Ran a test with 1:4 trains and the result doesn't seem ideal. (see attachment)
I got 42/33/48/41.Factoriointersection wrote: ↑Sun Nov 21, 2021 3:12 pm
That is pretty cool:) it gets a score of 53 which is better than an 2 lane unbuffered intersections. Regulary turning signals of, with combinators, causes trains to repath and try to avoid the intersection.
Nice, but it deadlocks.Lubricus wrote: ↑Sun Nov 21, 2021 5:59 pm I have a design that I haven't seen any other use and shouldn't be too bad. 4-way unbuffered
https://factorioblueprints.tech/api/str ... c6df053472
https://factorioblueprints.tech/bluepri ... 450d3521dd
1-4 trains results are not comparable to 2-4 trains.causa-sui wrote: ↑Sun Nov 21, 2021 8:27 pmRan a test with 1:4 trains and the result doesn't seem ideal. (see attachment)
Did I do something wrong? Or is this just a BP where the worst case is extremely bad but in practical situations, it does better?
Edit:I got 42/33/48/41.Factoriointersection wrote: ↑Sun Nov 21, 2021 3:12 pm
That is pretty cool:) it gets a score of 53 which is better than a 2 lane unbuffered intersection. Regularly turning signals of, with combinators, causes trains to repath and try to avoid the intersection.
I think there are missing signals one the screenshot from the test :Factoriointersection wrote: ↑Sun Nov 21, 2021 9:28 pmNice, but it deadlocks.Lubricus wrote: ↑Sun Nov 21, 2021 5:59 pm I have a design that I haven't seen any other use and shouldn't be too bad. 4-way unbuffered
https://factorioblueprints.tech/api/str ... c6df053472
https://factorioblueprints.tech/bluepri ... 450d3521dd
Unfortunately noFactoriointersection wrote: ↑Sun Nov 21, 2021 9:41 pm
Do you have the LHD version as well? I'll test and add this one
I know it has a pathfinding penalty, but does it really make trains repath if they are already travelling through the intersection?Factoriointersection wrote: ↑Sun Nov 21, 2021 3:12 pmRegularly turning signals off, with combinators, causes trains to repath and try to avoid the intersection.
This is the quality content I'm here for
No, not when they are already commited to traveling through the intersection. However, they do repath way before then while they still have multiple options available. Trains recalculate their path quite often while traveling through the train network, every time they have to slow down for example. When your rail signals are set to red during that recalculation, the train will probably chose a less optimal path that avoids your intersection, granted there is one available.DaveMcW wrote: ↑Mon Nov 22, 2021 12:44 amI know it has a pathfinding penalty, but does it really make trains repath if they are already travelling through the intersection?Factoriointersection wrote: ↑Sun Nov 21, 2021 3:12 pmRegularly turning signals off, with combinators, causes trains to repath and try to avoid the intersection.
Damn I was impressed when I first saw the compressed multicross. I got it from your comment on the previous forum thread in case you're wondering.Sopel wrote: ↑Mon Nov 22, 2021 12:03 am Thank you for your effort! This thread was really needed.
One note, Compressed Multicross can, I think, be tested with unsafe output. Even though the signals on the output lanes are placed closer than train length apart there are no chain signals there. Additional safe output would only be needed if the intersection ended earlier than the last chain signal could be influenced from. Same goes for other intersections with similar properties. It might change the throughput.
Also, the actual size is 128x128 (4x4 chunks), but the poles stand out (to be tilable). Here's the original reddit thread too https://www.reddit.com/r/factorio/comme ... tion_with/.
I suggest enabling comments in the doc, it would be easier to post corrections that way I feel
btw. I think it would be nice to know which intersections are best for UPS too. Possibly UPS/train measure could be useful. The pathing can also impact performance, as well as diagonal tracks. I was thinking about a setup where there's a grid of, say, 8 by 8 intersections and trains going randomly between all 32 possible ends.
Then just add the chainsignals then it won't deadlockFactoriointersection wrote: ↑Sun Nov 21, 2021 9:28 pmNice, but it deadlocks.Lubricus wrote: ↑Sun Nov 21, 2021 5:59 pm I have a design that I haven't seen any other use and shouldn't be too bad. 4-way unbuffered
https://factorioblueprints.tech/api/str ... c6df053472
https://factorioblueprints.tech/bluepri ... 450d3521dd
Checking the other intersections. It's only a slightly tighter version of this https://factoriobin.com/post/01B6XBlBFactoriointersection wrote: ↑Sun Nov 21, 2021 9:28 pmNice, but it deadlocks.Lubricus wrote: ↑Sun Nov 21, 2021 5:59 pm I have a design that I haven't seen any other use and shouldn't be too bad. 4-way unbuffered
https://factorioblueprints.tech/api/str ... c6df053472
https://factorioblueprints.tech/bluepri ... 450d3521dd
This is correct !Lubricus wrote: ↑Mon Nov 22, 2021 10:14 am Checking the other intersections. It's only a slightly tighter version of this https://factoriobin.com/post/01B6XBlB
So it's an improvment but slight.
.Made an variant for 2:4 trains the original is for 3:8 trains.Lubricus wrote: ↑Mon Nov 22, 2021 10:07 am 3-way instersection 4 lanes buffered. Should have hight throuput and the design is quite easily expanded to more lanes if needed.
https://factorioblueprints.tech/bluepri ... 83aac6d9ce
I asked Bocian to convert it to LHD and he did, because he's a swell guyFactoriointersection wrote: ↑Sun Nov 21, 2021 9:41 pm Do you have the LHD version as well? I'll test and add this one
).
