Belt layout affects transport speed

[Joe78man]

Guys, just noticed belt curves accelerate item tansport. I created this to test it:


I press R to let some resources pass and then R again and off they go


Iron got to the finish line first, almost 3 secods before

[5thHorseman]

Just change the graphics and we've got diagonal belts. It almost looks like it's about the correct distance for that.

[ssilk]

please upload the pics into forum, because that external link doesn’t work for some browsers. Thanks.

That is just due to that the outside lane has - I think it’s 10 items - but the inside lane only 5 item positions (Don’t nail me on the exact numbers). Straight belt has 8.

So it’s clear that it is a bit faster, because of this one position difference (5-8 is -3, but 10-8 is only 2).

And it doesn’t influence the throughput of the belts, don’t mix speed with throughput!

[wobbycarly]

https://en.wikipedia.org/wiki/Pythagorean_theorem

[DanGio]

https://en.wikipedia.org/wiki/Pythagorean_theorem

This doesn't apply here, Factorio world being structured in tiles, the setup is more like this :

triangle.jpg (35.46 KiB) Viewed 3390 times

There is the same amount of red squares & blue squares so both items should arrive together.

However, as "item travel time"' is by far less important than item throughput, I'm pretty sure this won't be fixed. This probably comes from the several years old fix that prevented belt turns to slow down item throughput (which was a major concern)

[SyncViews]

This doesn't apply here, Factorio world being structured in tiles, the setup is more like this :

Yes, but there are two lanes on the belt, and on corners the distance is not equal, and they are not completely "square", so counting the number of tiles is not exactly right.

If the items followed the red line at normal speed, then yes, each tile is always "1".

But they don't, the items on the right side take a much shorter route on the inside of the corner, and the items on the left take only a slightly longer route, not long enough to make up the difference due to it being "rounded".

[Joe78man]

I think the outter side of the belt accelerates to catch up with the inside side where items travel at regular speed, hence, accelerating it. If you then turn the other way arround you accelerate the other side, winnig a few millisecs of acceleration in both sides of the belt, one at a time.
Not reporting an issue or expecting anything. It's mere observation and sharing with you guys who probably like objective data and observation too.