Efficient LOADING/UNLOADING Train Station designs | Factorio Tutorial/Guide/How-to

[@Nilaus]

All Master Class Blueprints are available on FactorioBin Overview and direct links to all Blueprints: nilaus.atlassian.net/l/cp/HBEUm524 (Pastebin links no longer work)

[@tricorius9653]

“Don’t worry...I’m going to show something that is better than whatever you have later on.” Hahahahahaha. Ah Nilaus. You make my day with your snark, man.

[@HobbitGaming]

"I'll show you how to make one better than anything you have..." I nearly spat my coffee 😂

[@futurevendetta]

Glad you commented, else I would've said the same. Savage and yet so so true haha

[@gamefeed6134]

I have never thought about just having 4 chest unloading. It is an interesting concept and maybe you could mirror it and combine the two sides. Simple(maybe) and effective since we would have 6 [boxes] in and 8 out

[@LeSethX]

Honestly, that simple setup is better than what I have for my train unloading station...

[@gamefeed6134]

@@LeSethX same. Everything in his videos seem to be

[@MrLordZenki]

I watched a megabase tour where someone was using 4-100-4 trains. The idea was to more closely mimic the size of trains in real life. It was pretty amazing to behold.

[@ИльяВитцев]

omg link pls

[@MrLordZenki]

@@ИльяВитцев kzbin.info/www/bejne/sIO8gpaMl8RojZI

[@ИльяВитцев]

@@MrLordZenki thanks

[@maxwilson7001]

Who did it?

[@nakoda1610]

Cool idea But werent the trains super slow?

[@disjustice]

The advantage of unloading faster than you need (e.g. emptying into 48 chests and outputting on 4 blue belts) is that the train gets on its way faster, so it can return faster. This can be helpful if your outpost is far away. You'll need fewer trains to supply your base. It's not a huge issue, but it is a consideration.

[@calbsmoth4866]

@@karlsmink and higher UPS

[@dmikalova]

This actually becomes less and less relevant the farther away your stations are - for example consider if you could load/unload 5 seconds faster. If the station is 10 seconds away, then that loading savings is a 33% decrease in non-loading time. If your station is 45 seconds away, then it's only a 10% decrease.

[@mobius1qwe]

@@dmikalova Now forget about distance and just think about the ammount of inbound trains into your base. Forget that you not have just only 1 train from just only 1 outpost feeding just only 1 base station, but instead you are a regular factorio player that makes 10 outposts with at least 20 trains running around carrying stuff and while 1 train is unloading you have at least 3 waiting on the stacker. See if what the ppl in this part are commenting has any meaning or not.

[@kollegeknusprich8110]

I think there's another (albeit similar) point to be made about this. If we fully consume the belts on which we unload, and if we do so consistently (e.g. in a smelter), then we need to unload (and buffer) stuff faster than we can transport it by belt. If we don't, then the belts will run dry while the current train is leaving the station and a new one is not yet stopped. Even if the trains are stacked really well, there will be downtime in between where there's no train that's currently unloading. Edit: After some experimenting, scratch that (mostly). Four stack inserters per Wagon (at one blue belt per wagon) seems to be sufficient. The inserters filling the belt work slower than the inserters unloading the wagon, as a whole stack can be put into the buffer chests at once. So while the original point still stands, there is a natural buffer in the chests which may or may not be enough until a new train starts unloading.

[@robbert-janmerk6783]

@@pulsefel9210 Well, see it as a informative and practical lesson of the bottleneck concept :)

[@DavidLindes]

For future reference: while I don't at all mind that you didn't "go into details" (34:16) on the circuit stuff, it would have been nice to have had at least a quick visual on what those combinator settings were (each * -1 -> each, I presume?), for those who don't need an explanation per se, but who are trying to just take in exactly what's going on. If you're willing, please. :)

[@misterpanel9722]

I need help! Im making a Oli loading and unloading station and i don't understand the circuit

[@DavidLindes]

@@misterpanel9722 the circuit you made, or one from here? It's been 2 years since I posted that comment, so I don't quite remember what it was referring to... I imagine some factorio forums might be a better place to get help? Anyway, I can't help based on what you've said so far... if you have a more specific question, I _might_ be able and willing to, but I probably won't be fast about it.

[@misterpanel9722]

@@DavidLindes i don't understand circuits very well i.g. but i was refering to the one on the video

[@DavidLindes]

@@misterpanel9722 well, Nilaus doesn’t actually expose what this one even does, so I’m afraid I can’t help. I could probably guess, or figure it out in-game, but yeah. Sorry.

[@misterpanel9722]

@@DavidLindes well shit

[@bbrcher]

I totally get not going into the circuits in depth, but as someone who watches these days/weeks before I get a chance to play with it, just having the info box up for a few frames would let me see the circuits-I know them well enough to already know basically what you did from your verbal description, but confirming it with sight would be nice. That said, great explanation/reasoning on “big picture” planning over localized optimization. “Stay effective”

combinators oil * -1 -> oil, pumps >= 0,

[@gatekeeper258]

@ omg this hurts. Lol why multiply by -1???

@@gatekeeper258 to subtract. x - y = x + (-1 * y)

[@hakimahmad55555]

@@gatekeeper258 don't know if you figured this out yet, but the pumps actually sum the signal amount if entity is the same, thus the negative check, and multiplication by -1.

[@BEN-ys6gu]

​@ Nice trick: Instead of multiplying stuff by -1 you can substract it from constant value 0. Although I'm pretty sure this super small processor level optimization would't help with UPS

[@Jugglerman]

Yes this is what I needed. Doing a train world and trying to figure out train balancing.

[@Naruedyoh]

Just yesterday i started with trains. I was hit by a train like 10 times

[@jonilarsen-haikarainen8733]

You can actually build a guarded train crossing that will save you. (but damage the train if it is to close to stop when you cross) with walls, gates, signals and some wires.

[@mikepotvin668]

trains, the number 1 killer in facorio.

[@williamlooney6242]

We need a Nilaus' Master class on how to get hit by a train :)

[@masterphillips]

We call those 10 free pistols around here.

[@tricorius9653]

I’d suspect Nilaus could do a pretty killer Masterclass on being murdered via train. :)

[@beanstalk1309]

I love these explanations- I am not adept enough at this game to design these yet but being able to pause, rewind, and understand why something works this exact way is super helpful - been a follower for about a year and it's really upped my game!

[@brentcrobarger7872]

If I'm understanding correct your argument is that since you can only load a train at rate X you only need to unload a train at rate X or faster. But my train has multiple multiple mines. I need to unload as fast as possible because there is always another train waiting behind it to unload.

[@mobius1qwe]

THIS! Players don't have only 4 lines of ores into their smelting districts, they have 16, 32, 48 depending on the player. Restraining a station to only output 4 lanes no matter what while your base needs more so you make more stations is just a waste. Make an unloading station with much throughput as possible and diverge more trains to it, this is the true UPS preservation. Less overall entities in the world.

[@okuno54]

@@IngieKerr If it's not being consumed further down, the correct solution is _the factory must grow_! In any case, that wasn't the question. At 15:55 Nilaus argues that there is no way (and therefore no need) for an unloading station to receive higher throughput than a single loading station. Of course, this is plainly false: a single unloading station served by trains which load at different loading stations requires throughput equal to the sum of the throughputs of the loading stations. Indeed, this situation is common in real play: you will likely have several ore outposts all feeding a centralized smelter or other production facility.

[@vholes2803]

I think you're gonna need a bigger train...

[@NotBamOrBing]

19:20 fucked up how you're having the furthest one input to the closest belt instead of the closest one doing that

[@cshairydude]

For the fluid stations, the pumps feeding/draining the central tank use the same calculation, but inverted. So you only need one combinator for both, just set the pumps to opposite conditions.

[@seanpeacock4290]

or no combinators, just use both colors of wire for different color signals. if red > than green then left pump active and reverse it for the green pump. I do something similar for the heavy to light oil conversion and light to petrol. it may not be optimized but it prevents shutdowns from a full light oil and no petrol > plastics > red chips > blues chips. its not perfect but it means that if I run out of gas I've run out of everything.

[@Oberon4278]

@@seanpeacock4290Oh man, I was having the hardest time last night trying to figure out how to manage my petroleum output and I think you just gave me the solution. Thank you so much!

[@sirosis7858]

Watched the whole thing. Informative and WELL PRESENTED. The author doesn't try to act cutesy, goofy, or ham it up. The info is presented logically sequentially and is explained at each step. Good stuff. Subscribed.

[@BillCarlson]

For anyone who was struggling with weird issues using the fluid unloader blue print with LTN - the left and right tanks are hooked to the circuit using RED wire. Swap that for GREEN wire, then you can use red wire for each of the tanks (except the middle one) to combine and find the total amount in all of the tanks.

[@spc-tr]

at minute 25 you do the math for ore saying we need a train every 15 seconds, but, if we're talking about green circuits, the train only needs to come every 1 minute, since a circuit train carries 4 times the amount of items when compared to an ore train. So the "theoretical" maximum kind of becomes practical depending on the type (compression) of the material we're carrying!

[@csabanagy1879]

it does not matter how much the train can carry, what matters, that how fast you can remove those items from the storage. Even the item can hold more compressed items the belt cannot transfer more items. As Nilaus showed the 12 compressed blue belt is the limit. If the trains arriving faster the station will fill up.

[@melind82]

@@csabanagy1879 The number of items in the train is the same as how compressed the item stacks are (assuming full trains). If you need to move 4x as many items it will take 4x longer.

[@AlonCunningham]

This is exactly what I needed to learn, My trains were not loading and unloading efficiently, lol I'm just dreading the circuit network now. Thank you again

I prefer using circuits for item loading too as they fix themselves when something screws up. If anyone is wondering, here's how: connect all boxes together with green wire and send it to arithmetic combinator where it's divided by the number of boxes and outputs signal A which goes out with green wires to all inserters. Then each inserter is connected to it's respective box with a red wire, the condition on the inserter is item

[@grifferz]

I am surprised that Nilaus neglected to mention this MadZuri average contents method as it's very widely used. At first I thought it was because he wanted to avoid mentioning circuits, but then circuits were used on the fluid load/unload. EDIT I see in a later comment he said that he does not use it because it killed performance (UPS) in a larger base. Still, it is so often suggested as a solution that I think it merits a mention even if only as a warning about UPS.

@@grifferz oh, it has a name? I came up with it independently on my own. Nice!

[@Derzull2468]

@@grifferz While I like the MadZuri design it also drastically reduces the throughput, if you want to consume the belts asap, it's not really an option.

[@suit1337]

i've tried this, but it is slower as with splitters - to make it work, you need to add the current inserter Stack size to A to make it work, else the inserters will not grab items even if they could

@@suit1337 I've sometimes added a detector of full belts that just allows all inserters. If the belt is not full anyway slowness doesn't matter.

[@matthewjenkins914]

Love that idea for the computationally-efficient loader, whoever came up with that is a genius.

[@jeremysaklad6703]

The design I came up with for loading stations is somewhat similar, insofar as it uses 1-4 trains with 24 stack inserters and chests on one side. I use the same 4-belt balancer to distribute items to each set of six chests, but I just use three splitters to break each belt into six from that point forward. I use circuitry to disable the belt leading up to each chest if the chest in question has fewer items than the average across all chests, which means it actually adjusts itself if you suddenly have way more items in one chest for some reason. Wiring is actually free if done via blueprint, and the entire thing only needs one arithmetic combinator. I use basically the same trick for unloading, except I disable inserters if the corresponding chest has *fewer* items than average. I never disable anything actually interacting with a train, so turnaround is extremely fast.

[@NuriTANRIKUT]

great tutorial. a suggestion to make it easier to see: connect nixie tubes to chests/tanks to see how much in them while showing they are balanced

[@PURASAPECAGEEM]

I have like 50 hours in this game just learning things by myself and each step of the way I always thought "oh fuck this game can get really complicated". Today, I had the impression I may not be using trains properly and for the first time I came to KZbin to learn how I could do it more effectively. Now I know I'm like a newborn trying to take a quantum physics exam. But very helpful videos man, thank you.

[@StormCrow42]

With filter stack inserters, a little wire, and a constant combinator you can design an easily adjustable station that can help prevent mis-delivery accidents contaminating your base. I'm also curious if you considered the MadZuri style circuit controlled inserters for balanced loading and unloading.

[@ritinfek]

Indeed. A comparison which Madzuris loader would be really interesting

[@mobius1qwe]

Zuri uses combinators to control the inserter -> train part of the station, that hurts UPS by a lot. The idea is to make sure you will always feed the boxes evenly without combinators at all whenever possible.

[@bulldogg4]

I really wish you had gone through the circuit logic for the balanced fluid loading and unloading. You normally go into so much detail for topics it felt weird to see you just plop a blueprint down and not really have an explanation for how to set it up ourselves.

[@psychedelic951]

yes. i need to build this in my train megabase, too. so a little explanation would have been great. i think i have to figure it out myself which can take hours LUL

[@Nilaus]

Look at the runtime of the video. It is too long and adding 5 min of circuit would be detrimental to objective of the video. I do understand the request though 😏

[@actualperson1971]

My understanding is that there's one arithmetic decider for each pump. When the input-side tank is more than the output-side tank, the pump runs, vice versa. Pretty easy network tbh...

[@psychedelic951]

@@actualperson1971 yeah i figured it out already, was really not as hard as i thought 😂

[@mattc9683]

@@Nilaus Could you at least add the fluid station blueprint you used to the description? (edit: nvm figured out how blueprint importing actually works lol)

[@embrace_yt]

Finally, been waiting for this for sooo long

[@reda29100]

Having recently done a megabase, the consideration for even distribution across the same wagon is not really concern to be bothered about, even with heavily modded belts, wagons, miners and inserters. Let's first calculate how many trains a 6-chest station can buffer. 6 chests with 16 stacks (worst, 64 ideal) with 40-stack wagon gives 2.4, 4.8 and 7.2 trains. With more wagons comes more chests that don't decrease this number. What this means is, for whatever reason, gridlock or halt of any sort, this mine will function as to let the next 2.4 trains fully ready to be loaded right after the train issue or base consumption gets resolved. I haven't played vanilla purely put I can hardly imagine even with worst fuel types for a train to take so long this long before it arrives back from base or wherever this output is supposed to be unloaded at. And even then, stacker at unload and at load would solve this speed issue. Or used more chestsnif you will. A different type of advise would be, pollution is cool but have you tried limiting you chests for one train only? 40/6=6.666=7 stacks ideally for 1 train if your base is well managed for output and signals. Or 13.333=14 stacks (still manegavle by wooden chests) for 2 trains, or 20 stacks for 3 trains. Steel chests costing 8 steel *6*4 for 4-wagon train *10 iron per steel = 1920 iron plate. For a marathon run, wooden chests become a base savior even with highly functional base. True, wooden is burnt bit steel is kep later for logistic chests, iron is such a precious and valuable commodity that I don't enjoy "wasting" easrly on, on chests. And wood is almost abundant easily on. Coming to inserters, Wiki reports 0.83, 2.31 turns per second,and you can test that btw. Meaning, for normal inserters with stack size 1 has maximum 498 item per minute (assuming homogenous input, i.e., inserter does not pick 2 of item A (when it is waiting for item number 3 if you have stack size of 3 that is) then comes only item B where it waits and wastes time.) Keeping in mind belts have 15, 30 and 45 i/s, they have 900, 1800 and 2700 i/s. Slowest inserter is 49.8 and 138.6 turns/s. Meaning, 900/(50i/s) is 18 inserters per belt (more than 6 per wagon), or 6.52 (7) fast (filter, stack and filter stck have the same turning speed. Difference being stack size only.), or 3.26 (4) with stack size 2. You can do the math for you level of technology but I think only this is relevant, with 40/(138.6*12)= 1.62 stack inserters at max bonuses. I think this proves that for vanilla belts and inserters, if you're concerned about balancing this station, tbe issue is output consumption, not even distribution. Plus, all these 6 insertersns load a single wagon, which means if output is fine, next train is gonna balance it off and take the rest (which is very likely faster than belt throughput anyway.) And if the throughput itself is low, it really doesn't really matter how you put it inside the chests, plus the train would first load then buffer would build up, which is once again not an issue to be bothered about because that's the best that station can offer.) So my suggestion would be, as long as you balance wagons, then either limit your chests (especially for pollution concerns in early-mid game) or build stackers on input (i.e., mine) and output (base or smeltry or whatever uses this resource) if you have some issues in train network or are repurposing or expanding your base/smeltry.

[@sink53]

Hi. Thanks for the Master Class playlist it helped a ton and still do. I have less than 100 hours and thanks to the main bus video I went from tasty spaghetti that somehow makes a rocket into a scalable base where robots are actualy usefull. Never even used have Mk2 Power Armor and Robots properly before the lastest playtrough. (In the games I finished I never had portable fusion reactors and ran all games full coal steam with very low throughput) Back to the point, when it comes to unloading trains, I focked around and found out, at my own expenses, that running 6 or 8 lanes ot even more out of a drop station is not a good idea unless you balances all the lanes out. I use a bus of 4 lanes wide and I stumbled upon the issue that if a single train stop fills more than 4 lane, and thoses are not balanced, you will have a situation like 4 lanes are completly empty while 4 others are completly staturated, but the empty lanes will never get refilled because the unloading station has too much ressources (using LTN the calls for refill are not generated) That makes me wonder what is better, either putting more train stops or balancing the lanes out. As balancing the lanes with vanilla splitters becomes exponentially more complicated and space consumming the more lanes there are, I think going for more stations is a safe bet. I think I'll go for 4 lanes per import station. If you drop off the ressources on a single side you can have a very compact design hence cramping way more train stops. With 1/3/1 trains sounds like that will work nice.

[@MoonLiteNite]

Thank you patrons! And OMG 112k views??? I haven't watched any of your videos, been off factorio for about a year, since your death world series. Glad you took the plunge into full time youtube/twitch!

[@garyseal6721]

I like using trains, but do struggle. Was kool the loading and unloading. I have dabbled with LTN, just started over a base in 1.0; noticed some different items have need introduce. Would be really good to see one of your videos on LTN. Your explaining and getting me back through watching your 1.0 play through was a great help. Also these tutorial one fantastic. Cheers Gary

[@onemerlin]

At 4:50 or so, you say that you're limited by the belts, not the inserters. This is only *usually* true. As you note, the belt speed is 4*30 = 120/second. With 4 cars and 6 inserters each, you have 24 inserters. To allow your inserters to keep up with your belts, you need each inserter to sustain 5/second. However, inserters are spec'ed per swing, and a wiki check shows 2.3 swings/second. This means that blue inserters will be slower than your belts until your inserter capacity bonus is up to +2 for non-stack, which I think is level 8. At level 0/1, with no bonus, 24 blue inserters can only sustain 55 items/second; with a +1 bonus at level 2, they sustain 110 items/second - just barely too slow. Green inserters get to pick up even before bonuses, so will always be fast enough. But you need significant stack bonuses to sustain more than 4 red belts, and your favorite unload configuration with 4 inserters per car needs an even higher bonus to not be inserter-limited. (I can see this watching your trains unload on the deathworld; the belts pulse with the inserter swings and thus can carry more than the inserters put down.) To fill a red belt with 4 green inserters, your stack size needs to be 8+; to fill a blue belt with 4 inserters, your stack size needs to be 12+. (edit: typos)

[@Nilaus]

4 stack inserters with capacity 8 can completely saturate a blue belt

[@kevingriffith6011]

This is a massive step up from the circuit monstrosity I constructed to evenly load/offload trains. All this averaging, comparing, machine toggling madness... I think there are a total of 14 different logic machines for each train.

[@dmbadcat]

I keep coming back to this video... it's just too damn cool! Thank Nilaus for providing so many years of great content!

[@dmbadcat]

(though it would be cool to see more circuit designs)

[@tomaszromanowski9153]

Thank You for that video! loading and unloading designs are awesome but talking about using "full belts" was eye opeing for me. I had one attempt in transision to a mega base (following your guide from like 8 months ago) where I tried using 8 wagon setup and i now noticed that i was unloading to 8 blue belts, and all factories required at max 4. So now i can reduce the numer of them and make my builds more compact. Same thing with loading stuff, i was using 8 blue belts and output of a factory only covered 2 blue belts. Now, off to remodeling that stuff.

[@rhueoflandorin]

very good material, Nilaus. You are a true master, but I would like to add something you failed to mention.... a caveat of sorts: There is a difference between TRAIN LOADING speed (or train-exit speed), and PRODUCTION vs CONSUMPTION speed...both aspects are things to consider for whether you "require/should use" blue inserters or stack inserters on a single side or both sides..... You can have 20 ore outposts each only producing a few red belts of ore but there is VALID reason to set up the outposts to load DOUBLE-sided onto the trains that arrive. Although the chests will fill SLOWLY, trains can still LOAD at max "stack inserter speed", and only be at the station for a few seconds, assuming circuit conditions are utilized and you have multiple mining outposts for the trains to choose from.... Although GLOBAL consumption CANNOT exceed the total production across all outposts or the train loading speed will suffer... there still exists several factors that exist that give benefit to loading trains double-sided, one additional consideration is PARKING spaces...by setting up the stations to load as fast as possible, you reduce the need for larger stackers. But you must utilize circuit conditions in order to manage this properly. The rate trains can exit from any particular station will still be limited to the INPUT for that station, but by having MORE outposts than delivery spots (or just...more mining capacity than the smelters require), you can MINIMIZE the time the trains require to RE-LOAD at these outposts by "going ahead" and plotting down the double-sided loading station blueprints. This also helps future proof for the point where you begin adding speed modules and gaining more mining-productivity.

[@Clarke2316]

Absolutely what I needed!! Last few days I’ve discovered your channel, will be sticking around for sure.

[@MichaelCater]

Thank you for the video. This helped me with moving oil from a significant distance to my starter base.

[@gregd5569]

@Nilaus I'm following your base in a book series and I think I found an optimisation to the liquid loading station. When pumpjacks get depleted and flow lowers so that trains arrive when tanks are not full, I found that the tank wagons 1 and 4 are filled much quicker than 2 and 3. This is I think because the middle tanks have 2 outputs (wagon and the last tank) so the flow is halved. I changed the ratio at which the last tanks are filled by changing the "* -1" on the combinators to "*-10" and adding another combinator between the middle tanks and pumps with a "*5". Now all wagons are filled at a constant rate. Thanks for the vids.

[@ChucklesMcGee780]

nilaus, sir, you are the king of factorio. love the content man, please keep up the good work!

[@disjustice]

Great video! I've never tried balancing my liquids like that. I'm gonna start now!

[@figa5567]

that loading station override of the stack size trick was amazing. love it. Makes every station cheaper too, just as an added boost! But I'm not sure I got the stack sizes right. Are they 12-6-6-6-6-4?

[@ИльяВитцев]

yes, with 4 being at the most compressed part of the belt

[@Alex-xv6me]

I was surprised when you didn't mention Madzuri's balanced loading station. The liquid balancing was new information to me! thank you

[@mickaelmalove2553]

Thanks for providing the free factorio lessons :) I would have liked if there was a link or something to a more in depth view of the logic circuits being used for the fluids train. Much like trains, logic circuits are one of those parts of the game I don't feel like I grasp as well as I should to really master the game. I can manage a very simple check like an overflow to my light/heavy oil cracking, but I don't think I've ever used a combinator that wasn't part of some pre-fabbed blueprint.

[@MountHyjal]

This was perfect timing for me. I was just starting to struggle with my loading and unloading.

[@joe78man]

Neat trick for loading/unloading materials, I have a blueprint for them, full of balancers. To the toilet with them. For liquids I use 1 pipe that divides into 3 (I use 3 wagon trains) into Storage tanks and then pumped into the wagons, It works fairly well for me

[@oxion.]

For those using the liquid unloader: The pumps enable condition is set to water; You'll need to change it to "everything" (the red * symbol) to properly balance other liquids

[@deanlawson6880]

Excellent! Very cool video on train stations!! I have only been playing for a couple months now, and am in my 2nd run-through, this time with Krastorio 2 mode (and a few helper mods). My Train stations have gone through a pretty startling evolution and I currently have around half a dozen styles and types of loading and unloading styles for dry goods as well as Crude oil. I'll try out these designs, but honestly I'm a little baffled by that weird splitter arrangement for the loading stations. In my personal experience, I've found that a "stack" of 3-2-1 splitters works excellently and ensures even loading to all of the boxes for a cargo car on a train. But I'll try out this odd splitter arrangement. Thank you for providing blueprints so we can easily "print-in" these designs and try them out quickly! Thanks for the videos Nilaus!! Cheers!!

[@Naomitage]

Thank you so mmuch Nilaus for this Masterclass Videos... Trains were a nightmare for me for like forever but I'm slowly starting to get an Understanding of them thanks to this Videos

[@YammoYammamoto]

To the point. Effective. No nonsense. Thank you!

[@vladdembrovskyi9062]

Brilliant!! Love the tricks for ups-optimisation. It also saves lots of belts and splitters and looks simpler and easier.

[@theothercreare]

Thanks, I didn't know where to start with efficient stations When I get time ima watch the rest of the videos in your masterclass series

[@Hedning1390]

4:40 That calculated time is the time between the arrival of one train to the arrival of the next. I'm sure this is what you meant, but it's not what you said. You still want the actual loading to be faster, because the next train may not teleport in at the station. 6:21 Loading chests unevenly on the same wagon is absolutely no problem. If the capacity is needed it will be used. You do not need to waste time and resources building a splitter to load them evenly. If for example only 5 inserters are filling the train because the last chest cannot be filled, and this means loading the train is slower than the capacity the belts can provide, then the first chests will gradually fill up each train until it is full and then the last inserter will get access to ore and the train after that will have all 6 inserters loading it. 17:20 I was going to explain the thing here, but you used the same concept for your 6 belt design, so I'll just let you know that you can use the same trick for 6 to 1 belt. Though in practice the belts should stop every now and then, because it is better to have a machine stopped because its output is full than having a machine not work because it lacks inputs (well at least for ore since miners have varied outputs over time). When the belt stops the little hole will be closed. Also there is no point using 6 inserters to one belt so using a 6 to 1 belt merger is just in case you want to use a standard 6 chest design everywhere. 30:00 I think everyone knows to use pumps to increase fluid flow. However you don't need combinators. Setting the pumps to pump if previous tank is > 20k (for loading example) is good enough. The tanks are holding the same as a wagon, so you just want to prevent the tanks stealing capacity from a neighbor.

[@pmkaboo2446]

4 chests with 500 ores will fill the wagon way faster than 1 chest with 2000 ores, which is why uneven loading is in fact a problem.

[@Hedning1390]

@@pmkaboo2446 The train cannot take more items than the belts supply. If only 1 inserter is working then that means 1 inserter can empty the belt. Can't do any better than empty belts.

[@pmkaboo2446]

​@@Hedning1390 youre right, but only in situation where your buffer chests are always empty, or youre not using them. if you can collect 2k ore before the train arrives, then its better to have it in 4 chests than in 1.

[@Hedning1390]

@@pmkaboo2446 It will collect in as many chests as is needed to clear the belt. You do not need to worry about how many that turns out to be. Slowing down an inserter like he is doing just so that more inserters have to work to do the same job is the stupidest thing I've ever heard of. Then of course it is bad to have huge 2k stacks, because it creates a lag to the base while it builds up, but that can be solved much more efficiently by using wooden chests or limiting the steel chests.

[@pmkaboo2446]

@@Hedning1390 while limiting the chests does have the same effect, you need to be very precise with it, otherwise youll end up with uneven stacks. the limits will also require adjustments whenever you add new trains or miners run out of work.

[@malcolm_ocean]

I'm at 9:22 watching you lay out a mirrored version by hand and I'm like "bruh do you not know the G hotkey?" F flips horizontally and G flips vertically!

[@JETZcorp]

"Don't worry, I'm going to show something that's better than whatever you have..." Nilaus has the only accent that can make everyone believe a statement like this without question.

[@Baer1990]

for the unloading I always use 4 boxes in a 110011 pattern (Instead of 101101) I can put the first belt right next to the track and it will make it very compact

[@rctimmers]

But what about adding more pumps to pump the oil in or out of the fuelwagons? I think 3 per wagon is possible.. isnt that much faster?

[@tapiopuranen88]

For loading, balancing between the boxes, while neat and cool, is not really essential because as you mentioned earlier the main limiting factor is the total belt throughput.

[@gemstonegynoid7475]

can be considered that youre just replacing a length of belt with a train network. if the input is 4 belts full, it'll equal 4 belts at the output.

[@gamefeed6134]

The way I think of trains especially when I see a post online where someone is asking throughput of a train i always think ‘well if you get 4 belts(any) in and they don’t get backed up due to a full box then you get the source throughput. (Btw when writing this comment I was at 15:11)

[@minameismud83]

The fluid unloader station doesn't seem to be working. I placed it down from the blueprint with no modifications, and it's not hooked up to anything. I start with a 1-4 train fully loaded with oil and a completely empty unloader station. Train pulls up to the station, pumps all do their thing, then stop. From front to back, my fluid cars are: 1. empty 2. 5.4k 3. 5.4k 4. empty. The tanks are (same order) 1. 5.8k 2. 25k (full) 3. 25k (full) 4. 25k (full) 5. 5.7k. Middle to cars still have fluid. It's almost as if the balancer pumps shouldn't activate unless the train is completely empty, or maybe there needs to be backflow pumps. Not too familiar with what you can do with these circuit networks (or fluids or trains really), but this definitely doesn't work, and it's consistent. I've run this multiple times with the same result. Oh Grand Master @Nilaus, can you help?

[@stenoaux]

I've had similar experience. @Nilaus any tips?

[@minameismud83]

@@stenoaux Since I posted this, I've found that the orientation of the fluid station matters. I made my own circuit-driven fluid unloader designed for the right side of the train and place it twice - once with the train pointed north and once south, and in one case the rear two cars will unload first while in the other the front two cars will unload first. This was in 1.0 stable. Fluids are funky.

[@stenoaux]

@@minameismud83 I poked around with the circuits and noticed the blueprint is hard coded for a water signal. I changed that to the “everything” signal and it seems to be working better. Have you tried that?

[@SurfingZerg]

33:00 has anything changed since this video? I use the blueprints provided, and yes I did change it from water to oil in the pump inteface, but it just doesn't balance at all for me

[@pauljs75]

I tend to be more biased towards the buffering capacity at a station than total optimal throughput. As long as it doesn't run out before the next train shows then most of the time it's good. (Also dumbs down a bit what has to be designed for.) Combine that with something like a minimum 3-train stacker with full trains just before an unloading station, and you don't have to worry about resource demand at that station exceeding any delay by trains transiting to the resource locations. And to avoid any possible traffic jams, a simple "round robin circuit" based on a counter is put on such stackers to take care of queuing when a train leaves the drop-off. (I like a stacker setup as a particular kind of standby station vs. the usual, as they can be placed anywhere nearby the drop-off on the rail system rather than having to be just before it. But requires wiring stuff to tell when trains depart a drop-off. I suppose LTN may achieve similar without the wiring stuff?) Of course such isn't quite to the megabase level where people seem to worry about UPS, that I see elsewhere in comments. So the approach used may be a case of YMMV.

[@sullivanmkii]

To get a better understanding of the fluid mechanics myself, isn´t the decider thing somewhat of a performance drain? Recursively checking every unit of fluid twice, interrupted by updates for pumping it from/to the wagon. Wouldn´t it be better to rely on the automatic distribution between tanks and maximize input to the tank-filling pump?

[@justaweeb1884]

For the train liquid unloading station, the tanks in the middle are full but the tanks on the left and right are not and the 2 middle fluid wagons does not empty

[@bignose140]

So the fluid stations.... Do I just hook them up the same way I would a soild to an LTN station? With green wire to read the contents.... Wont that interfere with the pump's signals?

[@melind82]

Cool video, I'd definitely be down for another discussing station logic. I've starting using stackerless designs and am pretty happy with it. They don't scale to megabase speeds, but as long as you are ok with a train every 1-2 minutes per station they get the job done and are way less overhead.

[@JohnMeinel]

I've always used circuit based loading/unloading to help prevent inequity when you don't have full consumption/production. (eg one side is pulled more than the other side, you don't quite get fully compressed belts in, etc.) Essentially the same thing that you did for liquid loading/unloading but for chests. (red wire to box, green wire gives the negative average, and it only swings when this chest 'more full'.) I don't know that it is inherently better, it does have its own quirks. I'm also curious on the liquid loading/unloading, in that you can have 2 pumps per wagon, each one unloading into a separate storage tank (with the tanks joined on the opposite side). I can see a point that the total throughput isn't necessarily improved as it is rate limited by whatever output pipes you are doing, and how much you consume. It has been useful in modded play when you have much higher throughput. (The most annoying part of modded play is that pipe throughput is very hard to understand.) I really like how everything lines up nicely, though.

[@erikderfreak]

9:15 couldnt you use the Flip key to mirror it, instead of building it new? Or am I mistaken in that?

[@VenomVsTRex]

Excellent video, Nilaus. I eagerly await your circuit master class because it's basically Eldritch magic to me and I want to learn how to do fancy cool stuff like the big boys. Keep up the outstanding work.

[@CaptainLsc]

You should use F5 to show the grid. that'll help for placing as well as for us to count spaces :)

[@SublimeCS]

Oh shit just in time. I'm starting to reach the stage where I need trains.

[@jucom756]

This video made me realise i have an unloader design that has 16 belts capacity and actually even more if i would divide it, but i use it on 4 belts and that's good because having 16 belts from 1 station would become quite crammed so i'd rather have 4 stations that don't need a train every 30 seconds.

[@nickschaus7785]

30:20 The oil goes in your pockets obviously!

[@jonathanlanglois2742]

There is a simpler design possible for your tank setup. Simply connect all of the tank together and have a single combinator divide the total content by the negative amount of tanks you have with a red wire. The result is the average amount of oil you have in your tanks. From there, connect your pumps directly to your pipeline. Connect all of the pumps with a green wire to their respective tank. Connect all of the pumps to the output side of the combinator with a red wire. Your pumps are now getting the amount of oil from their respective tank on the green wire and a negative average amount of oil on the red wire. Setup your condition if oil < 1 then enable. Your pumps will turn on or off based on the average amount of oil in their respective tank. This system is easily scalable and also works for balancing a regular cargo station.

[@rysiii811]

liquid loading seems nice, but unloading isnt really necessary imho, its quite hard to use this much oil (for me at least), but if you just do a few tanks (one would work too) to buffer the oil with pumps sucking it out to larger oil tank field (oil tanks are super cheap and I like to keep a stable storage not to run short when the field is running low). You need as many pumps, you need no logic for that and while Im pretty sure that you lose some efficiency, do you need the 100% efficiency in oil?

[@EpicSpaghetti]

8:00 blew my mind! Thank you!

[@brendonhowbert9330]

Thank you. I do a lot of testing myself but this is one I didn't want to do.

[@memberHD]

"I'm going to show something better then you have" made me laugh. It's so true

[@usevaladgaponenka773]

Awesome guide for liquids especially

[@VacentViscera]

"Before I can show you the solution, I have to show you the problem. If I show you the solution without showing you the problem, you will not understand." Truer words have never been spoken. So many people care so much about "perfection" that they neglect the fact that any solution is replacing something else and that all things that exist to solve a problem solved it "well enough" for someone. Don't let the perfect be the enemy of the good. Also, don't twist my words to make it seem like I was endorsing something horrible. :P

[@TrabberShir]

This entire series is amazing. But, on the loading station I am not sure why you would use the splitter design after having discovered the stack size tweak? The top design is lower lag, and uses considerably fewer resources, and is faster to upgrade from red to blue, and can be tweaked for yellow if for some reason you want to do that. It also has what I consider to be one of if not the most important feature of a factorio design: you can see "at a glance" if something is broken. The splitter design hides most of the length of the belts in undergrounds so that a small mistake causing gaps in the belt or reduced flow to one car is hard to see.

[@seanpeacock4290]

My normal settings map has trains coming from all over the map to a central station from ore patches of different sizes, so my priority is to unload the train as fast as possible to clear the terminal for the next train of that type. Loading the trains takes very different amounts of time and I don't like having one unload station for each train. Some of the trains have very long paths to get to their destinations so even slow load times, yellow inserters to full train slow, are still the shortest part of the journey. I don't use anything slower than blue inserters for trains but you get the idea. of course now I want to see if I can mine an entire coal patch without electricity and what that would look like.

[@zeroqp]

Love you, Nilaus

[@americankid7782]

So your saying that with 2 stations I can get 24 compressed blue belts as long as I have enough trains and mining stations? I am about to create a logistical nightmare. Or get 24 belts with ONE 2-8 train? Then I can just double it whenever by adding extra stations? I’m sorry but I now need to make a central smelting district specifically for this.

[@MrMartinSchou]

For balanced in- and output to trains, you can use stationary cargo wagons instead of chests.

[@Nilaus]

no!

[@mikepettengill2706]

WOW, I learned what I needed to know watching this one.

[@odw32]

No MadZuri loader? I like the circuit approach because 1. small footprint, 2. cheap materials, 3. disturbances are rebalanced automatically, 4. can be easily adjusted/augmented for mixed contents

[@fastfiddler1625]

Ok, dumb question. I've been playing this game a couple years, with not nearly the amount of time and expertise as others. What's the reasoning for needing to do a loading balancer that way rather than just a 1, 2, 3 pyramid of splitters fed by one belt. After drawing it out and mapping the paths, is it because the first belt splitter creates a bottleneck of throughput if it is only fed by one belt?

[@fantafuzz162]

The point of using the balancer is to make sure that each of the outputs get 1/6 of the input. If you use a pyramid as you suggest for a 1-6 split the outer four lanes get 12.5% of the input, while the inner 2 get 25%. This is because the last 3 splitters do not get equal inputs. The first splitter in the pyramid outputs two lanes of 50%. Each of the second level splitters take one input of 50% and spits it into 25%. Now, by placing 3 splitters on the level below, the leftmost and rightmost splitter will take one input of 25% and split it to 12.5%, while tSo he middle splitter will have two inputs of 25%, and split those into two outputs, keeping the 25%. Adding the middle splitter on the lowest level of the pyramid doesnt actually do anything, as the input and output side of that single splitter is the same. A bit rambly explanation, but if you try it out ingame, you can clearly see the innermost two belts of the 6 outputs get double the amount each than the four outermost belts. This is fine if you don't care about uneven loading, but in order to load each chest evenly, you need a proper balancer.

[@fastfiddler1625]

@@fantafuzz162 No, rambly is good. I just drew out each type on paper with colored lines delineating the paths and I'm mad I didn't figure it out myself. I hadn't previously drawn attention to that fact. That in the balanced version, all but the first have one input, so equal. I get it now. Thanks!

[@NameNotAlreadyTaken2]

The trick at 22:33 to get compressed belts by setting the inner inserters to stack size 8 doesn't work for me. There's still always huge gaps. I'm not sure what it is you're doing to get a compressed belt.

[@theblackdragon7321]

Do you have a vid on how you set up the loading unloading logic circuit? like how many trains get called for a station and all the item requests?

[@Nilaus]

Many-to-many Train Network

[@theblackdragon7321]

@@Nilaus thank you

[@Villain_in_Glasses]

I liked most of this but I have no idea how that oil station works, coulda done with going over the calculations.

[@dnkbmc]

Hmmm...Why not just calculate the avarege amount of fluid and enable each pump when the amount is above average (for unloading) and below average (for loading)? That way you only need one combinator + as a bonus you will be able to output your station contents to the circuit (which is very tricky with your setup, since the signals mix)

[@civishamburgum1234]

I would likw to advocate for iron chest as station buffres over steel chests, since iron chests have (to my experience) more than enoughh capacity to fill up multiple trains, so steel chests end up just being a recource sink and you don't want to have dead material lxing around in some buffer spaces unnessecarily, especially when you move higher up in the production chain.

[@songs2424]

I have a complaint you can loaf and unload faster for liquids the max speed for it is 12k fluid per second per pump and you can stac 3 pumps per car maxing out at 36k fluid a second

[@jeffreyhightower5781]

Alright I've spent literal hours trying to set up these liquid stations and I can't seem to get it right. On the loading station, I keep filling up the center tank, and none of the pumps turn on. No issues on the unload station that I can see. My settings on the loading station: All wiring connections are identical to the video, so I know wiring is not an issue. Pumps: Anything > 0 Combinators: Input - Each * -1 Output - Each I've tried swapping the pumps to be less than 0, no dice. It seems to me that, because no comparison's are made with the center tank, once the other tanks empty out, there's literally no comparison being made anymore and the pumps will never turn on. What's the way around this? I'm friggin pulling my hair out trying to figure this out.

[@Timo4eus]

I don't know if you even can distribute evenly output from 6 boxes to 1 belt, because it converges to 3 belts, and balancers are 2 to 1, and divide 3-factor number by 2-factor number is overcomplicated.

[@512k]

I use parallel (2or4)stations to feed 36 lines with 1-3-1-3 trains powered by coal. kind of brut force solution. But Issues arise when there are too many trains, stations and intersections. Just start to implement logic solve some issues and add new one...

[@dkluempers]

In the immortal words of Mr. Burns "Excellent".

[@yanworkt]

Nilaus, why ever spread between boxes are important if we are bound by incoming belt speed? These uneven load is still fast enough. And compact. Or is it?

[@screamharsa]

Nilaus can u make explanation on IVTN? Or probably implementing it on your base so i can have the idea of how it works?

[@kaige498]

I love you and what you do Nilaus

[@Po-AnChen]

oh wow, I didn't know the lequid can't be balance by itself, thanks for the important information

[@SylvanasWindrider]

Nilaus, your videos are great!

[@wr0mb3l99]

I've just recently spent hours designing my trainstops with LTN in mind. I'm using 8 stack filter inserters per wagon (4 on each side so that the trainstop in total only has a 10-tile footprint) and manged to balance the loading using circuit network. I also want to build a megabase, so would you say figuring out stack limits for the inserters would be better?

[@actualperson1971]

Yes, I'd say so. Circuit networks do take some CPU to figure out

[@chris909090]

I use circuit balanced loading a lot and have been thinking about whether I want to pivot over to the stack size override technique, and where I’d want to do it. One of the big benefits of the circuit approach is that it always works regardless of conditions. The stack size override will only work (properly) if your belts are fully compressed, 100% of the time. So I’m probably going to be using a mix of techniques.