolyte plant.
3. Electrolyte plant will produce electrolyte, but even without output, the produced electrolyte will disappear.
Eventually, the plant will keep its output and stop producing due to full output. (While testing, plant would stop mostly at level 110 and once level 50, but I'm not sure this is a bug or related.)
Possibly, the issue I was trying to solve and this bug is related.

[The_Sane]

Check 119600 to see if it might be the same fluid mixing issue as what you are experiencing, if you aren't doing anything to cause it yourself. Currently happens every time the game receives an update.

As for fluid output disappearing, EM plants have multiple outputs of the same fluid and each seemingly has its own 'fluid box', but currently only one is displayed in the interface. See 119796. You are probably seeing the output draining into the other internal buffer that isn't displayed.

[SamsaraRazor]

Check 119600 to see if it might be the same fluid mixing issue as what you are experiencing, if you aren't doing anything to cause it yourself. Currently happens every time the game receives an update.

As for fluid output disappearing, EM plants have multiple outputs of the same fluid and each seemingly has its own 'fluid box', but currently only one is displayed in the interface. See 119796. You are probably seeing the output draining into the other internal buffer that has isn't displayed.

I tried my best to avoid duplicate issues. I've posted on the Issue here viewtopic.php?f=7&t=117703 because it was originally related to my thrusters. Could very well be due to upgrade today as in 119600 with the EM mixing. Quite likely what I'm seeing is also related to 119796, and that these invisible buffers could be the cause of the mixing in the first place. I think the invisible buffer doesn't correctly distinguish between fluids/pipe networks, so that it just outputs whatever if that pipe network is empty. The solution is either to use the delete button in a pipe connected to whatever is wrong, or pumps as diodes. But it is a pretty severe limitation that pipes cannot ever run out or you risk your entire pipe network the entire time.

[The_Sane]

But it is a pretty severe limitation that pipes cannot ever run out or you risk your entire pipe network the entire time.

From my experience running a pipe dry wont cause the "leak", but if the pipes are already contaminated (by 119600 or something else) then yes running a pipe dry will allow the errant fluid to take over and fully block the system.

[SamsaraRazor]

But it is a pretty severe limitation that pipes cannot ever run out or you risk your entire pipe network the entire time.

From my experience running a pipe dry wont cause the "leak", but if the pipes are already contaminated (by 119600 or something else) then yes running a pipe dry will allow the errant fluid to take over and fully block the system.

I'm getting a bit confused by all the different fault modes at play. But I'm pretty sure that is what is happening in case of the thruster.

In the case of my holmium/electrolyte mixing from 117703 there wasn't originally any pipes connected to output of electrolyte, it was just connected straight to the input of the electrolyte needing machine. But even with rebuild of the electrolyte machines, it outputs holmium solution in empty pipe from the electrolyte output. This is why I think there are more things at play at the same time, but without proper knowledge of actual inner workings of the fluid system, I'm just confusing myself thinking about different types of failure, theories, etc. I think something isn't right, but without a statement from the devs, I can't figure it out.

[spalch]

I have a similar bug, only my electromagnetic plants will randomly output heavy oil from the electrolyte ports.

https://imgur.com/a/nF1HyVk

[the_potty_1]

I'm getting the same bug, placing a tank on either of the electrolyte outputs shows the tank to contain 0 Holmium solution. The EM plant does contain 200 electrolyte, which does not move to the tank as expected.

Screenshot 2024-11-20 132808.png (737 KiB) Viewed 703 times

[the_potty_1]

So to add to my report above, I rotated the EM plant, and in the tank connected to the Holmium solution pipe in the above image, 5k of Holmium solution immediately changed into 17k of heavy oil. This is absolutely repeatable in the above save. An interesting point about that 17 is that prior to the switch, the intervening pipe contains 17/100 Holmium solution.

Sorry, I'm a coder, that kinda seems odd and might be useful. Can I just rave about this game for a short moment. Actually about the coders of this game. I've been playing factorio for ten years, I'm ashamed and a little proud to say I have almost 4000 factorio hours logged on steam, and this is the first bug I have ever encountered. Personally, I've released hundreds of bugs into the world, some cataclysmically bad. So you coders, you are goddam legendary in my eyes.

EDIT ok that might be basic flow-through of heavy oil through the electrolyte pipe. Personally, I'd be ok with only one electrolyte port, but I don't build big, perhaps the megabase maniacs need flow-through.

[Rseding91]

Everything described here is due to flow through machines having internal pipes. The internal pipes fill with the result of the crafting recipe even when no external pipes are connected which is the “disappears when no pipes are connected” (it’s not disappearing it’s going into the internal pipe).

The internal pipe also can fill with fluid the machine doesn’t want or need which is the “it’s making heavy oil when I rotate it” (because the internal pipe filled with heavy oil when rotated and then rotating it again connects it to the pipes containing not heavy oil)

That’s to say, there are no vanishing fluids or fluids created out of nowhere here. But it is apparently ultra extremely confusing to players and so they make bug reports.

I thought people would understand the internal pipe but I was WRONG.

[robot256]

I thought people would understand the internal pipe but I was WRONG.

The part no one understands is why assembling machines are allowed to have unfiltered pipe connections at all. It would make sense if the passthrough pipe would prevent the wrong fluid from entering and delete any wrong fluid that can't be ejected. I understand the limitations of how it's currently implemented but that's the root of the UX issue.

[computeraddict]

I thought people would understand the internal pipe but I was WRONG.

I think it's just a visibility thing. If there were some way to see the internal pipes it would alleviate most of the problem. I'll do some MS-Paint UI design tonight. The functionality provided by the internal pipes is great so I hope you don't decide to axe them.

[runamucker]

I'm getting the same bug, placing a tank on either of the electrolyte outputs shows the tank to contain 0 Holmium solution. The EM plant does contain 200 electrolyte, which does not move to the tank as expected.

Screenshot 2024-11-20 132808.png

A quick fix:
I had the same problem. When I opened the pipeline contents, there were both Electrolites and Holomium solution. When I deleted the Holomium solution, the Electrolites started to flow properly, so it's not a gamebreaking bug.

[the_potty_1]

The internal pipe also can fill with fluid the machine doesn’t want or need which is the “it’s making heavy oil when I rotate it” (because the internal pipe filled with heavy oil when rotated and then rotating it again connects it to the pipes containing not heavy oil)

OK forget about the rotation, that was just to gather extra data. The actual problem is that if I place a tank on an empty electrolyte port, the tank immediately contains zero Holmium solution, and the 200 electrolyte remains in the EM plant. I'll try deleting everything and rebuilding from scratch.

That's not to say there is not very weird fluid behaviour around EM plants. Flow-through I accept, but it's deleting existing fluid, where in every other case, if 2 fluids would come into contact, neither fluid would move until you correct the mistake. Actually I'm not sure what happens when you bring 2 fluids into contact, and that's really my point, whatever happens is not particularly bad, so I don't really know how it behaves.

[the_potty_1]

OK, I fixed it. Place your em plant somewhere else, rotate it correctly, copy it, then place the plan in position. Probably do this without robots itching to place an EM plant until everything else is in position.

Factorio EM plant 1.png (252.2 KiB) Viewed 535 times

Factorio EM plant 2.png (222.56 KiB) Viewed 535 times

[the_potty_1]

The part no one understands is why assembling machines are allowed to have unfiltered pipe connections at all. It would make sense if the passthrough pipe would prevent the wrong fluid from entering and delete any wrong fluid that can't be ejected. I understand the limitations of how it's currently implemented but that's the root of the UX issue.

It's taken a while but I now understand what you mean here. I agree, it is partly an ux issue. In the original game, and even in industrial revolution & Bobs & Space age, there was nothing that you couldn't place then rotate into position. Well I don't recall anything that didn't just work. as for flow-through, only the uranium miner did that? My first exposure to the more complex new flow-through was the spaceship engine. I went for the fairly obvious inverted V, after trying and failing to do a straight V. I did see a blade design, which looks very pretty but never occurred to be because I do tend to think in squares, or at least mirrored layouts.

The behaviour of the EM plant is just unexpected and a little buggy. If the plant is entirely full of 200 product, I assumed that product resides in the flow-through pipe, but the behaviour suggests the flow-through pipe still contained zero holm solution, and the product sits somewhere else. Coming from another of my favorite games - oxygen not included - I assumed two fluids can't occupy the same space. Well they still can't, but the product appears to be stored in some other non-intuitive place?

[BlueTemplar]

Steam engines are also pass-through.

Fluids can share the same fluidbox in 2.0 now. (Maybe not in practice since 2.0.16 though ??)

And fluid touching prevention has been introduced for a while now : since 0.17
(though empty fluidboxes being 'reserved' for a fluid, and also pass-through machines having a fluid icon on the pipes might be more recent ??)

----

You might have missed one of the biggest current discussion about related issues :
Mixed fluids in electromagnetic plant setup

----

See also what is potentially a big change introduced in 2.0.16 to try to fix these issues :

[...]
[*]Fluid mixing will prefer the fluid with more volume and discard the other.
[...]

[runamucker]

Fluids can share the same fluidbox in 2.0 now. (Maybe not in practice since 2.0.16 though
...

[...]
[*]Fluid mixing will prefer the fluid with more volume and discard the other.
[...]

I hope we agree that regardless, a pipe should never output a fluid other than what its label says.

[BlueTemplar]

Yeah, as you can see, I was asking for that change (as well as fluidbox contents' display) instead.

[MechBFP]

I would think it would just be better to not allow any fluids through unless a recipe is set, and then of course the only fluid it will take is the one set in the recipe.

Is there actually any practical situation where you would want any other fluids passing through it anyway? Even with circuit controlled settings of recipes I can’t see how that would be useful.