Changes to the pictures:
- The output filter on the top splitter is not relevant, just a remnant of the previous tests.
- The input priority on the lower splitter is not relevant.
- The bottom splitter feeding back into itself seems to be relevant. I was not able to trigger the bug without this condition met.
Additional infos:
- The size of the loop seems to be irrelevant.
- The top splitter is not relevant to the bug, just a convenient way to permanently top up the loop. Works also with side feeding belts or inserters topping up the loop.
- All information about belt sides are relative to the direction of the belt.
The "minimal requirement" for the bug:
- 2020-07-08 20_07_16-Window.png (1.25 MiB) Viewed 1214 times
After removing or adding the bottom belt at the splitter items somtimes "escapes" the loop on the left side of the splitter.
This behavior seems to be mostly random and can be triggered by removing any of the first 4 belts connected to the splitter in this location or by adding up to the forth belt to the line connected there.
So far experienced are:
- 0 items escape
- 1 item escapes on the left or right belt side
- 1 item escapes on each side
- 2 items escape on one side, one on the other side (both variations)
Also right next to the bottom splitter a gap can be seen on the right side of the belt. After the screenshot was taken and the gap moved all the way around another item "escaped" the loop on the right side of the belt.
- 2020-07-08 20_07_31-Factorio 0.18.35.png (1.22 MiB) Viewed 1214 times
On this last image the gaps on the belt can be seen. Those gaps only appear after modifying the short belt line connected to the bottom splitter.
Sometimes a gap can get "stuck" (for up to ~1 second) at a turn of the loop and delays the escaping of an item.
- 2020-07-08 20_09_34-Factorio 0.18.35.png (556.32 KiB) Viewed 1214 times
