Minimal size fully loaded circular belt stops rotating

[madtulip]

The title basically says it and im not sure if its a bug - its unintuitive and "wrong" somehow anyway. If you create the smallest possible belt circle (2x2 positions) and load that from the side with some other belt until its full the whole thing will stop rotating. Even if you disconnect the loading belt and just keep the small circle. The same thing applied for larger circles. Whats even more unintuitive is that if one side of the belt is blocked (by beeing full) the other keeps rotating (if not full).

A possible workaround is to ensure at least one free space in the circle with circuit networks reading the belt content.

[DaveMcW]

If you attach a side-load belt to a 2x2 circle, it is no longer a circle.

But you can also fill a transport belt using stack inserters, and the inner ring does stop moving.

2x2.gif (50.54 KiB) Viewed 2054 times

[daniel34]

I don't understand why the outer lane slows down or the inner lane keeps going normally and suddenly stops, but according to the devs this isn't a bug: viewtopic.php?f=23&t=29029

Its not a bug, the belt is full, it doesn't need to loop if its full, and it would be a lot of work and un-optimization to keep a full belt looping

I believe it has to do with the new sideloading technique implemented in 0.13 which tries to find gaps to insert their items and therefore blocking the belt. You can see it in the video above, the inner lane stops at the same time as the last possible piece of coal can be placed (on an otherwise already side-loaded and saturated inner lane).

EDIT: In my testing I've also established that if the inner circle is saturated before the outer circle it stops immediately, but if the outer circle is saturated first then it takes 2-4 seconds for the inner circle to stop after being saturated. This means that the lanes are affecting each other somehow.

[madtulip]

"The belt doesnt need to loop" - well that might be the case if each side of the belt only holds one type of item.

However this is not the case in a scenario where i encountered the behavior. In my case each side of the belt was loaded with a random type of up to 20 different items. 20 different filters inserter would unload one specific item type each from the loop. Whith the described behavior a setup like this has a possible deadlock. Its also counter intuitive to what you would exspect from physics. Its the extreme/edge case not handled correctly imo.

[Rseding91]

If you shove a belt so full of items that no one item can move without hitting the item in from of it then none of them will move.

That's just how it works.

Not a bug.