Nova_CI have the needThe need for speedRegistered Userregular
edited July 2020
Artillery is pretty rad, as it turns out.
And using my rail network is a no brainer, although it's nothing like Big Bertha. Just a single artillery car with a locomotive. I built forward stations that have a small defensive perimeter made up of two layers of laser turrets surrounding the station. It handles the biters and spitters no problem, so that pretty much nullifies the aliens. Whenever the turret runs out of ammo, it automatically returns to where it's being made to refill and then returns to station.
I'm about to start the process of building rocketry stuff. I think that's going to be complicated, but I'm at the last phase of the game. I'm at 68 hours on this playthrough. >.>
Edit: The three reactors are juuuuuuust enough to meet my power needs, so I'll have to do an upgrade before progressing. I think I'm going to completely redesign my reactor layout and set up a way to do progressive builds instead of the kind of haphazard layout I have now.
Nova_CI have the needThe need for speedRegistered Userregular
I live to serve.
I was up too late last night sorting out my reactor build. I never actually tested the intermediate steps, so I don't know how well it works as a progressive built, but I basically started with a four reactor design, saved the blueprint, and then worked backwards to have three, two and one reactor designs that require no changes, just additions, to upgrade each step.
But just before I quit for the night, I switched over to the new reactor build and decommissioned the old one. So I'm up to about 460 MW of power. I hope that's enough? It's almost double my previous output, so I'm guessing.
In terms of blueprints, I have some pretty solid building blocks for a hypothetical next playthrough, but my oil processing facility is a mess. And I'm not really sure what to do with it.
Maybe that'll be a project for tonight. Develop a tileable oil processing facility that hooks into my rail factory build.
Nova_CI have the needThe need for speedRegistered Userregular
Just in time to put it down while they release 1.0, my rail factory has launched the rocket.
And if you want to see it with the absolute craziness that is all the stations:
It is very much what I wanted to do, and the granularity of it made logistic robots superfluous. There was no point at which they made sense for anything except the powerplant. And even there, that was just for convenience more than anything.
Some of the problems that I ran into: While scaling things up was easy in most cases, low demand stuff was actually a problem. A single rail factory module could produce enough red computer chips to supply plenty of places, but then a single train running back and forth between the production area and all the delivery areas took time, and if one place went into high production for a time, it might run out, so large buffers were required. And that meant overproducing a metric fuckton of things. With more planning, I think I could do it a lot cleaner, but I would need to sit down with a pencil and a notebook to work out a more efficient system and ehhhhh, I do enough work at work. :P
My Green and Red Chip production modules:
And if you're curious, my powerplant (Turns out 450 MW can easily take you into and past the endgame):
I still had the old coal powerplant there as a backup just in case, but at four reactors, I'm not sure it would have achieved much if it needed to come online.
Oh, and you'll notice all those little outlying stations that look the same? Those are all artillery stations. A single artillery train set to go reload itself automatically when it runs out of ammunition. I arrayed them around at the end there to stop the biters from building any bases near me. The automatic function of the artillery would take out anything close and it worked well. At the end my evolution factor was at something like 0.96 or so. For the past few hours before launching the rocket, I didn't have to think about the biters even once. Double layered laser turrets guarding the artillery trains, each with dedicated repair robots and a supply of replacement parts, there was nothing they couldn't handle.
I will say, there were some times during my final expansion that got interesting. Taking out the larger nests would sometimes bring a massive surge of biters. If I had left it to the automatic defenses alone, they probably would have been fine, probably, but I definitely helped out to keep too many buildings from being wrecked.
Next time I do this, I'll probably go into creative mode first and work out the blueprints. There are some problems that were just not worth solving by the end with throughput, especially the green computer chips. I had six of those modules you see in the spoiler above. If I had kept playing, and needed to produce every science item type at once? I think it would have been fine for the first bit, but once all the buffers got low, I think my factory, sprawling as it was, would have been brought to its knees. I think the chokepoint would have been just getting enough copper to enough assemblers to make green chips. That part of my design needs some work.
Anyway, this playthrough took a ridiculous 98 hours. A lot of time was spent fiddling with the rail designs. However, it is the kind of game I always wanted to get out of Factorio, so I'm glad to have done it. We'll see how long until I come back after 1.0, I suppose.
Apparently the only way to do the thing I wanted to do, have trains dump directly into factories to abolish belts, was mod every recipie to force factories to preload a full stack of resources. So I've just been leaving it up to do a lazy tech rush so I can design my blueprin
I'll probably go into creative mode first and work out the blueprints
.. Really?
*Googles*
Oh, ok. It's a mod. You're off the hook, arbitrary vision.
This makes me very happy, as none of my pre-0.18 designs seem to have survived. My train direct-injection plan was mostly based on discovering I was going to have to remake my station buffer BPs.
Creative mode is a mod, but it's just an enhanced sandbox mode. You can enable cheats in that and spawn infinite chests / power / loaders to test stuff out
Ouch only 24 assemblers making green circuits? Yeah that's not nearly enough, not even a red belt's worth
My biggest stumbling block was that the way I was handling mining and smelting required exponentially more organization as I progressed.
I had four copper smelting facilities. Each one had two stations taking in ore, and two stations outputting plates. Each item that required copper had a dedicated copper train, like the green chip module you see in my screenshots. That meant it was super, super easy to have too many trains going to one station, and then the buffer at that one would run out, slowing down all the trains for that station, while other stations sat with full buffers and empty stations.
I would have to work out max throughput demand, and balance it, for each train.
Whereas when you run just a shit ton of conveyors, there are elaborate load balancers that handle that for you.
EDIT: My #1 feature for them to add on post 1.0 is programmable trains or stations. That is, make it so that train routes can be altered based on conditions so that if a station's buffer is low, it can automatically redirect an incoming train to a different station.
Nova_C on
0
Options
Nova_CI have the needThe need for speedRegistered Userregular
Ouch only 24 assemblers making green circuits? Yeah that's not nearly enough, not even a red belt's worth
My biggest stumbling block was that the way I was handling mining and smelting required exponentially more organization as I progressed.
I had four copper smelting facilities. Each one had two stations taking in ore, and two stations outputting plates. Each item that required copper had a dedicated copper train, like the green chip module you see in my screenshots. That meant it was super, super easy to have too many trains going to one station, and then the buffer at that one would run out, slowing down all the trains for that station, while other stations sat with full buffers and empty stations.
I would have to work out max throughput demand, and balance it, for each train.
Whereas when you run just a shit ton of conveyors, there are elaborate load balancers that handle that for you.
EDIT: My #1 feature for them to add on post 1.0 is programmable trains or stations. That is, make it so that train routes can be altered based on conditions so that if a station's buffer is low, it can automatically redirect an incoming train to a different station.
If I remember correctly you could tie your buffer chests to the train station with the circuit network and turn OFF the station conditionally. Like say if the buffer chest has less than x of ore or whatever. So then the train will bypass that station and go to the next one on the list
Ouch only 24 assemblers making green circuits? Yeah that's not nearly enough, not even a red belt's worth
My biggest stumbling block was that the way I was handling mining and smelting required exponentially more organization as I progressed.
I had four copper smelting facilities. Each one had two stations taking in ore, and two stations outputting plates. Each item that required copper had a dedicated copper train, like the green chip module you see in my screenshots. That meant it was super, super easy to have too many trains going to one station, and then the buffer at that one would run out, slowing down all the trains for that station, while other stations sat with full buffers and empty stations.
I would have to work out max throughput demand, and balance it, for each train.
Whereas when you run just a shit ton of conveyors, there are elaborate load balancers that handle that for you.
EDIT: My #1 feature for them to add on post 1.0 is programmable trains or stations. That is, make it so that train routes can be altered based on conditions so that if a station's buffer is low, it can automatically redirect an incoming train to a different station.
If I remember correctly you could tie your buffer chests to the train station with the circuit network and turn OFF the station conditionally. Like say if the buffer chest has less than x of ore or whatever. So then the train will bypass that station and go to the next one on the list
Yeah. And/or you can also run one logistics network (I used green) thru your entire power network and use that to send signals to trains. You can have them with a standby station on the schedule and an AND condition pair like Wait 15 seconds AND logistics signal 5, at which point when the 5 lights up it'll proceed to its next station.
Ouch only 24 assemblers making green circuits? Yeah that's not nearly enough, not even a red belt's worth
My biggest stumbling block was that the way I was handling mining and smelting required exponentially more organization as I progressed.
I had four copper smelting facilities. Each one had two stations taking in ore, and two stations outputting plates. Each item that required copper had a dedicated copper train, like the green chip module you see in my screenshots. That meant it was super, super easy to have too many trains going to one station, and then the buffer at that one would run out, slowing down all the trains for that station, while other stations sat with full buffers and empty stations.
I would have to work out max throughput demand, and balance it, for each train.
Whereas when you run just a shit ton of conveyors, there are elaborate load balancers that handle that for you.
EDIT: My #1 feature for them to add on post 1.0 is programmable trains or stations. That is, make it so that train routes can be altered based on conditions so that if a station's buffer is low, it can automatically redirect an incoming train to a different station.
It's not the cleanest API in the world, but you can do this.
Trains ignore any disabled stops on their schedule, so you can summon and cancel trains by turning stations on and off. If you want the iron train to divert because you need an emergency delivery from the fuel train, you would need to get a little creative based on the physical layout, (like having one stop per resource, or a holding station so they wouldn't both divert when you canceled the iron train) but it's doable.
This seems like it would have problems at scale, but I think they can be addressed by having substations that pull from larger scale distribution centers, and which then supply the factories.
Also if you name every stop the same name, it’ll pick up from whichever of those is on. Very very useful for picking up from remote ore stations.
+1
Options
Nova_CI have the needThe need for speedRegistered Userregular
Hmmm, that actually would work. Name them the same, then any station whose buffer is less than one full train load gets disabled. It doesn't really matter which station the trains go to in that case.
However, there is another possible issue - queues. If, for some reason, four different trains pick the same station, they'll queue up instead of diverting. Or can that be handled?
Hmmm, that actually would work. Name them the same, then any station whose buffer is less than one full train load gets disabled. It doesn't really matter which station the trains go to in that case.
However, there is another possible issue - queues. If, for some reason, four different trains pick the same station, they'll queue up instead of diverting. Or can that be handled?
Ultimately, you control this by ensuring that routes to like-named stops that already have a supply train inbound are calculated as longer. A combination of rail network topology and managed train release should be able to address this with sufficient forethought.
Example: Factory A1 enables it's generic "drop off" stop, but also broadcasts it's unique id. The depot flips some switches to ensure that the A1 'drop off' is the shortest (or only) route, then releases one train and nullifies the signal.
Trains that cross a decision point between like-named stations will automatically increase the calculated length of the route, but you could further influence it by activating dud stations or forcing red signals behind them:
When the rail block is occupied by a train -> Add a penalty of 2 * length of the block divided by block distance from the start, so the far away occupied paths don't matter much.
When the rail block is guarded by a rail signal set to red by the circuit network -> Add a penalty of 1000.
When the path includes a train stop that is not the destination -> Add a penalty of 2000.
(Assuming the latter means active stops, so that 'Speedbump Station (East)' wouldn't effect pathfinding when it was off)
Me all day at work: I can't wait to setup some simple tests to see if all that train theorycraft I posted works *continues to brainstorm*
Me at home: Oh dear, my ostensibly temporary blue beaker factory is suboptimal!
I've been stuck in this cycle for days now. Starting to realize the only way I'm ever going to actually advance this factory is by completely tearing it down. I absolutely refuse to shoehorn pink or yellow beakers into it.
Me all day at work: I can't wait to setup some simple tests to see if all that train theorycraft I posted works *continues to brainstorm*
Me at home: Oh dear, my ostensibly temporary blue beaker factory is suboptimal!
I've been stuck in this cycle for days now. Starting to realize the only way I'm ever going to actually advance this factory is by completely tearing it down. I absolutely refuse to shoehorn pink or yellow beakers into it.
Refactoring is substantially easier once you've started landing worker bots.
Anyway, this playthrough took a ridiculous 98 hours.
As someone who played Seablock for quite while and is dipping into Pyanodon, I'm used to just getting to automated green science when I hit 100 hours. Automated blue if I'm really ambitious.
I've probably done 6 100+ hour maps of various Bob's/ Angelbob / Seablock.
My big Seablock (.14 - .17) has 500 hours in it and I pretty much stopped because of burnout.
(Edit - no right way to play if you are having fun, this is an incredible game)
Currently on Factory #4 here after 1-3 fell to biters. Natural Evolution and Rampant mods do *not* fuck around, and that's with some of the nastier mod features turned off.
Given that I play Factorio half as the world's weirdest tower defense game I kinda love it.
"This weapon is so devastating it may crash the game"
Never change, mod authors, never change. The minimap shows it best but this is someone going "No, I'd like the nukes to do a realistic amount of damage"
+4
Options
Nova_CI have the needThe need for speedRegistered Userregular
Hmmm, that actually would work. Name them the same, then any station whose buffer is less than one full train load gets disabled. It doesn't really matter which station the trains go to in that case.
However, there is another possible issue - queues. If, for some reason, four different trains pick the same station, they'll queue up instead of diverting. Or can that be handled?
So, the good news is that Factorio 1.1 is adding a feature that makes my modular rail factory scalable. When a train paths to a station, the station counts it as an incoming train. You'll be able to set a maximum of incoming trains before the station no longer accepts incoming trains. This solves all my problems with train management in my design.
I REALLY want to start a new game since I wrapped my last playthrough a week before 1.0 dropped, but now I have to wait for 1.1.
+5
Options
Nova_CI have the needThe need for speedRegistered Userregular
I upgraded to 1.1 experimental last night and hoo boy, Factorio really gets those hooks into you.
So, new map, Rail World settings with 600% starting area (So I can mess around getting my rail system set up without being harassed), but I've increased biter density and frequency to 150%, and turned on a very slow biter spread (Normally Rail World doesn't have any biter spread). Turned off cliffs.
I just finished off getting what red/green science gets you. I'm in the process of switching over to my modular rail base, so everything has kind of ground to a stop while I build out the basic infrastructure to get it rolling.
What I'm doing this time around is making the rails work more closely to how megabases operate in terms of layout. I'm also building around 4 car trains instead of two for the ore trains, but will be sticking with one car trains for the movement of intermediate products. I didn't use the circuit network at all in my previous build, and I will be this time.
My first iron smelting facility is set up with 8 columns of 12 steel smelters, but they're laid out so I can upgrade to electric without having to redo anything - just remove the extra belts for coal and extra inserters. I tested out train automation using both logic to say "This station is only active if the storage has enough to completely fill a train" (IE, ore mining station only turns on if the chests have over 8000 ore, iron plate station only turns on if chests have over 16000 plates), and have train space limits set, but until I have multiple mines or processing plants, that's just theoretical.
What I DO like, though, is how I can keep the coal moving with a single train, and without complex route plans. Coal dropoff stations all have same name, only turn on if coal in storage is below a certain threshold. This means that the coal resupply train will sit at the coal mine until a station turns on, essentially 'requesting' a refill. Then the train goes, unloads, then returns to coal mine to await another request. It's so elegant! I wish I'd done this sooner.
And with the train limit on stations, that'll work across the board for all the kinds of stations I'll have.
I've also changed how the trains go in and out of their modules. I don't want the trains crossing tracks because that caused the odd pileup, so I'm using offramps and on ramps, and then roundabouts for when the trains need to turn around.
I'm much happier with the improved layouts I'm going with this time, but it means none of the old blueprints are usable, and I'm really missing having robots. Need to get them unlocked asap.
+6
Options
Nova_CI have the needThe need for speedRegistered Userregular
edited December 2020
So, I have been building up the modular portion with an eye on more effectively leveraging the trains, which means larger modules.
In my previous playthrough, if you remember, a single 'installation' of green chip production was four assemblers making green chips. This time there's....enough to max out a yellow belt. However many that took (Can't remember offhand). My red chip installation has 22 assemblers making red chips. This means each module will push much more throughput, which is great!
But it also means that setup has been extremely slow since I'm building things at a level of resource consumption that is much higher than it should be considering my tech level. This is okay!
Except for the biters. Because I'm taking so long, even with their slowed expansion, they are moving in and my copper mine is right by a couple really big nests, so some pretty big attacks are hitting. I'm hanging in there, but I running out of time. Luckily, last night, I got blue science finally in production.
And not a moment too soon. Power issues are happening and I'm burning through coal at an alarming rate. I need to get nuclear online. I could just build an entirely new coal power plant, but building out enough infrastructure to run that only to switch to nuclear pretty much right after seems like that time and effort would be wasted. I'm at 80MW of coal power, and power demands are nearly over that, but I have now unlocked nuclear power. The task tonight is to build a uranium mine and go ham on nuclear.
I'll post some screenshots of my modules soon - I'm really liking my approach this time. Most of them looked pretty haphazard last time and I'm trying to make things more elegant this time around.
EDIT: Just remembered, the green chip module is actually producing more than a yellow belt - I had to upgrade to a red belt for the last four assemblers. Which doesn't actually matter because the inserters at the train stop aren't able to keep up yet. Probably will be able to with my next inserter stack size upgrade.
Nova_C on
+2
Options
Nova_CI have the needThe need for speedRegistered Userregular
I've decided to do something foolish.
I've got robots working, so it's not, like, crazy dumb, but
I want to change over my modular designs to work with the city grid concept so I can completely change over to using construction bots for everything. Right now I need to set up the robo port, access to mats and add robots before a new module placed by blueprint will construct. If I adapt the modules to the city grid, then I can stop doing that.
PROBLEM is that most of my modules won't fit cleanly, so they'll all have to be redesigned.
Well, Nilaus talks about it in a 30 minute video on Youtube, but it's basically a pattern of roboports and power poles that follows a grid and has 100% bot coverage. It's nutty what Nilaus gets up to, and I won't take it as far as he does (He extends it out so that all his mines fall within the grid, which is insane), but I'd like to use it for my actual manufacturing base.
+1
Options
IanatorGaze upon my works, ye mightyand facepalm.Registered Userregular
Well, Nilaus talks about it in a 30 minute video on Youtube, but it's basically a pattern of roboports and power poles that follows a grid and has 100% bot coverage. It's nutty what Nilaus gets up to, and I won't take it as far as he does (He extends it out so that all his mines fall within the grid, which is insane), but I'd like to use it for my actual manufacturing base.
For the benefit of all reading this thread, I believe this is the video mentioned:
Twitch | Blizzard: Ianator#1479 | 3DS: Ianator - 1779 2336 5317 | FFXIV: Iana Ateliere (NA Sarg) Backlog Challenge List
+2
Options
Nova_CI have the needThe need for speedRegistered Userregular
edited December 2020
So, I dunno if my block is the same size as Nilaus'. He uses the 100x100 square, I just used the 2x2 roboport logistics square, but it's the same idea - power poles in two layers, four roboports, and made so it can be tiled. I then design a single production to fit within the block.
Iron processing (Copper and steel are essentially the same):
Green Circuit Manufacturing:
Flying Frame Manufacturing:
As you can see, I'm still in the process of converting everything over, but the beauty is that I can tile out the basic power poles/robo ports and the construction bots will slowly built it out without me having to do any of it. I can then add rails using blueprints so, again, bots are doing it. And then plop down whatever manufacturing I need to expand. Bots will do it. All I need to do at the end is make sure stations and trains are programmed correctly and then done. I've decided to add in a coal station to every single block so that I don't have to worry if a train is getting fuel or not.
Oh, and I'm using the logic circuits to further expand on the station activations. In the Iron Processing block, there are four sets of four chests for loading trains with iron plates. Each of the sets of four is connected to a decider combinator that will output a 1 if the count is below 4000 (A full traincar), 0 if above. If any of the deciders are outputting a 1, it will disable the station. That way, the station is only active if there is enough iron to fully load a train, making sure it is in the station for the minimum amount of time and there is no waiting and allows the freedom of having a train that is fewer than four traincars if that is a better for whatever purpose the train has.
All of the pickups work like that, but stuff like circuits and frames I only use single car trains, so it's just a check at the station instead of needing the combinators.
And, of course, every station is limited to one train so there is never a traffic jam. A train is only moving if it has somewhere it can go.
Nova_C on
+4
Options
Nova_CI have the needThe need for speedRegistered Userregular
And rocket launched.
The blueprints continue to evolve, but I think I have it figured out. I'm still experimenting with managing traffic, but so far only occasional traffic snarls happen, and I think I've got it solved - just waiting to see if it happens again.
The mall, or the central base where all the placeable structures are made, is a mess. It predates using the city blocks, so before I retire this map, I'll do up a set of blueprints for a mall that fits in the city block paradigm.
I'm going to be pushing further out with this map as well. I want to see about maxing out a single rocket silo and that will probably take a lot more infrastructure.
I'm actually looking forward to a new start now that I have the full suite of blueprints ready (Pending the mall blueprints). Launching the rocket took 75 hours this time, but a LOT of that was spent experimenting with design and redoing things. My plan next time is to rush rail and construction bots and then just start dropping blueprints everywhere.
+2
Options
Nova_CI have the needThe need for speedRegistered Userregular
I did a new rail world game, and found a massive black hole in the mid game where transitioning from the start to my blueprints was really, really difficult.
However, this time I launched in 38 hours, so they still offered a pretty good jump. I went solar only and man. Nuclear is so much easier. Solar more the doubled the size of my base.
I think I'm going to take what I learned this time around and redevelop my blueprint library to offer more flexibility, as well as solving the mid game problem.
I did a new rail world game, and found a massive black hole in the mid game where transitioning from the start to my blueprints was really, really difficult.
However, this time I launched in 38 hours, so they still offered a pretty good jump. I went solar only and man. Nuclear is so much easier. Solar more the doubled the size of my base.
I think I'm going to take what I learned this time around and redevelop my blueprint library to offer more flexibility, as well as solving the mid game problem.
Mix it up with a challenge run. Logistic network embargo, Raining bullets, Steam all the way. I had a ton of fun with that.
I did a new rail world game, and found a massive black hole in the mid game where transitioning from the start to my blueprints was really, really difficult.
However, this time I launched in 38 hours, so they still offered a pretty good jump. I went solar only and man. Nuclear is so much easier. Solar more the doubled the size of my base.
I think I'm going to take what I learned this time around and redevelop my blueprint library to offer more flexibility, as well as solving the mid game problem.
Mix it up with a challenge run. Logistic network embargo, Raining bullets, Steam all the way. I had a ton of fun with that.
My rail network makes logistic bots superfluous - I did the logistic network embargo cheevo this run just....without really doing anything different. Raining Bullets I got as well - that one just required me to keep on top of the spawner nests. Steam all the way? Yeah. After this run, I think I'll never do solar again. It was so much more work!
I actually want to get the <15 hour to rocket launch cheevo. That, <8 hours, Golem and build 111 items or less are the only remaining ones.
Golem I'll probably get by accident at some point, but the other three...hm. Maybe <15 hours, but <8, and 111 items seem probably out of reach to me.
IanatorGaze upon my works, ye mightyand facepalm.Registered Userregular
Figured I'd post a bit on my current Krastorio 2 run.
I'm finally researching the Interstellar Transceiver, a power-hungry structure which serves as K2's victory condition. And by researching I mean, "It took me eight hours to make maybe 50% progress". My current base is very organic in structure, starting with a main bus, transitioning into a mix of Transport Drones and vanilla robots. There were a few AAI trucks in there too but I think they got hit by trains. What this all adds up to is a base that has some difficulty scaling up production and quite a bit of latency in meeting demand.
This world has islands separated by large water features so a city block megabase will require massive investment in Landfill. I may have to use a Transport Drones main bus instead.
Twitch | Blizzard: Ianator#1479 | 3DS: Ianator - 1779 2336 5317 | FFXIV: Iana Ateliere (NA Sarg) Backlog Challenge List
0
Options
Nova_CI have the needThe need for speedRegistered Userregular
I didn't get close to the 15 hour requirement for the achievement, but this playthrough saw me test and iterate into my final set of blueprint designs. I have a set to hopefully help transition smoother in the future to my rail megabase, but as far as all the modules go, they have been tested and all the bugs worked out. I think I even solved the traffic jam issues for heavy use roundabouts, but that one is harder to know for sure.
I even added a feature! I made circuit wiring part of my basic power block design, and most of the modules hook into that so I can set up an alert for when a certain resource gets low. It basically alerts me when there's a production shortage. Not super useful for the end game items (Though I did design their modules to send out a signal, I just don't bother making an alert for them), but for any of the raw materials, or first stage thinks like petroleum or green circuit boards, it's super helpful.
And I have a final version of my mall. It's a bit messier than I wanted because I worked to pack it all into a single block, but I got there in the end.
How I set up my raw material processing:
And the full megabase in map view:
Each module was made to be fully placeable in all four orientations, so my next playthrough, I'm going to use 3x3 grids and then use the middle section for modules that don't use trains. Right now that would just be power related (Nuclear plant, accumulator block, solar block if I use one).
+3
Options
Nova_CI have the needThe need for speedRegistered Userregular
Bah, still getting traffic jams. Hm, I think I'll have to try something different.
Realistically I don't think you can do the timed achievements with a rail focused base very easily. Certainly not the 8 hour one, that's very tight.
Until you need to do long distance belts can scale perfectly well and the extra fiddliness of trying to balance loads and set everything up eats up way more time than you can save. Maybe if you already have all the blueprints.
Usually I plop down a unified mall + all science blueprint and just feed everything into that if I want to scale up quickly
So The Riftbreaker just released where it's basically Factorio in a mech suit. Main difference is a slightly less complicated logistics portion and much more prevalent PVE portion, not to mention 1000x sharper graphics.
Posts
And using my rail network is a no brainer, although it's nothing like Big Bertha. Just a single artillery car with a locomotive. I built forward stations that have a small defensive perimeter made up of two layers of laser turrets surrounding the station. It handles the biters and spitters no problem, so that pretty much nullifies the aliens. Whenever the turret runs out of ammo, it automatically returns to where it's being made to refill and then returns to station.
I'm about to start the process of building rocketry stuff. I think that's going to be complicated, but I'm at the last phase of the game. I'm at 68 hours on this playthrough. >.>
Edit: The three reactors are juuuuuuust enough to meet my power needs, so I'll have to do an upgrade before progressing. I think I'm going to completely redesign my reactor layout and set up a way to do progressive builds instead of the kind of haphazard layout I have now.
I was up too late last night sorting out my reactor build. I never actually tested the intermediate steps, so I don't know how well it works as a progressive built, but I basically started with a four reactor design, saved the blueprint, and then worked backwards to have three, two and one reactor designs that require no changes, just additions, to upgrade each step.
But just before I quit for the night, I switched over to the new reactor build and decommissioned the old one. So I'm up to about 460 MW of power. I hope that's enough? It's almost double my previous output, so I'm guessing.
In terms of blueprints, I have some pretty solid building blocks for a hypothetical next playthrough, but my oil processing facility is a mess. And I'm not really sure what to do with it.
Maybe that'll be a project for tonight. Develop a tileable oil processing facility that hooks into my rail factory build.
I wish I was good enough to pull this off.
And if you want to see it with the absolute craziness that is all the stations:
It is very much what I wanted to do, and the granularity of it made logistic robots superfluous. There was no point at which they made sense for anything except the powerplant. And even there, that was just for convenience more than anything.
Some of the problems that I ran into: While scaling things up was easy in most cases, low demand stuff was actually a problem. A single rail factory module could produce enough red computer chips to supply plenty of places, but then a single train running back and forth between the production area and all the delivery areas took time, and if one place went into high production for a time, it might run out, so large buffers were required. And that meant overproducing a metric fuckton of things. With more planning, I think I could do it a lot cleaner, but I would need to sit down with a pencil and a notebook to work out a more efficient system and ehhhhh, I do enough work at work. :P
My Green and Red Chip production modules:
And if you're curious, my powerplant (Turns out 450 MW can easily take you into and past the endgame):
I still had the old coal powerplant there as a backup just in case, but at four reactors, I'm not sure it would have achieved much if it needed to come online.
Oh, and you'll notice all those little outlying stations that look the same? Those are all artillery stations. A single artillery train set to go reload itself automatically when it runs out of ammunition. I arrayed them around at the end there to stop the biters from building any bases near me. The automatic function of the artillery would take out anything close and it worked well. At the end my evolution factor was at something like 0.96 or so. For the past few hours before launching the rocket, I didn't have to think about the biters even once. Double layered laser turrets guarding the artillery trains, each with dedicated repair robots and a supply of replacement parts, there was nothing they couldn't handle.
I will say, there were some times during my final expansion that got interesting. Taking out the larger nests would sometimes bring a massive surge of biters. If I had left it to the automatic defenses alone, they probably would have been fine, probably, but I definitely helped out to keep too many buildings from being wrecked.
Next time I do this, I'll probably go into creative mode first and work out the blueprints. There are some problems that were just not worth solving by the end with throughput, especially the green computer chips. I had six of those modules you see in the spoiler above. If I had kept playing, and needed to produce every science item type at once? I think it would have been fine for the first bit, but once all the buffers got low, I think my factory, sprawling as it was, would have been brought to its knees. I think the chokepoint would have been just getting enough copper to enough assemblers to make green chips. That part of my design needs some work.
Anyway, this playthrough took a ridiculous 98 hours. A lot of time was spent fiddling with the rail designs. However, it is the kind of game I always wanted to get out of Factorio, so I'm glad to have done it. We'll see how long until I come back after 1.0, I suppose.
..
Really?
*Googles*
Oh, ok. It's a mod. You're off the hook, arbitrary vision.
This makes me very happy, as none of my pre-0.18 designs seem to have survived. My train direct-injection plan was mostly based on discovering I was going to have to remake my station buffer BPs.
My biggest stumbling block was that the way I was handling mining and smelting required exponentially more organization as I progressed.
I had four copper smelting facilities. Each one had two stations taking in ore, and two stations outputting plates. Each item that required copper had a dedicated copper train, like the green chip module you see in my screenshots. That meant it was super, super easy to have too many trains going to one station, and then the buffer at that one would run out, slowing down all the trains for that station, while other stations sat with full buffers and empty stations.
I would have to work out max throughput demand, and balance it, for each train.
Whereas when you run just a shit ton of conveyors, there are elaborate load balancers that handle that for you.
EDIT: My #1 feature for them to add on post 1.0 is programmable trains or stations. That is, make it so that train routes can be altered based on conditions so that if a station's buffer is low, it can automatically redirect an incoming train to a different station.
8 assemblers per module x 6, so 48, but still not enough.
If I remember correctly you could tie your buffer chests to the train station with the circuit network and turn OFF the station conditionally. Like say if the buffer chest has less than x of ore or whatever. So then the train will bypass that station and go to the next one on the list
Yeah. And/or you can also run one logistics network (I used green) thru your entire power network and use that to send signals to trains. You can have them with a standby station on the schedule and an AND condition pair like Wait 15 seconds AND logistics signal 5, at which point when the 5 lights up it'll proceed to its next station.
Trains ignore any disabled stops on their schedule, so you can summon and cancel trains by turning stations on and off. If you want the iron train to divert because you need an emergency delivery from the fuel train, you would need to get a little creative based on the physical layout, (like having one stop per resource, or a holding station so they wouldn't both divert when you canceled the iron train) but it's doable.
This seems like it would have problems at scale, but I think they can be addressed by having substations that pull from larger scale distribution centers, and which then supply the factories.
However, there is another possible issue - queues. If, for some reason, four different trains pick the same station, they'll queue up instead of diverting. Or can that be handled?
Otherwise you really need to be able to buffer a lot of trains
Ultimately, you control this by ensuring that routes to like-named stops that already have a supply train inbound are calculated as longer. A combination of rail network topology and managed train release should be able to address this with sufficient forethought.
https://wiki.factorio.com/Railway/Train_path_finding
Example: Factory A1 enables it's generic "drop off" stop, but also broadcasts it's unique id. The depot flips some switches to ensure that the A1 'drop off' is the shortest (or only) route, then releases one train and nullifies the signal.
Trains that cross a decision point between like-named stations will automatically increase the calculated length of the route, but you could further influence it by activating dud stations or forcing red signals behind them:
(Assuming the latter means active stops, so that 'Speedbump Station (East)' wouldn't effect pathfinding when it was off)
Me at home: Oh dear, my ostensibly temporary blue beaker factory is suboptimal!
I've been stuck in this cycle for days now. Starting to realize the only way I'm ever going to actually advance this factory is by completely tearing it down. I absolutely refuse to shoehorn pink or yellow beakers into it.
Penny Arcade Rockstar Social Club / This is why I despise cyclists
As someone who played Seablock for quite while and is dipping into Pyanodon, I'm used to just getting to automated green science when I hit 100 hours. Automated blue if I'm really ambitious.
I've probably done 6 100+ hour maps of various Bob's/ Angelbob / Seablock.
My big Seablock (.14 - .17) has 500 hours in it and I pretty much stopped because of burnout.
(Edit - no right way to play if you are having fun, this is an incredible game)
Given that I play Factorio half as the world's weirdest tower defense game I kinda love it.
Never change, mod authors, never change. The minimap shows it best but this is someone going "No, I'd like the nukes to do a realistic amount of damage"
So, the good news is that Factorio 1.1 is adding a feature that makes my modular rail factory scalable. When a train paths to a station, the station counts it as an incoming train. You'll be able to set a maximum of incoming trains before the station no longer accepts incoming trains. This solves all my problems with train management in my design.
I REALLY want to start a new game since I wrapped my last playthrough a week before 1.0 dropped, but now I have to wait for 1.1.
So, new map, Rail World settings with 600% starting area (So I can mess around getting my rail system set up without being harassed), but I've increased biter density and frequency to 150%, and turned on a very slow biter spread (Normally Rail World doesn't have any biter spread). Turned off cliffs.
I just finished off getting what red/green science gets you. I'm in the process of switching over to my modular rail base, so everything has kind of ground to a stop while I build out the basic infrastructure to get it rolling.
What I'm doing this time around is making the rails work more closely to how megabases operate in terms of layout. I'm also building around 4 car trains instead of two for the ore trains, but will be sticking with one car trains for the movement of intermediate products. I didn't use the circuit network at all in my previous build, and I will be this time.
My first iron smelting facility is set up with 8 columns of 12 steel smelters, but they're laid out so I can upgrade to electric without having to redo anything - just remove the extra belts for coal and extra inserters. I tested out train automation using both logic to say "This station is only active if the storage has enough to completely fill a train" (IE, ore mining station only turns on if the chests have over 8000 ore, iron plate station only turns on if chests have over 16000 plates), and have train space limits set, but until I have multiple mines or processing plants, that's just theoretical.
What I DO like, though, is how I can keep the coal moving with a single train, and without complex route plans. Coal dropoff stations all have same name, only turn on if coal in storage is below a certain threshold. This means that the coal resupply train will sit at the coal mine until a station turns on, essentially 'requesting' a refill. Then the train goes, unloads, then returns to coal mine to await another request. It's so elegant! I wish I'd done this sooner.
And with the train limit on stations, that'll work across the board for all the kinds of stations I'll have.
I've also changed how the trains go in and out of their modules. I don't want the trains crossing tracks because that caused the odd pileup, so I'm using offramps and on ramps, and then roundabouts for when the trains need to turn around.
I'm much happier with the improved layouts I'm going with this time, but it means none of the old blueprints are usable, and I'm really missing having robots. Need to get them unlocked asap.
In my previous playthrough, if you remember, a single 'installation' of green chip production was four assemblers making green chips. This time there's....enough to max out a yellow belt. However many that took (Can't remember offhand). My red chip installation has 22 assemblers making red chips. This means each module will push much more throughput, which is great!
But it also means that setup has been extremely slow since I'm building things at a level of resource consumption that is much higher than it should be considering my tech level. This is okay!
Except for the biters. Because I'm taking so long, even with their slowed expansion, they are moving in and my copper mine is right by a couple really big nests, so some pretty big attacks are hitting. I'm hanging in there, but I running out of time. Luckily, last night, I got blue science finally in production.
And not a moment too soon. Power issues are happening and I'm burning through coal at an alarming rate. I need to get nuclear online. I could just build an entirely new coal power plant, but building out enough infrastructure to run that only to switch to nuclear pretty much right after seems like that time and effort would be wasted. I'm at 80MW of coal power, and power demands are nearly over that, but I have now unlocked nuclear power. The task tonight is to build a uranium mine and go ham on nuclear.
I'll post some screenshots of my modules soon - I'm really liking my approach this time. Most of them looked pretty haphazard last time and I'm trying to make things more elegant this time around.
EDIT: Just remembered, the green chip module is actually producing more than a yellow belt - I had to upgrade to a red belt for the last four assemblers. Which doesn't actually matter because the inserters at the train stop aren't able to keep up yet. Probably will be able to with my next inserter stack size upgrade.
I've got robots working, so it's not, like, crazy dumb, but
I want to change over my modular designs to work with the city grid concept so I can completely change over to using construction bots for everything. Right now I need to set up the robo port, access to mats and add robots before a new module placed by blueprint will construct. If I adapt the modules to the city grid, then I can stop doing that.
PROBLEM is that most of my modules won't fit cleanly, so they'll all have to be redesigned.
I think this is the better idea in the long run.
But maaaaan
Well, Nilaus talks about it in a 30 minute video on Youtube, but it's basically a pattern of roboports and power poles that follows a grid and has 100% bot coverage. It's nutty what Nilaus gets up to, and I won't take it as far as he does (He extends it out so that all his mines fall within the grid, which is insane), but I'd like to use it for my actual manufacturing base.
For the benefit of all reading this thread, I believe this is the video mentioned:
Twitch | Blizzard: Ianator#1479 | 3DS: Ianator - 1779 2336 5317 | FFXIV: Iana Ateliere (NA Sarg)
Backlog Challenge List
Iron processing (Copper and steel are essentially the same):
Green Circuit Manufacturing:
Flying Frame Manufacturing:
As you can see, I'm still in the process of converting everything over, but the beauty is that I can tile out the basic power poles/robo ports and the construction bots will slowly built it out without me having to do any of it. I can then add rails using blueprints so, again, bots are doing it. And then plop down whatever manufacturing I need to expand. Bots will do it. All I need to do at the end is make sure stations and trains are programmed correctly and then done. I've decided to add in a coal station to every single block so that I don't have to worry if a train is getting fuel or not.
Oh, and I'm using the logic circuits to further expand on the station activations. In the Iron Processing block, there are four sets of four chests for loading trains with iron plates. Each of the sets of four is connected to a decider combinator that will output a 1 if the count is below 4000 (A full traincar), 0 if above. If any of the deciders are outputting a 1, it will disable the station. That way, the station is only active if there is enough iron to fully load a train, making sure it is in the station for the minimum amount of time and there is no waiting and allows the freedom of having a train that is fewer than four traincars if that is a better for whatever purpose the train has.
All of the pickups work like that, but stuff like circuits and frames I only use single car trains, so it's just a check at the station instead of needing the combinators.
And, of course, every station is limited to one train so there is never a traffic jam. A train is only moving if it has somewhere it can go.
The blueprints continue to evolve, but I think I have it figured out. I'm still experimenting with managing traffic, but so far only occasional traffic snarls happen, and I think I've got it solved - just waiting to see if it happens again.
The mall, or the central base where all the placeable structures are made, is a mess. It predates using the city blocks, so before I retire this map, I'll do up a set of blueprints for a mall that fits in the city block paradigm.
I'm going to be pushing further out with this map as well. I want to see about maxing out a single rocket silo and that will probably take a lot more infrastructure.
I'm actually looking forward to a new start now that I have the full suite of blueprints ready (Pending the mall blueprints). Launching the rocket took 75 hours this time, but a LOT of that was spent experimenting with design and redoing things. My plan next time is to rush rail and construction bots and then just start dropping blueprints everywhere.
However, this time I launched in 38 hours, so they still offered a pretty good jump. I went solar only and man. Nuclear is so much easier. Solar more the doubled the size of my base.
I think I'm going to take what I learned this time around and redevelop my blueprint library to offer more flexibility, as well as solving the mid game problem.
Penny Arcade Rockstar Social Club / This is why I despise cyclists
My rail network makes logistic bots superfluous - I did the logistic network embargo cheevo this run just....without really doing anything different. Raining Bullets I got as well - that one just required me to keep on top of the spawner nests. Steam all the way? Yeah. After this run, I think I'll never do solar again. It was so much more work!
I actually want to get the <15 hour to rocket launch cheevo. That, <8 hours, Golem and build 111 items or less are the only remaining ones.
Golem I'll probably get by accident at some point, but the other three...hm. Maybe <15 hours, but <8, and 111 items seem probably out of reach to me.
Penny Arcade Rockstar Social Club / This is why I despise cyclists
I'm finally researching the Interstellar Transceiver, a power-hungry structure which serves as K2's victory condition. And by researching I mean, "It took me eight hours to make maybe 50% progress". My current base is very organic in structure, starting with a main bus, transitioning into a mix of Transport Drones and vanilla robots. There were a few AAI trucks in there too but I think they got hit by trains. What this all adds up to is a base that has some difficulty scaling up production and quite a bit of latency in meeting demand.
This world has islands separated by large water features so a city block megabase will require massive investment in Landfill. I may have to use a Transport Drones main bus instead.
Twitch | Blizzard: Ianator#1479 | 3DS: Ianator - 1779 2336 5317 | FFXIV: Iana Ateliere (NA Sarg)
Backlog Challenge List
I even added a feature! I made circuit wiring part of my basic power block design, and most of the modules hook into that so I can set up an alert for when a certain resource gets low. It basically alerts me when there's a production shortage. Not super useful for the end game items (Though I did design their modules to send out a signal, I just don't bother making an alert for them), but for any of the raw materials, or first stage thinks like petroleum or green circuit boards, it's super helpful.
And I have a final version of my mall. It's a bit messier than I wanted because I worked to pack it all into a single block, but I got there in the end.
How I set up my raw material processing:
And the full megabase in map view:
Each module was made to be fully placeable in all four orientations, so my next playthrough, I'm going to use 3x3 grids and then use the middle section for modules that don't use trains. Right now that would just be power related (Nuclear plant, accumulator block, solar block if I use one).
Until you need to do long distance belts can scale perfectly well and the extra fiddliness of trying to balance loads and set everything up eats up way more time than you can save. Maybe if you already have all the blueprints.
Usually I plop down a unified mall + all science blueprint and just feed everything into that if I want to scale up quickly
10% off release sale