Pipe fullness bug

[swiftdeath007]

Depending on pipe fullness, any machine with a fluid input will create X amount of product based on how full the input pipe in. In the pictures we can see that without a pump the lube underproduces, even though it has plenty of fluid in the machine and pipe. But adding a single pump increases pipe fullness and now the lube fills up the tank and performs as it did in 2.0.77. In 2.1.9 this issue occurs with the new fluid mechanics. (if someone else has posted this bug LMK)
This is a very weird bug to find, and massively borks a bunch of previously working machines. And is unreasonable in the future to include pumps in every large scale build. The only difference between the correct production and underproduction of lube, is the pump for heavy oil. If need be, I can provide video comparing 2.0.77 and 2.1.9 and comment the link to it, if that'll help fix the bug. Thank you in advance. (world is a test world for submitting a benchmark. The last user is the host)

07-07-2026, 19-45-35.png (51.73 KiB) Viewed 659 times

07-07-2026, 19-45-59.png (57 KiB) Viewed 659 times

07-07-2026, 19-46-09.png (26.65 KiB) Viewed 659 times

07-07-2026, 19-46-48.png (51.95 KiB) Viewed 659 times

07-07-2026, 19-46-58.png (26.52 KiB) Viewed 659 times

07-07-2026, 19-47-09.png (60.53 KiB) Viewed 659 times

07-07-2026, 19-47-27.png (133.27 KiB) Viewed 659 times

07-07-2026, 19-47-40.png (144.35 KiB) Viewed 659 times

[Hurkyl]

I don't understand what you're trying to demonstrate. What are you actually measuring? What, specifically, is the difference between 2.0.77 and 2.1.9? And what does the whole relevant system look like?

If the only thing that's changed is the fluid level in the heavy oil and lubricant pipelines, that doesn't imply you are underproducing lube: it may well just be that the steady state of the system has different fluid levels than the previous version.

[swiftdeath007]

What I am trying to demonstrate is that based on how full the pipe is, any product with a fluid input will make more output products based on how full the input pipe is. So if you're only 20% full at all times, then you make less product than if you're 90% full. In this specific design I am showing off my competition yellow science build, which has zero issue in 2.0.77. But in 2.1.9 with the newer fluid mechanics (they are better) The exact same design under produces. And the solution is to add pumps to increase pressure, when before the entire build worked flawlessly. For context, the build produces 2880/s yellow science in 2.0.77, but makes significantly less in 2.1.9, with the only significant difference that would affect the two builds, being the newer fluid system. These pics are in 2.0.77 showing off the exact same lube machine. And included circuit logic. The pics originally posted are 2.1.9, same build. Attached is also the

07-09-2026, 00-43-43.png (214.2 KiB) Viewed 527 times

save file with my yellow science BP, so you're welcome to see the exact same issue pop up. As you can see, the pipes and build will work without issue in 2.0.77, but with the newer and better fluid mechanics it simply breaks, and the only solution to get it 98% fully working is adding pumps in key areas to increase pipe pressure which was never an issue before in 2.0.77. The save file is in a state where the BP is already working, and is already in editor mode to speed up time and watch it never under produce. Jump to 2.1.9 and within 3 min the build underproduces (excluding the two heavy oil plants that get flipped when switching over, it'll still underproduce)

07-09-2026, 00-38-27.png (511.73 KiB) Viewed 527 times

07-09-2026, 00-38-47.png (434 KiB) Viewed 527 times

07-09-2026, 00-38-57.png (96.75 KiB) Viewed 527 times

07-09-2026, 00-39-08.png (61.37 KiB) Viewed 527 times

07-09-2026, 00-39-19.png (67.89 KiB) Viewed 527 times

07-09-2026, 00-39-33.png (74.61 KiB) Viewed 527 times

07-09-2026, 00-41-47.png (79.87 KiB) Viewed 527 times

07-09-2026, 00-42-25.png (76.41 KiB) Viewed 527 times

[immibis]

The speed that oil transfers into the machine depends on how full the pipe is. That means machines might work slower when the pipe is nearly empty, because they have to wait for enough oil to transfer into the machine to start the recipe.

If this isn't your problem, then I don't understand the problem. Surely someone would have noticed if the recipe would produce less output from the same amount of input. They have unit tests for this.