|
Agent355 posted:Is there a good way to take down forests quickly? Poison capsules, grenades, construction robots. E, re: modules, slot 2-3 efficiency 1s in common, constantly running and individually unimportant things (miners, furnaces, gear & circuit assemblers) if you care about pollution or not covering the map in solar panels. Slot speed if you want more stuff and are pressed for space. Slot production if you want more stuff and care about using more input materials (advanced modules, capsules, maybe robots). Note that production modules on their own actually slow down the rate of production: they just give you more stuff out per thing put in, at the cost of production speed, pollution and energy consumption. Efficiency is the most useful, especially for furnaces and miners; for a constantly running electric furnace an efficiency 1 is about as much energy gained as a solar panel + accumulator. Speed and production are less good and generally paired to get more at the end of long production chains. If you want to mass produce module 3s or destroyer capsules then slot production and surround with a couple of beacons slotted with speed. These are much more situational, though: most of the time you're better off just adding more assemblers to get more stuff. Xerophyte fucked around with this message at 05:53 on Mar 13, 2016 |
#
¿
Mar 13, 2016 05:24
|
|
|
|
| # ¿ May 7, 2024 14:03 |
|
|
Truga posted:https://www.youtube.com/watch?v=hdBTknCZnGY Well, if we were in it for the money rather than just for the joy of killing all the wildlife and covering all their stuff in glorious concrete, railways and smog. And if Factorio had parts of The Tempest grafted to it for some reason. The best part* of Factorio is that we are essentially playing Captain Planet villains. The only thing missing is a magical ring that would let us dramatically shout "Deforestation!" to get rid of all the pesky trees. Nth Doctor posted:I started making the jump to using bots last night. Is it wise, now, to throw down provider and storage chests for a bunch of intermediate and final goods? 2. Put requester chests as inputs to assemblers that you want fed by bots. 3. Put a field of storage chests in some central area for junk deliveries and temporary storage of wood and deconstructed things. 4. Put active providers in your train station if you want to have it emptied by robots (you probably don't unless very pressed for space), or if there's some product you want immediately moved to your central storage area for whatever reason. Active providers are generally not that useful. Try to not go overboard on running your entire factory on logistic bot transport, it can potentially get very boring. Belts are the best part* of Factorio. *not actually the best part.
|
|
#
¿
Mar 14, 2016 17:19
|
|
|
RattiRatto posted:Hypothetically speaking, is there a way of coupling combinators, wires and active/passive chests in order to come up with a centralized storage with only of a finite number(example: only one cell of the inventory) of multiple items which i decide? (belts, bullets, sonal panels, etc..) Without just actively providing at the distributed assemblers as the above? You can technically do this is if the centralized storage is a different logistic network than everything else. Say the centralized storage has a Network B and the factory has a Network A. You'd have chests requesting one bot stack limit of items in network A, belt or dumb inserter those items over to network B, then use smart inserters to fill up centralized passive providers in B to your threshold using the logistic filter. Having two disjoint networks probably defeats the purpose of the exercise, though.
|
|
#
¿
Mar 14, 2016 18:17
|
|
|
Agent355 posted:I want some way to measure total throughput of a particular place and total input capacity of another so I can be more optimal. What's "a particular place"? Individual assemblers list what they require, what they generate and how much time they take so you can arithmetic your way to know what they need. For arbitrary regions of factory? Not so much. Frankly, my general advise on good ratios is gently caress ratios and gently caress efficiency. The only thing it costs you to screw up your ratios is that you will have some assemblers, furnaces and mines along the chain that lie idle while either their output is saturated or their input is empty. This is perfectly okay. It costs you a small amount of idling power and some space. Getting perfect efficiency costs a bunch of time and effort in setting up your factory, which are much more precious as resources go. It's worth adhering to some trivial equivalencies like 3 wire assemblers feed 2 circuit assemblers, but beyond that just build everything in easily expandable assembler lines with inputs on belts on one side and outputs on belts on the other side. If you need more output, expand the line. If you need more input, expand the line that feeds the input belt. Repeat until you get enough stuff out at the business end.
|
|
#
¿
Mar 15, 2016 23:21
|
|
|
Loopoo posted:Is there a particular reason loops are bad? For a small network it's not going to matter. If you've got a lot of trains later on then it can be painful to have a section of track with half the capacity of the surrounding rail, though.
|
|
#
¿
Mar 20, 2016 22:32
|
|
|
Loopoo posted:I don't understand accumulators. How can I accurately determine if my stored up energy is enough to keep laser turrets firing at peak performance for 20 seconds? They're both in different units. Watts and Joules. Watt is defined as Joule/second so you could ideally math it, however I'm not sure the turret interface is always accurate. The way it works is that each shot costs 800 kJ in addition to the 6 kW idling cost, so consumption starts at 2.4 MJ/second extra while firing 3 shots/second and increases to 6.24 MJ/s at maximum research and 7.8 shots/second. Estimate: 1 full accumulator feeds 1-2 seconds of continuous fire for 1 laser turret.
|
|
#
¿
Mar 22, 2016 16:34
|
|
|
CanOfMDAmp posted:For those of you with a bus going on, how do you pull from 3+ belts of one material while keeping it balanced? If you really want to you can add a perfect N-to-N balancer downstream on the bus from the splitoff point. For reference, the perfect 3-to-N balancers are:  And the perfect 4-to-N balancers are:  In practice, you shove a few splitters in every here and there after splitting things off to keep things flowing and call it a day. Perfectly balancing the bus after pulling resources is not important or necessary; the bus cannot maintain full throughput after resources are removed anyhow. You might want to favor parts of the factory that come later down the bus, but doing count-perfect balancing after each split will always favor an earlier split-off over a later one. Really, as long as the slack is spread out sufficiently that all the source furnaces/assemblers for that resource are working and none of the sinks are starved, who cares?
|
|
#
¿
Mar 26, 2016 03:40
|
|
|
zedprime posted:That's the practical method of throw splitters at it vs bulkier perfect balancing schemes as explained in this reply last night To clarify: using perfect balancers for your bus isn't just impractical, it gives zero additional value over imperfect balancing. A resource bus can essentially be in one of three states: 1: All your resource sinks have full inputs. The belt balancing of the bus is irrelevant. 2: Some of your resource sinks do not have full inputs and some of your sources have full outputs. There is objectively some problem with how your belts are organised and you should fix it. 3: Some of your resource sinks do not have full inputs and all your sources are outputting at full pace. Where and how you balance the bus is purely a matter of which sinks you want to prioritize over others. Belt balancing is a trap of sorts in that unbalanced belts are not themselves a problem. A furnace or assembler not being able to put a requested resource on its output belt is a problem and one you can in some cases fix with a balancer, but even then there's usually no advantage to doing something "more perfect" than slapping down a splitter or two to clear the logjam. About the one unambiguously useful place to have count-perfect belt balance is on busy train stations in order to make the unloading even.
|
|
#
¿
Mar 26, 2016 22:16
|
|
|
I decided to start 0.15 by playing on the higher recipe difficulty setting, which apparently doubles the amount of iron used for gears and steel and triples the copper used for circuits. This is the mess I ended up with in order to saturate a single green circuit assembler prior to researching stack size bonus. Not sure if I should be proud or dismayed: Friends don't let friends put copper wire on belts. Getting stacksize bonus 2 allowed me to get rid of most of the excess inserters so it's a bit less messy, but still kind of horrifying. In general, my current factory -- I just finished basic liquids and blue science -- is pretty squirrely. There's an ad-hoc bus and a mess of things around it, which is a lot more fun to work with than the well-planned and spacious factory I made last time I played. Embracing beltgore and solving the resuling problems as they happen is more fun, especially now that the recipes throw me off again.  That's 2 full iron lines feeding half a gear belt, which can barely keep up with science when it's running full tilt. I'm happy that they didn't go quite as ham on the recipe costs as I recall the old marathon mod doing, but I'm still going to need significantly more mining and smelting than usual. Seems like a nice enough balance.
|
|
#
¿
Apr 28, 2017 05:25
|
|
|
Breetai posted:Always remember that you can put a short length of rail down and use a train carriage as an extra long super crate if you need to feed multiple assemblers into one. Yeah, that'll work. It does require that you've researched rail and gotten some pieces built, which you haven't when setting up green circuits. It also feels to me like bypassing the fun belt logistics problem of providing 20 wires/second to an assembler with a somewhat dull warehouse mod-like solution, but that's admittedly just me. The "correct" early game approach is probably to ignore the actual ratio and do some simpler 3-1 or 4-1 setup that takes less attention to build. It's not like adhering to the no-idling ratio gives you anything in the early game, and player effort is by far the most constrained resource. When you've got beacons and module 3s everywhere then sure, a having idling assemblers in your design is actually costing you some measurable amount of resources, but then you've also got a lot of other tools to deal with the logistics issues.
|
|
#
¿
Apr 28, 2017 07:24
|
|
|
If you're concerned about wasting your uranium you could always build a couple of fuel cells worth of accumulators, then set up a circuit network to only insert fuel cells in your reactors when the accumulators are nearly empty. One fuel cell is 1600 accumulators' worth of charge* and I doubt it's worth spending that much stuff on accumulators for better uranium efficiency when uranium isn't that uncommon, but you can. E: *at 1x efficiency, with adjacency bonuses you presumably get more.
|
|
#
¿
Apr 29, 2017 04:25
|
|
|
Alkydere posted:Yeah, in 0.14 and before it was "You need iron to start and build your infrastructure, but afterwards everything is increasingly circuit hungry." My expensive recipes railworld game has been a mix as I've advanced. Iron was definitely the main limiting factor up to military science, but with blue and production science my copper requirements shot up and with hi-tech science I'm definitely more in need of circuits than anything else. I've only got a token amount of processing unit production as is and I'm going to need a spectacular amount of copper to get it revved up, as well as for module 3s and beacons. All the advanced circuit production is also pretty hard on plastic, but oil seems less problematic than before so I haven't even needed to ship any by rail yet unlike 0.14. Not sure if I just lucked out, I had a very large starting deposit. More iron is also going to be a thing for blue belts, but that's more a luxury than a requirement.
|
|
#
¿
Apr 30, 2017 21:40
|
|
|
Alkydere posted:Yeah, Purple (Production) Science takes a pumpjack (5 steel), an electric furnace (10 steel) and an electric engine (another steel) so each purple beaker takes 16 steel to make. Not to mention a whole lotta other stuffs. Each craft produces 2 purple science packs, fortunately, so half of that. 52 furnaces smelting steel is still a lot, especially at 10 iron/steel on expensive.  E: FWIW, my solution to purple science was to alternate researching things that require purple packs with things that don't so that the packs could buffer. Then settle for having half-ish the purple production that I do of everything else. This let me get away with having a mere 2 steel furnace columns, fed by 4 iron furnace columns, for my entire factory. Cheap! E2: Also, note that military and hi-tech science packs are also 2/craft, which I completely failed to notice for military so I'm massively overproducing them. Oh well... Xerophyte fucked around with this message at 22:38 on Apr 30, 2017 |
|
#
¿
Apr 30, 2017 22:00
|
|
|
Loopoo posted:6 hours later and I still haven't migrated my factory The personal roboport mk1 is pretty early now, I ran around with modular armor and a solar powered one for a long while before I got production and hi-tech science sorted. Using solar power and batteries to power it was surprisingly OK; sure, it ran out of power if I did any major projects but for those I could usually just plunk down a roboport and a storage chest on demand. It also let me make use of the new night vision goggles, which are pretty nice. The only real annoying part was laying rails which I had to do manually since having robots lay down a long line would drain my batteries. At least I could blueprint down the signals, lights and power poles afterwards, and putting down track isn't that bad since 0.14.
|
|
#
¿
Apr 30, 2017 22:15
|
|
Only thing stopping me is the lack of a personal roboport. I can't bear the thought of setting everything up by hand. I wish the game had an option to at least start you off with one roboport and 20 construction bots!
|
Someone made an automated shuttle train:
|
|
#
¿
Apr 30, 2017 22:37
|
|
|
super fart shooter posted:How are you getting 52 furnaces? When I figured out I'd need 105 furnaces making steel, I already factored in that it's a 2x science pack recipe and was calculating for steel/electric furnaces. I just wanna know cause I'm usually pretty bad at doing the math for this kinda stuff and I'd love to find out that I calculated wrong. It's 16 steel per craft of 2 packs. You want 0.75 packs per second. You need to produce 16 / 2 * 0.75 = 6 steel/second. It's 17.5 / 2 seconds per steel smelt in an steel/electric furnace. 6 * 17.5 / 2 gives 52.5 smelters to satisfy. I think that's correct, anyhow. I could be totally off with the time to smelt. E: Okay, I was slightly off with the time to smelt. Not sure where I got 8.77s from. Xerophyte fucked around with this message at 22:53 on Apr 30, 2017 |
|
#
¿
Apr 30, 2017 22:50
|
|
|
M_Gargantua posted:Yeah I know they're not actually hashes. What I'm curious about is since they track orientation, reverie, and I think most likely to cause trouble, specific power distribution links and circuit network connnections. Each individual poles connections need to be tracked. Gotta be some fuzzing you can do against it. They already need to have every single piece of data in the game serializable for saves and network, blueprint strings are likely just encoding their underlying common binary format in Base64 or similar.
|
|
#
¿
May 1, 2017 07:55
|
|
|
I've mined out a couple of 1M-2M iron ore fields and made a decent dent in a 10M ore one. I've got all research packs under production -- although not a lot of yellows -- and have managed to (very slowly) research the Kovarex enrichment process tech which is of similar cost to rockets. That's on the "expensive" difficulty so my numbers are likely bigger than those, but they seem reasonable. I don't have the circuit production to do productivity modules (I've made 2 prod3s by hand and I shoved them in my nuclear fuel assembler), but the things that really enable you to scale to 10x whatever your current size is are thousands of construction robots and the infinite power of the atom. Modules are just there to not make you go insane as you scale further.
|
|
#
¿
May 1, 2017 09:12
|
|
|
Buses are nice as somewhere guaranteed to always have the stuff you need to make the stuff that makes you more stuff. I don't know how people got this idea that bus factories are by necessity neat and tidy, though. I made a neat and tidy bus factory once. Every branch was for a specific purpose, made straight and infinitely expandable. Any useful product was belted back to a chest on the bus for easy access. I found it horribly boring, and resolved to only branch squirrely messes of beltgore off my buses in the future. Job successful! Of course, now I'm at the point where I try to replace most of my bus with trains and dedicated crafting outposts for each component type. Not so much because doing so is efficient as because trains are cool and I want more of them. I've got an outpost for uranium processing, another with the reactor, a third for belts and so on. Belts turned out to be kinda problematic. Blue belts, and especially undergrounds and splitters, are litterally hundreds of iron/unit on expensive. I'm mining two deposits of about 10M iron total, if I mine them out I'll have about enough to replace my existing belts. I should not have skipped reds and am regretting my life's choices. Alternatively, I should strip mine more lands and shove their bounty into giant fields of furnace hell fueled by the radiant heat of the atom like a good Factorio player.
|
|
#
¿
May 3, 2017 05:22
|
|
|
Skip might be putting it strongly, but on standard I make very limited use of them and generally just use 2 yellows unless otherwise constrained. You don't need both throughput and space efficiency until beacons and modules. Doing the same on expensive has been slightly more problematic.
|
|
#
¿
May 3, 2017 05:48
|
|
|
I started .16 on a railworld-ish setting with lots of water because water looks neat now and I wanted a little more of a routing challenge.  I may have slightly miscalculated on my ore frequency settings. That's the first iron patch I've found. I've got 320k iron ore and 133k coal back at base, better make 'em count...
|
|
#
¿
Dec 23, 2017 12:51
|
|
|
Alkydere posted:Yeah, when adding a lot of water you have to turn up resource volume or else it just spawns in water and you get nothing. Yeah, I know that's the mechanic. I'm kind of curious how they've implemented them. Best guess: each resource gets its own Voronoi tesselation with one deposit/cell, then they use gaussian_on_voronoi_centra * noise - threshold to form the resource values, maybe? Anyhow, I did check the cool new preview thing and thought the map looked fine for the level of scarcity I wanted, but I wasn't really aware of the scale and I promptly forgot where all the deposits were. It doesn't help that I can't take a straight line anywhere, either. That iron deposit is hooked up now, I just need to add a coal line and I should be able to get to bots and tanks and reactors before I need to scale up again. I'm happy I set the biters to mostly harmless on my little Lakeworld, building my giant circuitous rail lines would be a lot more annoying if I had to clear out serious biter camps.
|
|
#
¿
Dec 23, 2017 19:02
|
|
|
Dr. Arbitrary posted:If Factorio wins this steam award, I hope they put the game on sale for $19.99 for one day to celebrate. It's up against 3 of the most popular games on the platform and a dark souls. It's pretty amazing that a niche game about steam powered logistics for weirdos is up against the 3 of the most popular games on the platform and a dark souls in a popularity contest, but that's likely as far as it'll go.
|
|
#
¿
Dec 30, 2017 17:06
|
|
|
Olesh posted:Yeah, this is also why adding speed modules to things with productivity paradoxically generates significantly less pollution. There's a fun second part to that due to how the individual category speed, production and power use bonuses and maluses from modules stack additively, but production and speed bonuses combine multiplicatively for overall throughputs. Speed 3s are +50% speed, +70% power use. Prod 3s are +10% productivity, -15% speed and +80% power. Say you fill an assembler with 4 Prod 3s (a good idea). You get +40% productivity, -60% speed, +320% power. You produce at 1.4 * 0.4 = 0.56x normal speed, at 0.56 / 4.2 = ~13% power efficiency. Say you add 1 Speed 3 through beacons. You now get +40% productivity, -10% speed and +390% power. You produce at 1.4 * 0.9 = 1.26x normal speed, at 1.26 / 4.9 = ~26% power efficiency. Even though individual speed modules are not efficient, adding a speed module to something that has a speed malus actually improves your power efficiency! In principle this continues indefinitely, as you add more speed modules to an assembler with +40% production you come ever closer to 1.5 * 1.4 / 1.7 = 100% power efficiency, i.e. just as power efficient as just building more assemblers. Only without needing a ton of space and with 40% better resource use, which saves you power in all earlier steps in the production chain and a ton of ore if you do it everywhere. Which leads us to the end game factory design rule of thumb: try to alternate rows of assemblers (or refineries or chemical plants) with Prod3s with rows of Bacons with Speed3s. Get more stuff for less things in less space faster.
|
|
#
¿
Apr 18, 2018 22:02
|
|
|
greatZebu posted:If there's a bug in a 9 9s system and it goes down for an hour, it just blew its error budget for the next 100,000 years. If it goes down for one tenth of one second, it just blew its error budget for the next year. I don't think there's any system in existence that can be reliably shown with confidence to be 9 9s and definitely not 8. But if there is one, I would love to read about how they handle their monitoring. The only time I can remember coming across the "nine nines" statement is for Ericsson's AXD301 switch so here's Joe Armstrong's thesis if you really want. The case study is in chapter 8. I can't speak for how accurate their measurements are but it's by all accounts a pretty reliable switch. In general, I think there are engineering systems in reality that have extremely high reliability. You can find situations where one second of reactor containment failure is probably one second too much, for instance. They're just typically not IT systems.
|
|
#
¿
Oct 9, 2018 05:15
|
|
|
Magus42 posted:I'm in a no-biter game for now while I get the hang of things Radars give remote vision around them, which you'll be able to use to remotely build things when you get robots, and also passively "explore" surrounding areas for you. They're great no matter if you have biters or not, you should throw them up around any mining outposts and train tracks you build just to find more stuff.
|
|
#
¿
Jan 5, 2019 23:38
|
|
|
Vic posted:Bang-for-buck wise you'd be looking at ATI Radeon RX560 - 580 or Nvidia Geforce GTX 1050 - 1060. You might also consider the new -- as in, one week old -- GTX 1660 Ti which is basically nvidia's value card for the newest hardware generation. It's a cost bracket up from that list at around $280, but still very good price/performance. Going that high is more a future-proofing thing or wanting to play other games than specifically for Factorio, though. It's not a perfect tool and SHSC would probably yell at me for suggesting it but for a basic benchmark ranking of absolutely everything you can use userbenchmark. Sort by value%, look at the price and avg. bench%. Any newer card in the $150+ range will run Factorio without any problems, though I'd suggest avoiding the 3 GB ones.
|
|
#
¿
Feb 28, 2019 17:43
|
|
|
I'm simplifying very wildly here but there's very little distinction between 2D and 3D in terms of what your hardware does. Graphics cards accelerate 3D graphics by compressing the entire process down to 2D graphics as soon as they possibly can, which they do by taking each triangle of the 3D triangle meshes you ask them to render, projecting those triangles down to 2D screen triangles and then drawing those 2D triangles directly. The projection from 3D to 2D is typically cheap and the vast majority of the work happens when drawing the 2D triangles. A "2D" game like Factorio is still fundamentally drawn in 3D. Their 3D data is flat terrain, flat sprites above the terrain and a flat camera so it's rendering a lot of squares floating above a lot of other squares. That means the projection from 3D triangles down to 2D triangles is even simpler than usual, probably even trivial, but since that projection is almost always negligible it doesn't meaningfully make the rendering work much cheaper. In Factorio the actual drawing part isn't very complex either so you still don't need particularly high end hardware, but that's not exactly guaranteed. How much graphics card you need for 2D games depends basically on how many and how complex textures the game needs to load to draw a screen and how complex the drawing process is. For a game using pre-rendered backgrounds you can easily need way more GPU memory than equivalent fidelity 3D graphics, for tiles probably not so much. Things like shadows, dynamic lighting, smoke and fog are just as complex to do in a 2D game as a 3D game since in both cases they're done in a 2D triangle drawing step (again, simplifying wildly). They can cripple any GPU if your game turns them up to 11; for Factorio I know smoke has been a persistent performance issue on the lowest end hardware, which is not surprising because transparent layering can eat fillrate like mad. Gravy Jones posted:Out of my price range (in the UK where cards especially newer ones are a lot more expensive). I'd almost pulled the trigger on something ummm, RX 580, but my PSU may not be up to the job so will need to do more research. I may end up not bothering as I don't really play anything that troubles my current card other than the vram thing for hires factorio and I do fine with the regular graphics. Thanks for the tips though. Sure, that's perfectly sensible. A 4 GB RX 580 is an excellent card for its cost and should be more than enough for Factorio. Looking it up and it looks like the RX 580 has a 200W max draw vs about 60W for the 750 Ti you said you had so there's definitely a difference. The GPU is usually the hungriest component in the system so you should be able to run on a 450W PSU, but 550W or more would be more efficient (less actual watts drawn from your wall).
|
|
#
¿
Feb 28, 2019 19:44
|
|
|
The Locator posted:Thanks, I briefly considered re-doing my labs to do that when I noticed that there were more than 4 science colors, but then I realized that I had accidentally left enough room on either side for that purple/yellow belt feed, and that was just easier to do than rebuild it all. It's worth noting that with red inserters and undergroundies there's no real problem in feeding an assembler or lab with 3 belts on a single side. So if you really want to have a purely belt fed 22 component recipe, you can. E: I guess with weaving you can in fact have Xerophyte fucked around with this message at 16:25 on Mar 9, 2019 |
|
#
¿
Mar 9, 2019 16:21
|
|
|
Dancer posted:There's that one mod with 30 different science packs... And, unless I'm mistaken, weaving won't let you use more than 3 belts per side because there's only 3 spaces for inserters. You can get 5 belts/side with classical parallel belts:  It doesn't actually work with yellow belts, even with all the undergroundies pointing the right way, but you get the idea. Of course, you can trivially just run 2 separate weaved belts straight at the assembler for the aforementioned 24 belts in so not sure why I bothered to proof of concept the parallel thing above.
|
|
#
¿
Mar 9, 2019 16:46
|
|
|
Syka posted:Which is more of a concern now that solid fuel is a required part of your progression. Though I guess you can turn petroleum into solid fuel if you really have to. FWIW, I've used primarily solid fuel for furnaces and trains and so on after I set up oil for a couple of factories now. My pump setup is: code:Essentially I crack to keep the 3 types balanced, turn light oil to solid fuel if I have a good supply and turn petroleum to solid fuel if I'm almost full. It's pretty easy to implement: you just need to wire the pumps to the storage tanks and set the condition you want directly on the pumps, no need for combinators. You could do more and less complex things like skip the solid fuel conditions entirely or set up conditions to prioritize solid fuel over plastic/sulfur if the supply for your power plant is running low, but this works well enough for me. If you have flamethrower turrets you probably want to prioritize light oil for them I guess; I play with biters set to not expand and pollution off so they're not really a thing I use. Xerophyte fucked around with this message at 18:30 on Mar 14, 2019 |
|
#
¿
Mar 14, 2019 12:15
|
|
|
Mr. Powers posted:I guess I really don't get the advantage of the splitters in this design. It looks like the same basic belt side feed design that I hate except now it has splitters. The unload is slower because you're only using 8 inserters per car rather than 12. Far as I understand the benefit is that stack inserters actually unload to both outputs of the splitter, which then makes the item drop part of the inserter loop twice as fast. 8 inserters on splitters will end up with higher throughput when taking from the chests than 12 inserters directly on belt at some inserter stack count breakpoint. If chest to belt is the throughput bottleneck more than train to chest it makes to prioritize that step. If my understanding is correct it still seems like there are a lot of individual variables in all that: what fuel your trains use, how long they are, your station design, train brake research, inserter research, etc. E: This said, you should be able to make a(n inefficiently and stupidly) larger station design that uses 12 stack inserters/wagon and still outputs from chests to splitters by chaining together chests with stack inserters to make space. Hop to! Xerophyte fucked around with this message at 13:17 on Mar 23, 2019 |
|
#
¿
Mar 23, 2019 13:11
|
|
|
Evilreaver posted:Not quite, like I said mine as signaled can handle pairs (either both lateral or both right-hand-turning). It can only handle one left-handed solo at a time or one U-turn. They have very nice throughput and they've never caused me grief except when I deleted my blueprint book accidentally then had to rebuild them and re-signal them That roundabout is a little dangerous. You have a deadlocking problem in for instance situations like this:  (I am not great at paint) How dangerous depends on train length, routes and so on, but this deadlock can also happen when there's no turnaround for either train and just two simultaneous left turns. It probably wont happen often unless the train network is both very busy and has a lot of routes where long trains cross over, which isn't typical in Factorio where your trains are usually all going to or from a central hub. Anyhow, to be safe in general you need to use chain signals for signals that are inside the roundabout itself, same as for internal signals in an intersection. Internal chain signals will still allow trains to pass and allow simultaneous right turns, but they mean entering trains will reserve the entire roundabout for U-turns and 3/4th of it for left turns which slightly reduces throughput when these things happen. Since you need to use chain signals inside both roundabouts and intersections unless they're large enough to park entire trains there's not a lot of difference in performance between them.
|
|
#
¿
Mar 24, 2019 02:33
|
|
|
This isn't the best factory ever by any means but it launches rockets and after my runs for the last couple of patches just kinda petered out once I got to modules I'm happy to finally have gotten to end game again. main base:  (click for giant 8000x4000 version, thanks /c game.take_screenshot) map:  Most of the messy parts of the base are intentionally messy; I made a clean and logical base once and hated it so now I try my best to do all manufacturing at least slightly ad-hoc and just cram in new things where they fit and the right resources are nearby. Some of the messy parts are not intentionally messy, like the terrible awful copper belting that's really limiting my output. I also don't use blueprints from other people or even my old factories because, well, designing assembler setups is the fun part of the game to me. Some of the designs I made this time turned out pretty good, like the red chip layout. Others not so much, like doing all the rocket parts on artificial islands which in practice was more annoying than cool. I am particularly happy with how my enrichment setup turned out:  First time I set up enrichment since priority splitters were added and they let me do it in a way that's nice and compact, and looks pretty neat in motion. The circuit logic is relatively simple. The enrichment loop has buffer chests at its input (left) and output (right). The output buffer is set to hold a max of 600 U235 and U238. If it has >500 of both types then the ore processing is turned off. As a failsafe I also turn off the enrichment centrifuges if the input buffer ever has more U235 than U238, but I don't think that'll ever happen.
|
|
#
¿
Feb 7, 2020 17:34
|
|
|
I like Satisfactory well enough but it and Factorio are -- fortunately! -- different games and the playerbases aren't going to 100% overlap even if the central concept is very similar. Factorio has a very strict primary focus on expanding the factory to meet the needs of the expanding factory, with a secondary focus on defense (for most map settings) and exploration very tertiary. One consequence of pure procedural terrain is that it's not like you're going to come across any cool landmark or game-changing things in the world. All you really need for exploration (again, for most map settings) are a couple of radars on your perimeter to tell you where the next ring of deposits are. Satisfactory removes the defense aspect entirely and intentionally simplifies the logistics to add a lot more exploration. There's a ton of lovingly hand-crafted terrain with all sorts of secrets and game-changing loot, many resources are intentionally hard to get to, etc. The logistics game is intended to be done fairly small scale: most of my factory expansion spurts in Satisfactory have me hook up maybe 10-15 new structures to produce just a few items of a newly researched end product per minute from some baser resources. Then I put those structures inside some walls, spend a long time pondering the best shade of pink to use for this new expansion and finally go off exploring for magic worms for an hour or two while stuff accumulates, research completes and elevators fill up back in the base. If the appeal of Factorio for you is as a large scale logistics puzzler or esoteric programming language and you hate collect-a-thon platformers, fixed worlds and non-orthographic projections then Satisfactory probably won't be your thing. Even with the wonderful people-tubes, and people-tube cannons.
|
|
#
¿
Apr 2, 2020 04:19
|
|
|
Got to the rocket launch in my K2 run  ...which I guess means I'm about halfway?   Click for 8660x11547 version 1.0 turning my lake into a green polluted hell-hole is wonderful and I 100% approve. Map, not that I've needed to expand very far yet:  Very proud of my mall, which is a splendid, unplanned, compact mess of beltsnakery. K2 has been pretty nice so far. Less of what I think is the repetitive bullshit that made me bounce off most other mods, mostly interesting new puzzle wrinkles. There are some missteps: the sand-quartz-silicone chain is pretty pointless, "radioactives" damaging you is stupid, and they've massively increased the uranium fuel cell cost (30x the U235!) so running a reactor before enrichment now requires a gargantuan raw ore income. I extra disliked only noticing the new fuel price after booting up what I thought was a modest 2 reactor power plant; I ended up just cheating myself some more U235 to keep the lights on while enrichment got moved up the research queue. [E: Resized my giant screenshot to 8660x11547 instead of 6Kx8K since that's closer to drive's 100 megapixel limit for panoramas and I'm a giant nerd.] Xerophyte fucked around with this message at 23:06 on Aug 14, 2020 |
|
#
¿
Aug 14, 2020 22:46
|
|
|
Fuzzy Mammal posted:I like the train world settings. No expansion but if you let your cloud expand you'll still get attacks. I like to also change the evolution settings to have no evolution from destruction or pollution but 2x or 3x the evolution from time. Pollution still triggers attacks but there's no pressure on me to handle the biters in some specific way to avoid them leveling. I can turret some parts of the base, clear all the nests near the cloud elsewhere and it all works out the same. There is some pressure on me to not AFK for several hours while the base works I guess, but even 3x is still mild biter evolution without the other factors.
|
|
#
¿
Aug 17, 2020 22:27
|
|
|
The easiest way to get the speed achievements is to play with as many other people as possible and brute force the hell out of it. If you want to go at it alone, I think the most important thing is to make a plan. It doesn't have to be a perfect plan, but make a rough list of what you're going to do in what order (e.g. initial power -> initial smelting -> automate gears/chips -> red/green science -> double smelting -> double power -> oil...), what you're going to research and stay relentlessly on task until done. Buildwise it might help to watch one of the speedruns, takes less than an hour on 2x, but they can be confusing since they build very compact bases to minimize running. The big thing to take away from them is to favor simple factory designs that are easy to physically build and expand. Do not do fiddly things for 2% more throughput or the like. Beyond that, try to build and buffer pre-requisites before you need them, don't research things that you do not need and go big. Getting a rocket out in less than 8 hours is honestly not too hard on custom settings if you know the game. It requires a somewhat odd playstyle and some preparation, but shouldn't be too bad. 8 hours on default settings is way more difficult.
|
|
#
¿
Aug 19, 2020 11:47
|
|
|
The way the nuclear "ratios" shake out in vanilla is that you need 3.333 miners to saturate 1 centrifuge doing ore processing, and 1 centrifuge produces on average enough U235 to run around 1 reactor (1.166 to be more exact). Since U235 is random and you want to save some for enrichment down the line you will want more than 1 centrifuge/reactor, but you can absolutely run a starter 1-2 reactor nuclear power plant on a very modest ore patch and still be virtually guaranteed to run a surplus.
|
|
#
¿
Aug 20, 2020 20:26
|
|
|
|
| # ¿ May 7, 2024 14:03 |
|
|
DreadCthulhu posted:How do you all deal with constantly respawning bug bases? Do you keep running around the map taking them down, or is there a way to make it less of a nuisance? At some point you're big enough that there's always something under attack :/ Well, personally I turn off biter base respawning at worldgen. If not doing that you end up needing to set up a big perimeter wall. Enclose your base(s) in walls, turrets, flamers and an autosupplied ammo belt if using gun turrets (generally a good idea). You want to get roboports and construction robots set up pretty fast to automatically repair your defenses, getting them should be your first goal after doing initial oil processing and setting up blue science. The goal is that, yes, you'll be constantly under attack but you don't personally have to do anything about it since your defenses have enough firepower to kill the attacks before they destroy anything, and your bots automatically repair any damaged wall segments or turrets. In the late game you'll have artillery and a platoon of angry spider robots, but it doesn't sound like you're near launching rockets yet.
|
|
#
¿
Aug 25, 2020 06:50
|
|