I’m working on an in-depth understanding of belt balancing, and discovered some behavior that I find pretty counterintuitive! Thought y’all might enjoy!
Edit: reuploaded video that wasn’t absurdly huge
This kind of headache is why I always copy other people’s tried and tested balancers instead of making my own :D
Wanna know something awful? I found this by analyzing the shortcomings of the 3-3 balancer in one of the more popular community blueprint books XD
Nice example. What you are basically doing is splitting the belt into two, where none of the belts backs up. Hence you get a throughput of 50% only.
The backwards loop does not back up since the splitter splits 50-50, and the “joiner” joins 50-50 as well. So the rate at which stuff goes into the backwards loop is the same at which it leaves.
At the point where you break the backwards loop, it does back up, and then the splitter splits 100-0, i.e. puts everything on the outward belt.
