Old chem plant setup...fixable?

[GlassDeviant]

I have this old array of chem plants that I used to switch between cracking light oil to petroleum and making solid fuel from light oil, simply by changing the 'recipe'. This doesn't work in 0.17.

https://imgur.com/a/XPdDeeJ

I get an error message when I try to set it to solid fuel, and I can't set cracking yet as I don't have the tech for that researched.

Any way I could revive this, or will I need to make separate arrays for the two tasks? I would prefer to have the switchable setup if it can be fixed so I don't have to waste resources and space on two separate arrays, but if whatever the update did to break my old setup can't be fixed I haven't got a choice.

[Durentis]

You can't mix fluids anymore, and you have water input where it changes to light oil in the solid fuel recipe.

You'll have to make a separate array unless you can pump out the water just before you switch to solid fuel. Perhaps something like having a water pump leading into the array get disabled, and have a water pump leading out of the array turn on to pump the lines into a tank, which would go back into the array when the input pump is re-enabled. But that said, I'm not sure even that would completely drain the line or set it such that it wouldn't count as a water pipe for fluid mixing checks.

[GlassDeviant]

Thanks, I missed that. I'll try pumping the water out first, as it's just a simple step rather than having to build an entirely new array and rearrange everything else.

Edit: No luck, even with the water sucked out the game still sees them as "water pipes". Why was this change made anyway? It seemed to work fine before.

[Hannu]

Edit: No luck, even with the water sucked out the game still sees them as "water pipes". Why was this change made anyway? It seemed to work fine before.

It prevents wrong fluid going to pipe. Pipes are easy to connect unintentionally and there was no more practical way to clean contaminated fluid line than disassemble it and build again. It was extremely annoying, especially in huge modded refinery or chemical setups with tens of liquids and hundreds of plants. In my opinion this was very good change.

[5thHorseman]

In situations like this, I just delete the pipes and relay them.

Just like I used to do when fluids mixed, only now I am 100% sure I won't screw it up and remix them.

[BaggyK]

In situations like this, I just delete the pipes and relay them.

Just like I used to do when fluids mixed, only now I am 100% sure I won't screw it up and remix them.

I would change that to delete the pipes, change the recipes and then relay the pipes. Due to how detecting the mixing is done, changing the recipes whilst they are connected together will not be possible.

[GrumpyJoe]

It prevents wrong fluid going to pipe. Pipes are easy to connect unintentionally and there was no more practical way to clean contaminated fluid line than disassemble it and build again. It was extremely annoying, especially in huge modded refinery or chemical setups with tens of liquids and hundreds of plants. In my opinion this was very good change.

imho, it dumbed down the game, with no real benefit. I dont like the feature.
Specially in "huge modded games"
Im playing pY, and it has lots of weird fluid in and outputs. When you try to rotate a machine, it often doesn´t allow it cos its the wrong fluid, without anything in it yet.

Its annoying. Since the the 0.17 release, i was swearing at the game alot more often than i was angry at myself over the previous last 1.5 years combined for actually mixing up fluids. Which, as soon as you had bots, was actually VERY easy to fix.

It should a least ignore machines, cos they wouldn´t take in the wrong fluid anyway.

[Koub]

[Koub] Moved to Gameplay Help.

[Hannu]

Its annoying. Since the the 0.17 release, i was swearing at the game alot more often than i was angry at myself over the previous last 1.5 years combined for actually mixing up fluids. Which, as soon as you had bots, was actually VERY easy to fix.

I see that annoyance level depends on player's accuracy and playstyle and current solution is far from perfect. Maybe cleaning function would have been a better option. You could open a small control window by clicking any pipe, just like many other entities, and there would be "clean line" -button, which would remove all liquids in all pipe components connected to that pipe. It may be unrealistic, but if we accept that we can hold dozens of locomotives, factories and orbital rocket launching equipment in pocket for practical reasons, in my opinion we could also accept disappearing of few liters of chemicals by a button press for practical reasons.

[GrumpyJoe]

I see that annoyance level depends on player's accuracy and playstyle and current solution is far from perfect. Maybe cleaning function would have been a better option. You could open a small control window by clicking any pipe, just like many other entities, and there would be "clean line" -button, which would remove all liquids in all pipe components connected to that pipe. It may be unrealistic, but if we accept that we can hold dozens of locomotives, factories and orbital rocket launching equipment in pocket for practical reasons, in my opinion we could also accept disappearing of few liters of chemicals by a button press for practical reasons.

I dont see how that would be possible
The new system marks a pipe(system, as far as connected) with the fluid of the first connected machine. There is actually no fluid in it if the machine isn´t already working, still its reserved for one fluid only, showing 0.0 of that (lets say 0.0 light oil, at a single pipe, connected to a refinery that never ran cos crude is not connected)

You couldn´t even prepare those unmodded refineries water intakes, as before Advanced Oil Processing, both inputs are oil and it doesn´t allow a water connection.
And thats an input, there cant be crude oil coming out of it

[5thHorseman]

I don't see any reason that the game can't check an entire pipe once every time the recipe changes, and if the pipe is completely empty and not hooked up to any other inputs, just change it. Of course, this wouldn't allow you to switch from regular to advanced oil processing with multiple refineries, but at least it'd allow you to clear the recipes and reset them without removing the pipes, so long as you didn't allow oil in them in the first place.

[Hannu]

I dont see how that would be possible
The new system marks a pipe(system, as far as connected) with the fluid of the first connected machine.

It would not be possible with current detection system. In my opinion it would not be needed at all if there were simple and reliable method to clean contaminated pipe.

[Mr. Tact]

Yeah, I have to agree the new way of marking pipes with the type of fluid when they are connected to a source or destination isn't a great solution to the problem they were attempting to fix. The part which annoys me is previously an empty pipe looked empty, now you can be "fooled" into thinking the pipe has fluid in it when it doesn't. I understand the reason the change was made, they were addressing a problem, but their solution only causes different problems which are no better than the previous problem.

[GlassDeviant]

Something I noticed from a fresh perspective this morning is that after I pumped the water out AND disconnected the pipe-to-ground between the pump and the offending water pipes then got the recipe working, said water pipes became light oil pipes, even though the only thing connecting them to the light oil pipes is the chem plants themselves.

Somehow the chem plants which have a water input and a light oil input while running the cracking recipe, but both as light oil inputs when running the solid fuel recipe, are allowing flow between the two inputs when the latter is running. Before the pipe fix, the chem plants didn't care whether there were two light oil inputs supplying them while making solid fuel. They accepted the light oil from one and blocked the water from the other.

The flaw is not in the pipe changes, but in the chemical plants, which was not evident before the pipe changes and now needs to be fixed if my design is to be soft toggleable again without having to make hard changes to the piping every time the recipe is changed. I'm not demanding a fix, but I think it should be done, since it was a pre-existing flaw in the chem plants that was only revealed/became a problem after the pipe changes.

[GrumpyJoe]

Somehow the chem plants which have a water input and a light oil input while running the cracking recipe, but both as light oil inputs when running the solid fuel recipe, are allowing flow between the two inputs when the latter is running.

are you sure about it, that the light oil is really flowing out of the input? or is it just marking both inputs for light oil, even tho one is not connected to a light oil source, whithout actually getting any.

I understand the reason the change was made, they were addressing a problem, but their solution only causes different problems which are no better than the previous problem.

Thats an understatement imho.
I think it turned an almost non existing problem into a real one.
The game flow for people that take a bit care and hardly ever mixed up fluids, is now interupted very often.
Like i said, i was never swearing that much before. This is screwing me up while building, ALOT more than it will ever actually help me

[Ultros]

You can't mix fluids anymore, and you have water input where it changes to light oil in the solid fuel recipe.

You'll have to make a separate array unless you can pump out the water just before you switch to solid fuel. Perhaps something like having a water pump leading into the array get disabled, and have a water pump leading out of the array turn on to pump the lines into a tank, which would go back into the array when the input pump is re-enabled. But that said, I'm not sure even that would completely drain the line or set it such that it wouldn't count as a water pipe for fluid mixing checks.

I was trying to switch oil type between light and heavy for my rocket fuel plant over the weekend and it didn't work out, the pump removed all the light oil, but the pipe refused to change to blank even after the plant itself was removed and the pipe was showing as light oil even though it was showing 0 for its entire length. I had to rebuild the entire system in order to get the fluid switch working.

It's more likely that you'll have to rebuild two separate arrays for the two fluids, as I am doing for my cracking build (two sets of crackers crack heavy/light oil respectively)

[GlassDeviant]

Somehow the chem plants which have a water input and a light oil input while running the cracking recipe, but both as light oil inputs when running the solid fuel recipe, are allowing flow between the two inputs when the latter is running.

are you sure about it, that the light oil is really flowing out of the input? or is it just marking both inputs for light oil, even tho one is not connected to a light oil source, whithout actually getting any.

As it turns out, the light oil was not flowing out, only the assignment of those pipes to carry only light oil. This is ridiculous. I've seen design decisions that made me scratch my head or even question what the devs were thinking before, but this smashes right through everything else. I've never before seen anything that I would even consider calling bad, but this is bad. Browsing the forums, I can't even keep track of all the complaints that boil down to this cart-before-the-horse process of preventing fluid mixing. The devs are brilliant and have achieved something far beyond my capabilities in making this game, but this time they've out thunk themselves.

If the new system only prevented you from putting one fluid INTO a pipe that already contained another, or was reserved by another SOURCE for a particular other fluid, that would make sense. But reserving a pipe for a fluid because of what the DESTINATION machine expects makes no sense at all. The chem plants always ignored the water input when they were expecting two heavy oil or light oil inputs, and it worked fine for everyone. Being able to switch a cluster between cracking and making solid fuel was common practice and there were no problems in that particular area. Learning how to place pipes without getting fluids mixed up was an art and part of the learning process of the game. Now we have policing and hand holding (the kind of hand holding a game for OCD people should never have) and it is causing more problems than it ever solved.

[Durentis]

I see the fluid mixing change as a very good one and a solid step forward for the game.

What OP is doing is an awkward edge case design. Sure it's a sandbox but, as far as "you're doing it wrong" can go in a sandbox (not very), this is a good example. I don't see wanting to do something like this as a good reason why the fluid mixing change was a bad one. I do see it as an example of a design that has been forced to be improved by the improvement of its environment.

• Changing recipes is not automated and is clumsy.
• The extra space for another row of chem plants is insignificant.
• Solid fuel should* be made from Petrol not Light. (Heavy -> Lubricant, Heavy -> Light -> Petrol -> Solid)

I'm not surprised that pumping out the water didn't work and I'm actually kinda glad. That would have just made the design even more clumsy in an attempt to keep it working more or less as before without improving it.

As far as the source/destination argument goes, I think it's good as-is with both reserving the pipes. One or the other could just cause confusion whereas both keep it simple and intuitive.

[GlassDeviant]

It is neither awkward nor edge case.

Not awkward because it was not quite trivial to design, but close enough not to matter. It got the job of allowing two processes to be performed by one chem plant array done simply and in a relatively compact space while allowing for extension at the far end. That is the same definition of elegance as is enshrined in the open source philosophy of simple elegance.

Not edge case as I have seen plenty of examples of switchable chem plant arrays, and they are much more capable and automated than mine, allowing for pretty much any chemical process to be performed by each array instance. Since my goal was only to switch between light oil products, I didn't need that complexity.

Also, space for another row of chem plants is not the primary issue, it's the pipe connections that are the biggest pain. The connections in my existing setup are specifically placed to make them as uncomplicated as possible. Space isn't even a secondary issue, it's tertiary after pipe connections and the extraction of solid fuel from the entire refinery and processing area.

What solid fuel should be made from is not only a matter of opinion, but a completely irrelevant issue, though I will point out that this entire setup is from an earlier version of Factorio where making solid fuel from light oil was the most efficient choice.

As for whether pumping out the water would have made the design clumsy or not, that also is irrelevant as it was a test along the way to figuring out a solution.

One final point, I did not say that the entire fluid handling system was bad, just the part where a target machine reserves a pipe for particular contents. Your post implies that I suggested the entire system should be thrown out.

While we're at it, let's see your idea of a petroleum setup that is neither clumsy nor awkward.

Edit: I just looked at the wikia, which seems to be outdated, and github, which specifically refers to 0.17. Both show that light oil is the more efficient source fluid for making solid fuel as it takes 10 light oil to make one solid fuel block and 20 petroleum.

From github:

40H -> 30L -> 20P -> 1SF
40H -> 30L -> 3SF
40H -> 2SF

Making petroleum the WORST fluid to convert to solid fuel. There are much better uses for petroleum, like plastic and sulphur.

Light oil is clearly the best, not only from a pure materials cost perspective, but because it takes additional power to crack the light oil to petroleum. And yes, it also costs to crack the heavy to light, but you're going to be doing that anyway, and the additional energy cost of cracking to light gives you MORE solid fuel from that initial heavy oil, whereas converting to petroleum cuts your end product by two thirds, a foolish waste of resources. The only excuse for turning petrol into solid fuel is when you have such an excess of it that your refineries are no longer able to produce heavy or light oil because of a backlog. In that case you are simply disposing of excess, which you should never have gotten because you should have automated your light oil to petroleum cracking to stop when petroleum stockpiles began to far outstrip light oil supplies.

edit 2, since I missed this part of your post before:

As far as the source/destination argument goes, I think it's good as-is with both reserving the pipes. One or the other could just cause confusion whereas both keep it simple and intuitive.

You seem to have a strange idea of what intuitive means. Things go from a source to a destination. What ends up at the destination depends on what is sent from the source. A waterfall drops water down to a pond; the pond does not restrict the waterfall from dropping water simply because there happens to be tar seepage from a nearby bog encroaching on the pond. Defining the reservation for a pipe by the destination is not intuitive, it doesn't make sense at all. That's backwards twisted logic.

[Durentis]

I have a strong suspicion we're simply not going to agree, and that's fine since much of this game is personal preference.. that sandbox thing. But I'll address a few things above since you took some time to respond so thoroughly.

I think you're misapplying/overusing your notion of "simple elegance". And to continue drawing from software design, the Single Responsibility Principle says roughly "entities should do only one thing". The Open Closed Principle states that "entities should be open for extension but closed for modification". Your design does two things (crafts Petrol and/or crafts Solid Fuel) but could be said to do one thing (crafts a recipe) if recipe selection were automated, but it isn't. As-is, the recipe is like a hard dependency (tight coupling) that is a part of the design itself. If the recipe selection were automated, that would be more like Dependency Injection (loose coupling), which would be great. Your design is perhaps open for extension, in a manner of speaking, but is also open for modification (because you must change the recipes). You have to physically revisit the array of chem plants and change the recipes manually when you want/need to. There is nothing elegant about creating a long-term maintenance burden of any complexity.

My suggestion that Solid Fuel should be produced only from Petrol comes from an old solution to problems caused by fluid backups and is attached to a couple premises. First, raw oil is free. Because oil patches are plentiful and each gives infinite supply (especially once surrounded in speed beacons), the amount of oil it takes to create an end product becomes irrelevant. Second, power is free. Coal power is very easy to scale early on to whatever you need and the moment you go nuclear you have infinite power requiring practically no resources and very little surface area. So again, the amount of power required to crack Light -> Petrol is irrelevant, and only increasingly so as time goes on. As a result, all that matters is that the amount of solid fuel created from some amount of Petrol is greater than zero and that the refineries cannot back up unless the Petrol does, which means that your factory has no need for an oil refinery at the moment. You're not wrong that Light is the most efficient source of Solid Fuel but I see that efficiency as irrelevant and the results of ignoring that efficiency far preferable. If this were real life, the waste would be horrific.. but it's a game where trains take tight flat corners at top speed without derailing. I have seen a refinery solution that slightly simplifies the circuitry to what I have been using and I'll probably switch to it, but it more or less amounts to the same both in terms of layout and buffered fluid levels.

Let's assume that inputs/destinations shouldn't reserve pipes to their fluid type, as you say. I lay down a chem plant and set it to crack Light-> Petrol and then decide to start connecting pipes from the Light input. Obviously I need it to go to pipes connected to a Light output, but somewhere along the way I misplaced an underground pipe between others and it connected to filled Water pipes. Oops. Now I have to sever it, drain or remove them, and start over because it's insufficient to just disconnect the offending pipe and continue towards a Light input. Part of the point of the change is to prevent this and to do so the input/destination has to reserve the pipes. Is it more reasonable to create pipes from output/source to input/destination than vise versa? I don't think so, regardless of the ultimate flow direction when the connections are complete.