|
Sydin posted:Yeah the bot mall implications of being able to switch assembler recipes on the fly via circuit conditions is huge. Not sure which I'm more excited for: the expansion, or the inevitable Dosh video of him doing some absolutely insane circuit bullshit with the expansion. New challenge run: Launch a rocket with only a single Assembler
|
#
?
Jan 19, 2024 18:23
|
|
|
|
| # ? May 26, 2024 15:23 |
|
|
Tenebrais posted:New challenge run: Launch a rocket with only a single Assembler New achievement: Crazy Bastard
|
|
#
?
Jan 19, 2024 18:53
|
|
|
ymgve posted:Did they say the expansion would have an early access period, or are they planning to wait until it’s complete? What they said was no less than a year. Everything in Factorio development has taken more than twice what they expected. That said, this time they didn't even announce what they were doing until they were like a year into it, so maybe early fall 2024 will happen. I wouldn't really expect an early access style of release, just because it would be weird to revert a finished game into early access state. I think it'll be more or less done when they put it out, but they will probably do things like they did with 1.1 afterward.
|
|
#
?
Jan 19, 2024 18:53
|
|
|
K8.0 posted:What they said was no less than a year. Everything in Factorio development has taken more than twice what they expected. That said, this time they didn't even announce what they were doing until they were like a year into it, so maybe early fall 2024 will happen. I wouldn't really expect an early access style of release, just because it would be weird to revert a finished game into early access state. I think it'll be more or less done when they put it out, but they will probably do things like they did with 1.1 afterward. Didn't Oxygen Not Included do an early access for their expansion?
|
|
#
?
Jan 19, 2024 19:15
|
|
|
Tenebrais posted:New challenge run: Launch a rocket with only a single Assembler One of the first things I saw someone do with recursive blueprints after it came out was launch a rocket and make 1000 of every other science pack in something like a 13x13 area (including power), with everything being made in stages, and one belt/pipe of each raw material coming in. But that involved deconstructing and reconstructing everything (except for a couple of chests) between stages; this will allow you to do it in a space that can hold a rocket silo, refinery, chemical plant, smelter, assembler, and boiler/engine (or perhaps a single solar panel if you don't care how long it takes). e: Found it https://www.youtube.com/watch?v=9dzQge6pe2o
|
|
#
?
Jan 19, 2024 20:15
|
|
|
ymgve posted:Did they say the expansion would have an early access period, or are they planning to wait until it’s complete? I forget which 3F it was, but the plan was “2.0+expansion shows up one day fully released”
|
|
#
?
Jan 19, 2024 20:25
|
|
|
It was 367:quote:Step 6 - Beta test The "not less than a year" was from #365, which was almost 3 years ago now.
|
|
#
?
Jan 19, 2024 20:35
|
|
|
celestial teapot posted:You're losing a lot of throughput that way. Put belt balancers after your unloading station. Check out raynquist's balancer book. It seems like 4 stack inserters set up this way can still saturate a blue belt and it avoids the chests furthest from the balancer from draining first? The issue I had with them all just going for it is the chests wouldnt drain at the same rate, eventually leading to the train sitting in the station with 3/4 cars empty as it cleaned the dregs out of the last one.
|
|
#
?
Jan 19, 2024 21:46
|
|
|
That should never happen if you're using a proper balancer. Sounds like you're using one that doesn't balance evenly and is pulling from some cars more than others.
|
|
#
?
Jan 19, 2024 21:58
|
|
|
It's the standard 4-4 balancer from the book 
|
|
#
?
Jan 19, 2024 22:28
|
|
|
FWIW I also use the 4-4 balancer from that blueprint book, and I've repeatedly encountered the same thing - if I don't force the inserters to all swing at once, the center wagons get unloaded before the outside wagons are even halfway done. Don't know how or why, just that it's not just the one person having the issue.
|
|
#
?
Jan 19, 2024 22:45
|
|
|
Tenebrais posted:New challenge run: Launch a rocket with only a single Assembler They said unused components get dumped into a local inventory, which from the screenshots is only 4 items big though. So that seems like it would still be very tricky to switch between more than 2 recipes unless they have a lot of common ingredients. Or else you'd have to resort to a lot of very careful counting of inputs to prevent there being excess getting dumped out when it changes recipe
|
|
#
?
Jan 19, 2024 22:50
|
|
|
Just to be clear, we're talking about an unloading setup something like this, right? Ignoring exactly how the items get onto a belt, but you've got 4 belts going into and out of the balancer, one belt per train car? Squibbles posted:They said unused components get dumped into a local inventory, which from the screenshots is only 4 items big though. So that seems like it would still be very tricky to switch between more than 2 recipes unless they have a lot of common ingredients. Or else you'd have to resort to a lot of very careful counting of inputs to prevent there being excess getting dumped out when it changes recipe I have a feeling those are ephemeral inventory slots and the machine will have as many or few as needed to hold the items. What happens if you're making satellites and switch the recipe? In other news, preliminary testing on the 192x192 grid is looking good. I can even fit 16 smelter lanes.  I wanted a wider road so that I would always have room to underground past the 4 tiles that the large power poles / lamps are going to occupy. This means that inside the block I've got 8 tiles occupied on each side with path or power, but everything within that should be usable. And it fits the 4x4 roboport plan pretty well too. The inner 4 roboports could be shifted inward a bit and still have full coverage; if you move them outward then the center of the block starts losing logistics coverage. There's only 4 tiles of wiggle room, though, but it should be enough to allow builds to fit better without sacrificing coverage. On the smelter front, unfortunately I have to avoid the roboports in one direction or the other, so I went with this middle gap design (ignore the starter base):  I can't move the headers (at the bottom) back too much farther, otherwise there won't be room for 16 lanes of incoming ores (just barely enough right now, and that involves skirting around the roboports in an ugly way). Unfortunately this also means I can't double up on the other side of the block and have 32 lanes all emptying to the center of the block (but this is almost certainly possible with 256x256 blocks, which is just too big). The extra space isn't wasted though, this leaves ample room for fitting in balancers on both the input and output sides and there's more than enough space to extend by another 12 furnaces per side and run blue belts, something I was only barely able to do with the 128-sized block system (and at the cost of having to do balancing outside the block). My next problems are figuring out how to run power to the inner roboports and starting to think about steel/brick smelter designs. I think I can get 10 lanes of steel which would produce 2 full yellow or red belts depending on the belts/furnaces used. Stone brick only uses 2 blocks more width than a standard smelter lane so 12 lanes of brick from 24 lanes of stone would almost certainly fit. After that is starting to look at power blocks, steam and solar for now and eventually a nuclear build (where I am aiming to have the build take up 1/4 of the block, with the reactors in the center so you can place multiples in one block and get the reactor bonus). And then the production blocks. I should be able to fit an entire set of non-space science in one block, not sure how much it will make but a couple hundred per minute would be more than enough. 900spm on the last base took about 8 blocks of actual production, not counting refinement or smelting.
|
|
#
?
Jan 19, 2024 23:29
|
|
|
Squibbles posted:They said unused components get dumped into a local inventory, which from the screenshots is only 4 items big though. So that seems like it would still be very tricky to switch between more than 2 recipes unless they have a lot of common ingredients. Or else you'd have to resort to a lot of very careful counting of inputs to prevent there being excess getting dumped out when it changes recipe I can think of like a dozen fun ways to try and manage what's sent to the singular assembler before you unlock a loginet. And good news, there'd be enough downtime to try them all!
|
|
#
?
Jan 19, 2024 23:41
|
|
|
All you have to do is filter an inserter to blacklist the current production item and it will automatically remove all ingredients of the prior recipe. Then you either belt or bot them back to their inputs to the mall. Sushi belts are hard because you need ratios, but in this case you don't need to manage ratios. You just need an output belt. Actually, now that I think about it for some cases you don't even need to filter the output, you just throw everything that comes out of the assemblers onto a belt and then you let inserters filter it into input/ output chests.
|
|
#
?
Jan 19, 2024 23:58
|
|
|
I like the latest FFF's text description on their steam announcement.Wube Software posted:Hello, I am personally a little wary of the refinery (and I presume chemical plants) flipping. It is convenient, but does feel like a reduction in the required complexity of building. The circuit controlled assemblers are cool but I'm not sure I'll use them myself. We shall have to see!
|
|
#
?
Jan 20, 2024 00:34
|
|
|
Having to gently caress with a few squares worth of pipe isn't the important kind of complexity. It's just busy work.
|
|
#
?
Jan 20, 2024 00:38
|
|
|
I think we're missing one key component that seems likely to show up in a yet-to-be-seen component at this point: a way to take a signal and get its ingredients. That way we can set requests on a requester chest or allow certain items to enter a belt. For the single assembler challenge idea, though, all you need is 2 chests and 3 inserters, looping all the outputs back to the input chest. I think it's an interesting combinator problem to start with a desired end product (such as a power armor 2) and recurse through the ingredients all the way back to raw plates, assembling each component in turn. Although there's another problem with that specific example: it needs electric engines (which take lube) and blue chips (which take acid) - and as far as I know we don't have a way to purge a pipe network via circuit logic, so that seems pretty difficult to do with a single assembler unless you somehow meter very exact amounts of fluid into the pipes.
|
|
#
?
Jan 20, 2024 00:41
|
|
|
Pumps can purge pipes as long as they have space to push it into.
|
|
#
?
Jan 20, 2024 01:22
|
|
|
Xerol posted:Just to be clear, we're talking about an unloading setup something like this, right? Ignoring exactly how the items get onto a belt, but you've got 4 belts going into and out of the balancer, one belt per train car? Yep, slightly different arrangement of inserters and belts but 1-4 cars unloading 4 belts into a 4-4 splitter into downstream consumers.
|
|
#
?
Jan 20, 2024 01:25
|
|
|
Magus42 posted:Having to gently caress with a few squares worth of pipe isn't the important kind of complexity. It's just busy work. You're not entirely wrong, but on the other hand allowing access to both left and right-handed refinery layouts feels slightly off. Thinking about it a bit more I think that making sure that flips and rotates work in all cases is better than perhaps confusing a new player with two types of mirrored oil buildings. I can just imagine the ridiculous spaghetti now, a new player's refinery complex with the odd flipped refinery for unknowable reasons is definitely better that occasionally having buttons just not work.
|
|
#
?
Jan 20, 2024 01:30
|
|
|
Are you certain the trains are coming in evenly loaded? The outcome you're describing should not happen.
|
|
#
?
Jan 20, 2024 01:30
|
|
|
K8.0 posted:Are you certain the trains are coming in evenly loaded? The outcome you're describing should not happen. My setup is also the same as what you displayed there (slightly different "unload from trains onto the belts", but the exact same balancer) and yes, evenly-loaded trains still asymmetrically unload with it. Consistently.
|
|
#
?
Jan 20, 2024 01:46
|
|
|
K8.0 posted:Are you certain the trains are coming in evenly loaded? The outcome you're describing should not happen. That, or the buffer chests are unbalanced. You'd have to stop the process, get buffer chests filled up, then manually send the unevenly unloaded train back to a patch to reload.
|
|
#
?
Jan 20, 2024 01:47
|
|
|
Can you post an image of the setup? Maybe there's something weird about your unloading.
|
|
#
?
Jan 20, 2024 02:10
|
|
|
the real scary thought is "what are modders going to do with the new circuit capabilities?"
|
|
#
?
Jan 20, 2024 02:10
|
|
|
Xerol posted:I think we're missing one key component that seems likely to show up in a yet-to-be-seen component at this point: a way to take a signal and get its ingredients. That way we can set requests on a requester chest or allow certain items to enter a belt. Look at the assembler in the first image. You can read the recipe off the assembler.
|
|
#
?
Jan 20, 2024 03:21
|
|
|
There's also the question if they're going to make recursive blueprints official. That one mod adds so much megaproject potential that it'd be nice to be official.
|
|
#
?
Jan 20, 2024 03:32
|
|
|
 pretty basic unloading station afaik. Only thing I can think of is sometimes with intermittent sinks the next train will arrive before the buffers empty and some kind of early unbalance in the chests is compounding or something, or I'm missing something obvious.
|
|
#
?
Jan 20, 2024 04:26
|
|
|
If you let that kind of station completely fill up it will get slightly unbalanced because of the additional belt buffering for the far wagons, but otherwise there should be nothing causing an imbalance. The 4x4 bowtie balancer is an ideal I/O throughput unlimited balancer, it won't get messed up regardless of which inputs or outputs are being used.
|
|
#
?
Jan 20, 2024 04:45
|
|
|
I doubt they will add recursive blueprints to vanilla. It's fairly advanced and in a FFF about Space Age they compared that to Space Ex and kept Age to a more friendly logical extension instead of the big leap that Space Ex goes to.
|
|
#
?
Jan 20, 2024 14:27
|
|
|
Abandoned my Exotic Industries playthrough where I do in most late-game oriented mods: at the expanding rocket infrastructure stage. There's some limit my brain hits where 'the factory must grow' loses to 'building another furnace stack is absolutely mind numbing'. Cool mod, though. I liked the new uranium processing and the energy crystal bootstrapping / replication but the tech cost scaling was a little too much for me. Switched over to Ultracube and holy poo poo this is wild. It feels like it fits the factorio mold so well - it's this super simple idea that just turns the whole game on its head. I've got my cube production bus that then shunts all the resources out to an ugly rear end not-cube-needing bus. But every process you add to the cube bus means it takes longer for the cube to circle back, so then you add more buffer but then the cube stays at each process longer which poses a similar problem. I just got blue science automated (limping along at like 20spm lmao) and unlocked cube splitting which seems neat in theory but I can't make too many products with it right now and it is a huge power hog. The very last thing I unlocked before taking a break was loaders (finally!) and I think those should be a really big help with some throughput problems on machines.
|
|
#
?
Jan 20, 2024 20:30
|
|
|
Since I finished cubing I needed a new mod to dive into until I got the cube bug again and I picked Nullius. I dunno, when the first concept one has to learn is organic chemistry to form hydrocarbons one starts to wonder how they got here. The conceit of the mod is highly appealing I just don't know if I can commit to learning its complexity.
|
|
#
?
Jan 21, 2024 04:19
|
|
|
Nullius is incredible! It also took me 4 attempts to beat it. A lot of the production chains are deep, even mid-tier materials like aluminum and batteries have a lot of inputs. The two big things in Nullius is that many processing recipes have two or more outputs which are required in different ratios, and if you can't figure out how to get rid of the thing you don't want, the whole chain will stall. Also due to the depth and complexity of the production system, it can be difficult to pinpoint what exactly is the limiting factor. Generally, you'll have to find an easy and automated solution to the dumping problem. Also, there are multiple ways to make the same material, and higher tier recipes are much faster and more efficient in every way. The end-game biology recipes are incredibly liberating. Some stumbling blocks I hit on my final run. You can stay on a belt+outpost+bot mall until you unlock Tier 3 buildings, at which point it's worth switching to a city block base. Everything that goes into a rocket has to be mass produced, because you'll wind up launching thousands, so be prepared to scale rockets and satellites. And of the hydrocarbons, propene is the one you'll need the most of, so figure out a blueprint for a self-contained propene factory.
|
|
#
?
Jan 21, 2024 06:54
|
|
|
I'm currently trying to set up a city-block-ish layout for Nullius, so somebody please check my work on a priority train system, to prioritize consuming byproducts over raw production. https://factoriobin.com/post/Y41_2tED1to_ET8s-EXPIRES How it works: All stations take 1-2 trains and output their capacity to a global network, with demand on the red wire and supply on the green wire. (Although TBH I might just reverse that, since red wire is cheaper to make than green wire.) All stations will also set a train limit of 0/1 based on that capacity being above 0, so I don't have to deal with fitting stackers into the city blocks. * High priority demand stations just output their capacity and set their train limit the usual way, based on how far below the desired stock level they fall. * High priority supply stations set their train limit as usual and multiply their capacity signal by the H-for-high signal, which is broadcast globally and has a value of 100, so I can separate high-priority signals from low-priority ones. This means that 4 trainloads of iron ore each at a high and a low priority station will give a combined signal of 404 for iron ore on the network. * Low-priority supply stations set their train limit by reading the global network and comparing demand to high-priority supply for their product, so if ((supply/100) - demand) < 0 that station is activated. EG, given a supply signal of 404 and a demand of 3, that gives a high-priority supply of 4, and (4-3) = 1 so the station is deactivated. If the demand was 5, then it'd be 4-5 = -1, so the station would be activated. * Low-priority demand stations do the same calculation as low-priority supply stations and activate if supply exceeds demand, but do not output their capacity, since they will just exist to void polluted liquids and/or for symmetry's sake. The upside is that this significantly simplifies train operations, since I can just do basic schedules like "Iron Loading <=> Iron Unloading" and rely on the usual train routing to handle everything. The downside is that it takes (up to) 7 combinators per station, which costs both space and resources.
|
|
#
?
Jan 21, 2024 07:28
|
|
|
I would just mod out the wire cost anyway because it is changing like that for 2.0
|
|
#
?
Jan 21, 2024 14:03
|
|
|
meowmeowmeowmeow posted:
The circuits are completely superfluous I assure you
|
|
#
?
Jan 21, 2024 16:41
|
|
|
Using inserters and chest and splitters will always work, but without a circuit you can't always guarentee full speed throughput of both the train and the belts. A stack inserter moving from wagon to chest moves a full stack each swing, but if the inserters pulling from the chests are even slightly unbalanced, as often happens with belt-only systems, then at times you can be down one or two inserters pulling from the car. Now that car unloads slower, and the train has to wait for the one wagon to finish after the rest, and this starts compounding after hours of operation. The standard balancer circuit avoids this by ensuring that each Train->Chest->Belt is always operating equally fast. The circuits are not superfluous, but you are correct in that they are not necessary as passive balancing gets you close enough. Here is a vanilla circuit unloader if you need one, slap a 4x4 balancer somewhere after it to complete it. code:
|
|
#
?
Jan 22, 2024 18:27
|
|
|
Ultracube is great. It does two things that make it really addictive 1. Force you to use the cool mechanics like circuit network & fluid mechanics that you can easily skip in vanilla 2. Introduce real resource scarcity so the game isn't just about building bigger and bigger The tech tree is non-linear, there's not strict order of prerequisites that you have to run through and you can put some real thought into what to research next. And since there is resource scarcity it feels like that decision actually matters. There's a tech I've been able to research for about 10 hours but I'm ignoring it because I'm a little afraid of it - it has been gnawing at my mind while i build my dumb little cube factory
|
|
#
?
Jan 23, 2024 20:23
|
|
|
|
| # ? May 26, 2024 15:23 |
|
|
M_Gargantua posted:Using inserters and chest and splitters will always work, but without a circuit you can't always guarentee full speed throughput of both the train and the belts. A stack inserter moving from wagon to chest moves a full stack each swing, but if the inserters pulling from the chests are even slightly unbalanced, as often happens with belt-only systems, then at times you can be down one or two inserters pulling from the car. Now that car unloads slower, and the train has to wait for the one wagon to finish after the rest, and this starts compounding after hours of operation. The throughput-unlimited balancer at the end ensures this already without the need for circuits limiting inserter throughput. Your problem might be that you're not sideloading the belts like meowmeowmeow is, which privileges some buffer chests over others. Regardless, meowmeowmeow can certainly remove the circuits in that screenshot without any loss in performance - belts or trains.
|
|
#
?
Jan 23, 2024 21:04
|
|