Olfah wrote: Thu Oct 24, 2024 12:32 amits only getting like 62 TPM. thats ass for a 4 lane train setup
We're all busy trying to make pretty and small intersections with the new elevated rails and curves, but a lot of best practices for intersection design still matter and aren't being applied. Buffers within an intersection that match the length of your trains (with some specific extra for accel/decel) are still an important part of making a high speed intersection. It looks like you've made your intersection with all chain signals, so only one train can be on each path through the intersection at a time, enforcing a *huge* gap between trains on your network.
Take a straight rail and put a gap between signals as long as this intersection is wide. How many TPM can go along that straight track? Whatever that number is, it's a cap on the possible performance of an intersection with no rail signals (interior buffers) of the same width.
Re: 2.0 Intersection Compendium
Posted: Thu Oct 24, 2024 12:23 pm
by mmmPI
sparr wrote: Wed Oct 23, 2024 6:24 pm
If 4 trains enter at the same time, each turning left, and they are immediately followed by 4 more trains, each turning left, the cloverleaf will fill up and jam with no train being able to proceed. The first four trains will be on the curved rails waiting to enter the straight rails on their way out of the intersection, while the second four trains will be on the straight rails waiting to enter the curved rails on their way into the intersection.
True, i didn't know for what to test, i got some crash using the mod for some reasons, manually setting groups of 4 synced left turning trains caused deadlocks.
Re: 2.0 Intersection Compendium
Posted: Thu Oct 24, 2024 12:28 pm
by mmmPI
hansjoachim wrote: Thu Oct 24, 2024 9:21 am
but you also copied the fucking layout of our intersection post you even added the roundabout for comparison.
That is not exactly true, the original layout was more precise as it included different categories wether the intersections were buffered or not.
Not sure who is reponsible for setting up the exact same test setup, but i still see no reason to use 2-4 trains as the base for troughput calculation on a junction and trying to rank them all given that train size as it doesn't reflect properly what to use those junctions for.
Curious to hear the arguments for it in both case
Re: 2.0 Intersection Compendium
Posted: Thu Oct 24, 2024 12:41 pm
by Arcus
EDIT: Merged into first post.
Re: 2.0 Intersection Compendium
Posted: Thu Oct 24, 2024 1:25 pm
by sparr
Once again, I *strongly* recommend that we as a community shift from putting all intersections in one thread to using one thread per intersection. Then we can have separate compendium threads for best-in-class, different categories, different testing methodologies, etc.
Re: 2.0 Intersection Compendium
Posted: Thu Oct 24, 2024 2:50 pm
by TiMatic
I created another design that is exactly 96x96 and fits perfectly into the 32x32 chunk grid.
I used a cubic outer frame and a knot in the core where the tracks switch sides like a diverging diamond.
Hi,
I tinkered around a little bit and realized that when there are no crossings in the intersection, you dont need to use chain signals, because the split or merge would be blocked by the train anyway.
All that aside, here is my design:
Snímek obrazovky 2024-10-24 183010.png (452.96 KiB) Viewed 4191 times
And it performed wonderfully
Snímek obrazovky 2024-10-24 183053.png (40.41 KiB) Viewed 4191 times
I know its pretty big, but the throughput really surprised me.
Anyway, happy designing!
Screenshot 2024-10-25 005334.png (549.34 KiB) Viewed 4203 times
So something I was thinking about is that so far most of the designs have been assuming that the intersection exits will always be at ground level, but this may not always be true. Might it be worth adding a category for whether the exits are elevation agnostic, ground level, or elevated?
Re: 2.0 Intersection Compendium
Posted: Thu Oct 24, 2024 5:15 pm
by Factoriointersection
.
Re: 2.0 Intersection Compendium
Posted: Thu Oct 24, 2024 5:26 pm
by TBTerra2
i have shut down this thread (as much as I can cos i dont know if threads can be deleted), and made it point towards the main one.
Thanks everyone, and remember, the factory must grow
Re: 2.0 Intersection Compendium
Posted: Thu Oct 24, 2024 6:24 pm
by hansjoachim
I'm an idiot. Sorry
Tberra is great at a lot of things and i went too far.
Re: 2.0 Intersection Compendium
Posted: Thu Oct 24, 2024 8:21 pm
by khift
TBTerra2 wrote: Tue Oct 22, 2024 12:39 pm
to add one of my own designs to the list I present the streamline flyover 3-way
its rather wide as a whole, but narrow in profile. I could make it smaller but that would involve making it manual driving un-friendly, or changing the rail spacing.
it gets a score of 83.6 which is around the same as the best 2lane 3 way buffered intersection from 1.1
i suspect the 2lane 3way limit for score is around 90, with the 2lane 4way limit around 115-120
Can you repost this blueprint? The link in the main thread links to the wrong blueprint.
Re: 2.0 Intersection Compendium
Posted: Thu Oct 24, 2024 8:30 pm
by TBTerra2
khift wrote: Thu Oct 24, 2024 8:21 pm
Can you repost this blueprint? The link in the main thread links to the wrong blueprint.
im sure hans will fix it shortly, but until then the slimlines bp is: here
TBTerra2 wrote: Tue Oct 22, 2024 12:39 pm
to add one of my own designs to the list I present the streamline flyover 3-way
its rather wide as a whole, but narrow in profile. I could make it smaller but that would involve making it manual driving un-friendly, or changing the rail spacing.
it gets a score of 83.6 which is around the same as the best 2lane 3 way buffered intersection from 1.1
i suspect the 2lane 3way limit for score is around 90, with the 2lane 4way limit around 115-120
Can you repost this blueprint? The link in the main thread links to the wrong blueprint.
Yeah, there was an exception not added in the forum generation script. Fixed it=)
Re: 2.0 Intersection Compendium
Posted: Sun Nov 03, 2024 1:50 pm
by Maelstromeous
Slightly off topic but how are you guys getting the stats for your intersection throughputs like in the image here?
Re: 2.0 Intersection Compendium
Posted: Sun Nov 03, 2024 4:44 pm
by coppercoil
Maelstromeous wrote: Sun Nov 03, 2024 1:50 pm
Slightly off topic but how are you guys getting the stats for your intersection throughputs like in the image here?
I wanted something that fits both substation grids and 100x100 city blocks (Nilaus style) and looks reasonably symmetric/balanced. Pretty happy with this and now working on an elevated mainline version.
Solid throughput, cheap with a minimal 4 ramps and 2 supports.
11-10-2024, 12-16-50.png (778.81 KiB) Viewed 2607 times
TiMatic wrote: Tue Oct 22, 2024 2:53 pmBocian2 has published some of his ideas for elevated rail interchanges in this Reddit post.
I created blueprints for some of his Celtic Knot designs. Celtic Knot (Variant A)
RHD
LHD
Celtic Knot (Variant B)
RHD
LHD
Variant A is near identical to the design from LuziferSenpai, but Variant B is both different and pretty good with a score of 89 for RHD and 86 for LHD
as an added note: i went and tried to make the intersection from FFF417
Im not 100% sure on the north and south outputs as we cant see them so this is just what i got to work
as the largest intersection tested and the first 4 lane design, it unsurprisingly did rather well. what i was not expecting was a score of 209
Do you happen to still have the blueprint string for this?
