3 and 4 way intersections

Smart setups of railway stations, intelligent routing, solutions to complex train-routing problems.
Please provide - only if it makes sense of course - a blueprint of your creation.
Avona
Long Handed Inserter
Long Handed Inserter
Posts: 61
Joined: Sun Dec 26, 2021 8:50 am
Contact:

Re: 3 and 4 way intersections

Post by Avona »

FuryoftheStars wrote:
Wed Jun 01, 2022 2:04 pm
Erm, ok. I'm a native English speaker and I've always understood RHD/LHD to be the side of the road you're driving on, not the steering wheel. Dunno, maybe I learned differently? I don't have anyone to compare with.
I am also a native speaker but don't recall using RHD before Factorio. Talking about which side of the road we drive on wasn't a common subject for me. A user on discord had been suggesting RHT/LHT for years at that point so Hans brought it up. Hans, Kano, and I had a very long conversation about this and after reviewing the evidence, switching seemed simple and persuasive.

mmmPI
Smart Inserter
Smart Inserter
Posts: 2675
Joined: Mon Jun 20, 2016 6:10 pm
Contact:

Re: 3 and 4 way intersections

Post by mmmPI »

farcast wrote:
Wed Jun 01, 2022 2:05 am
I've done it! I've made a working junction out of scribbles!
This is art !
Avona wrote:
Wed Jun 01, 2022 10:47 pm
FuryoftheStars wrote:
Wed Jun 01, 2022 2:04 pm
Erm, ok. I'm a native English speaker and I've always understood RHD/LHD to be the side of the road you're driving on, not the steering wheel. Dunno, maybe I learned differently? I don't have anyone to compare with.
I am also a native speaker but don't recall using RHD before Factorio. Talking about which side of the road we drive on wasn't a common subject for me. A user on discord had been suggesting RHT/LHT for years at that point so Hans brought it up. Hans, Kano, and I had a very long conversation about this and after reviewing the evidence, switching seemed simple and persuasive.
I've used RHD = RHT = people drive on the right side of the road, exclusively in the context of factorio, without ever knowing i was doing a mistake and i've used it a lot :roll: but i am convinced too now.

Factoriointersection
Burner Inserter
Burner Inserter
Posts: 9
Joined: Sat Sep 25, 2021 4:58 pm
Contact:

Re: 3 and 4 way intersections

Post by Factoriointersection »

farcast wrote:
Wed Jun 01, 2022 2:05 am
Abstract 4-lane 3-way
Abstract in action.gif
I've done it! I've made a working junction out of scribbles!
Nice man, I have put it on the page, got the same scores as you.
Tested with 5.1.1

User avatar
hansjoachim
Fast Inserter
Fast Inserter
Posts: 239
Joined: Wed Apr 26, 2017 7:03 pm
Contact:

Re: 3 and 4 way intersections

Post by hansjoachim »

The goal was to make the highest throughput 4-way unbuffered intersection. This intersection now holds the highscore for 4-way unbuffered intersections.
It looks larger than the others, but when Including the safe outputs it really isn't.
set 1: 52
Set 2: 41
Set 3 46
Average: 47
https://factoriobin.com/post/-UFOZDli
Image

ElderAxe
Fast Inserter
Fast Inserter
Posts: 129
Joined: Thu May 18, 2017 8:04 am
Contact:

Re: 3 and 4 way intersections

Post by ElderAxe »

I have tested a slightly modified version of my "Almost squareabout OCD" blueprint with testbench 5.1.1. It would be nice if you update the post with the new numbers. Also, Darkelder is my previous nickname. It should be ElderAxe as well.
almost-squareabout-ocd.png
almost-squareabout-ocd.png (2.08 MiB) Viewed 6121 times


Almost squareabout OCD Results
almost-roundabout.png
almost-roundabout.png (508.98 KiB) Viewed 6121 times
-------------------------------------------------------------

Also designed another intersection which has slightly better throughput.
It would be great if you can add this to the list as well.

Name: Expanded Straights
expanded-straights.png
expanded-straights.png (2.01 MiB) Viewed 6121 times


Expanded Straights Results:
run-2.png
run-2.png (503.42 KiB) Viewed 6121 times
run-3.png
run-3.png (504.82 KiB) Viewed 6121 times
run-4.png
run-4.png (530.98 KiB) Viewed 6121 times
-------------------------------------------------------------

Since these numbers are better than Christmas BP, I did a test with Christmas as well and its test results came up better than listed.

Christmas Results:
christmas-1.png
christmas-1.png (507.98 KiB) Viewed 6121 times

Avona
Long Handed Inserter
Long Handed Inserter
Posts: 61
Joined: Sun Dec 26, 2021 8:50 am
Contact:

Re: 3 and 4 way intersections

Post by Avona »

ElderAxe wrote:
Wed Jul 06, 2022 1:53 am
It would be nice if you update the post with the new numbers.
Hello ElderAxe!

I tested both of your intersections as well as Christmas as you had and all of my results were lower than yours with 5.1.1. Since our results vary by about 3 tpm, I thought you might want to see how I tested, so I included it in the spoiler below.
Testing Setup and Results
I've updated the spreadsheet and OP.

ElderAxe
Fast Inserter
Fast Inserter
Posts: 129
Joined: Thu May 18, 2017 8:04 am
Contact:

Re: 3 and 4 way intersections

Post by ElderAxe »

Thanks Avona.

The only difference I've noticed in our setups is that I had full-length train spacing on entrances and exits.
I did another test with 2 car spacing signals on the entrance and the numbers were lower than the previous. Putting full-length spacing on the entrance significantly boosts Set 3 numbers.
run-6.png
run-6.png (561.42 KiB) Viewed 6060 times
Also, the testbench map I downloaded is tiled with gray lab floors. It's not grass with the grid as your SS shows.

This is the save I tested with:
Throughput tester v5.1.1 spacing-4.zip
(2.4 MiB) Downloaded 138 times

Avona
Long Handed Inserter
Long Handed Inserter
Posts: 61
Joined: Sun Dec 26, 2021 8:50 am
Contact:

Re: 3 and 4 way intersections

Post by Avona »

ElderAxe wrote:
Wed Jul 06, 2022 11:27 am
Also, the testbench map I downloaded is tiled with gray lab floors. It's not grass with the grid as your SS shows.
Yes. Changing the tiles does not change the version, which is 5.1.1.

Avona
Long Handed Inserter
Long Handed Inserter
Posts: 61
Joined: Sun Dec 26, 2021 8:50 am
Contact:

Re: 3 and 4 way intersections

Post by Avona »

Well I finally got around to messing with the high throughput merges that Hans used in his latest buffered intersections. I took the Parallel Multicross and wrapped it with the lengthy merges to get a TPM of 112 for RHT/LHT. It's technically the same 250x250 but each arm is thicker.

Image

I also tried a smaller Hurricane with the merges but it only got 107 (as well as a version with the Echo Triple which got lower) so I switched to the multicross.

Zaflis
Filter Inserter
Filter Inserter
Posts: 414
Joined: Sun Apr 24, 2016 12:51 am
Contact:

Re: 3 and 4 way intersections

Post by Zaflis »

How does a pair of T-intersections actually compare to compact 4-way intersections? Mainly thinking 2-lane (1 rail per direction). There is 1 such design in the buffered category but none in the compact.

User avatar
hansjoachim
Fast Inserter
Fast Inserter
Posts: 239
Joined: Wed Apr 26, 2017 7:03 pm
Contact:

Re: 3 and 4 way intersections

Post by hansjoachim »

Zaflis wrote:
Wed Jul 13, 2022 6:26 pm
How does a pair of T-intersections actually compare to compact 4-way intersections? Mainly thinking 2-lane (1 rail per direction). There is 1 such design in the buffered category but none in the compact.
They perform the same. But if you place the unbuffered 3 ways close to each other, =<2 train lengths, an unbuffered 4 way performes better.
For buffered it's a different story. The 4 way wins since the lanes connecting the two 3 ways becomes a bottleneck.

https://www.reddit.com/r/factorio/comme ... urce=share

FireDuckz
Manual Inserter
Manual Inserter
Posts: 2
Joined: Fri Jul 15, 2022 12:32 pm
Contact:

Re: 3 and 4 way intersections

Post by FireDuckz »

4 Way 2 Lane Unbuffered intersection.

It uses LHD on the vertical axis and RHD on the horizontal axis

I know, it's cursed

It should have a pretty decent throughput tho.

Image
Blueprint:

Code: Select all

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

Zaflis
Filter Inserter
Filter Inserter
Posts: 414
Joined: Sun Apr 24, 2016 12:51 am
Contact:

Re: 3 and 4 way intersections

Post by Zaflis »

FireDuckz wrote:
Fri Jul 15, 2022 12:40 pm
4 Way 2 Lane Unbuffered intersection.
You can optimize the east and west going exiting rail signals by replacing the chain signals with them, otherwise it should run good.

Afternote, i think it's not vertically symmetric. Try placing the blueprint over itself after rotating it 180 degrees.
Edit: I see what you mean by "cursed" :P There is no symmetry to be found.

FireDuckz
Manual Inserter
Manual Inserter
Posts: 2
Joined: Fri Jul 15, 2022 12:32 pm
Contact:

Re: 3 and 4 way intersections

Post by FireDuckz »

You can optimize the east and west going exiting rail signals by replacing the chain signals with them, otherwise it should run good.

Afternote, i think it's not vertically symmetric. Try placing the blueprint over itself after rotating it 180 degrees.
Ye I know, I posted it on reddit, and someone made a new more symmetrical design, that also has better signaling due to it having more space

AristomachosCZ
Burner Inserter
Burner Inserter
Posts: 12
Joined: Mon Apr 19, 2021 9:04 am
Contact:

Re: 3 and 4 way intersections

Post by AristomachosCZ »

Hello,
it seems like the 4-way/2lane/compact unbuffered/4. Super compact Celtic knot doesn't have a RHD with 6 spacing, so there it is.

Score is: 1) 50,40,45 = 45 ; 2) 51,40,42 = 44 ; 3) 50,41,43 = 45 (testing version 5.1.2, testing included safe exit block).

https://factorioprints.com/view/-N7f-xEMfLRZ7csac09q

AristomachosCZ
Burner Inserter
Burner Inserter
Posts: 12
Joined: Mon Apr 19, 2021 9:04 am
Contact:

Re: 3 and 4 way intersections

Post by AristomachosCZ »

Also there is a version of previous intersection with separated turning lanes – I wouldn't mark it as buffered (because these separated lanes can hold only one train per direction, so I don't think it's a buffer), but it can be a subject of a discussion – and it has a score 59,45,50 => 51 (average values from 5 tests).

Blueprint of this intersection (4-way, 2 lanes, RHD, 6 spacing) here: https://factoriobin.com/post/wJA55Pls

Avona
Long Handed Inserter
Long Handed Inserter
Posts: 61
Joined: Sun Dec 26, 2021 8:50 am
Contact:

Re: 3 and 4 way intersections

Post by Avona »

AristomachosCZ wrote:
Sat Jul 23, 2022 2:19 pm
I wouldn't mark it as buffered
Not only is it buffered, it's also not compact. It's 118x118. The current smallest buffered intersection is the Tiny Starfish at 82x82, scoring 57, which is still large when compared to compact intersections. Would you like me to test and add it as the lowest scoring 2 lane 4 way buffered intersection?

I will also test and add your 6 tile space RHT celtic. I didn't notice your posts yesterday but I'll get to them tomorrow or the day after.

User avatar
Tallinu
Fast Inserter
Fast Inserter
Posts: 129
Joined: Sun Jun 14, 2015 8:14 am
Contact:

Re: 3 and 4 way intersections

Post by Tallinu »

Hey guys, long time no visit! Nice to see this stuff is still alive and kicking in a new home! Some really interesting new designs in the list...
Avona wrote:
Tue Jul 12, 2022 9:58 pm
I took the Parallel Multicross and...
I had worked on a more parallel version of the Multicross myself (as well as a "turbo" circuit-controlled exit-merge compacting system, but it was really finicky, nowhere near as elegant as the loopy thing in that Michigan Left video a few pages back). I thought I had posted it in aaargha's old thread, but I never really focused on it because I think I was running head-on into the throughput limits of the old test map. And it just seemed too bulky for too little gain compared to the two-lane Multicross. Avona, your expanded versions look downright epic, I'm quite impressed!


I noticed in the current listing here for my 2-lane Multicross, the "expanded" version is lumped in with it on the same entry and even though it had a significantly higher score, the score for the smaller version is what's listed on the heading, and determined its position in the list. Now part of the design goal with those was that you could mix the two as desired to make a hybrid that had different amounts of buffering on different entrance/exit legs (and even an unbuffered segment for if you had to connect a lower-traffic outpost really close to the intersection and didn't want to waste all that space and resources on a full-size segment just to connect to that one thing). So listing the different blueprints together is great, but at the same time it was mildly saddening to see it listed only in the lower-scoring location in the list. ;)

Also, I made so many different size variations for longer trains, it seems a shame to not have those available. Would you be willing to add a (current) link to my drive share folder that has all of those? (I still get emails occasionally from people clicking on the old "no longer valid due to security update" link from the OP in the old thread, even though I put an updated link in my signature quite a while ago. I guess I'll have to unshare it and make a new one to get that to stop. Considering it was set to allow public viewing, and it clearly still associates the link with the right folder, I don't know why it has to pester me to manually allow access every single time...)

hansjoachim wrote:
Wed Jun 08, 2022 7:55 pm
The goal was to make the highest throughput 4-way unbuffered intersection.
Really throwing symmetry to the wind there, eh? Wonder how much of its high score is attributable to that... :D

Avona
Long Handed Inserter
Long Handed Inserter
Posts: 61
Joined: Sun Dec 26, 2021 8:50 am
Contact:

Re: 3 and 4 way intersections

Post by Avona »

Tallinu wrote:
Fri Jul 29, 2022 12:57 pm
I noticed in the current listing here for my 2-lane Multicross, the "expanded" version is lumped in with it on the same entry and even though it had a significantly higher score, the score for the smaller version is what's listed on the heading, and determined its position in the list.
...
Would you be willing to add a (current) link to my drive share folder that has all of those?
I'm glad you like my expansions of your multicross. I'd previously used your intersection and it's what got me into designing intersections! I actually felt a bit excited that the Tallinu responded to me lol... :P

I believe the lower scoring multicross was used because it is 262 x 262, very large sized in other words, compared to 378 x 378, astronomically sized. I thought that no one would use the expanded version, since I don't recall anyone using the regular sized MC other than myself. Considering that it is a 11% increase to throughput compared to a 44% increase to the length of the intersection, the utility seemed dubious. You can see in my parallel MC, hurricane, and starfish intersections that I shoot for as small as possible with the hope that people won't have difficulty fitting them into their factories. But alas, not many use buffered intersections lol...

That was the logic anyhow. That's what we decided in your absence, but you're the designer, and I'm quite happy to flip the listing if that's what you want. I'm sorry for the dismay.

Is this the link that you want associated with the multicross listings?: https://drive.google.com/drive/folders/ ... RQD_SExbhW

User avatar
Tallinu
Fast Inserter
Fast Inserter
Posts: 129
Joined: Sun Jun 14, 2015 8:14 am
Contact:

Re: 3 and 4 way intersections

Post by Tallinu »

Glad to hear it, I got my start on these in basically the same way! :D

Your logic is quite reasonable, and I was mostly joking anyway. I wonder if you could put in a header at the 92 position for "MultiCross Expanded: 92" with a "See below" link to the spot the existing entry is in? That way the two sizes are still listed together, since they can be used together.

As for the length, I figured people using even a smaller version of buffered intersections like these would most likely be running something like RSO, or at least expanded Railworld or other non-default ore patch settings where distances would be longer anyway, so there'd be a good chance of having room for the longer version in some cases. And keeping them linear made them easily adjustable for different train lengths, unlike the loopy-swirly style designs. Also I tended to make use of the space surrounding the intersection for solar farms, which is a lot easier when there's a nice right-angle corner to tuck the blueprint into! ;)
Avona wrote:
Fri Jul 29, 2022 8:12 pm
Is this the link that you want associated with the multicross listings?
Didn't realize there was a different link available for the subfolders. But yeah, that one will work. I'm fine with the way the four listed blueprints are linked to factoriobin. Just wanted to have a link added below those to the "full listing including larger train sizes" or something to that effect. (And yes, those really do get ridiculously long! I didn't bother scaling up the "expanded" blueprint for many of the train sizes, it just got a bit too ridiculous.)

Post Reply

Return to “Railway Setups”