I fail to see a reason why it is happening. But I can reproduce the very same behavior in Sandbox mode. The principle may be same on any surface, but on space platform the circuit cycled assemblers are more likely having this behavior.
The involved layout is very basic:
- {34A9765A-A15D-42E4-B3AC-7CEEF5C43463}.png (967.01 KiB) Viewed 355 times
To fill the Tank with Fluoroketone while platform being in Aquilo orbit. By picking up FK barrels and emptying them into the Tank to certain Tank level.
To fill back some barrels from a Tank when in orbit of planets with requests of barreled FK.
Consume some FK from a Tank otherwise (not being in orbit to provide/process barrels).
How to replicate:
In Sandbox mode, build Assembly machine 3 bellow the hub with two inserters in both directions between the hub and Assembler.
Add couple of Solar panels to supply power. Add a pipe to the south side of Assembler and deploy a Tank right bellow that pipe.
Put some Fluoroketone (Cold) barrels into hub inventory, so Assembler may eventually get those barrels and empty them into the Tank.
Set Assembler manually to Foundation recipe, and rotate it so it is connected to the pipe and Tank.
Build one Constant combinator and one Selector combinator near the Assembler.
Configure CC it with three recipe signals:
Empty Flouroketone (Cold) barrel
Fill Flouroketone (Cold) barrel
Foundation
Connect CC output to the input of SC, configure SC in Random Input mode, and Update interval 500.
Connect CS output to Assembler, and configure Assembler to "Set recipe" mode.
What happens:
Circuit will start cycling the recipes, but switch of recipes causes Assembler to rotate it's pipe interface to opposite direction of the pipe actually having a required fluid for unknown reason. Pipe itself and Tank behind it always contains only the same type of the fluid, and that should be compatible with all involved recipes, but may differ with inbound / outbound flow directions.
What I expect to happen instead:
Assembler probably should not change the orientation of it's pipe interface, or at least not if the fluid contained in the pipe is very same kind of fluid the recipe is using.
It does happen whenever the recipe is updated, either manually or by the circuit. It does even when I explicitly manually rotate the Assembler to the correct orientation while having every involved recipe set manually before having it set by the circuit afterwards. On next switch of the recipe the pipe interface is rotated to the opposite side of the pipe than it is desired.
I have build many quite complex single Assembler, multirole, circuit controlled recipes having nonfluid/fluids recipes in the mix. But this is first time Assembler rotates away from the pipe. Does it have something to do with the fluid flow direction or pipe having "something" in it (not considered as usable for active recipe)?
