Space Age MEGA Bases

[majere119]

Alright addicts, lets see what y'all are cooking up for idiotic and completely unreasonable SPM numbers to test the limits of the sim now.

[solublefish]

Are these hiding someplace else? I'm polishing up my first large Space Age project and wonder where that community is and what people are interested in seeing shared. Also I don't know what metrics people are using, really.

I've designed for all 12 sciences, targeting one per tick of each for now. (12*60*60*60 = 43200 spm) I'm running at about 36000 spm and just tuning and testing to be sure it's stable over many hours and to push it closer to the design target.

In 1.x I believe this would comfortably be "megabase" terratory (though I've always thought they should be "kilobases" until we're at 1 million science per second). But of course with the new tools it's far easier to achieve these numbers. And of course in Space Age it's really multiple bases. My big Nauvis base only handles the 6 "easy" sciences, of course, so it's rating would be half. But my Aquilo base is hardly "mega" at all - it's pretty much the minimum for the target and mostly just hacked together.
I'm also thinking many people won't even be interested in doing all 12 sciences - there are legitimately different challenges with each.

[Tertius]

I don't understand your 43200 number. As far as I understand, SPM is counted per science, so if you're consuming all 12 different science packs at the same time, you have 1000 SPM if you consumed 1000 of each of them. You count this as 12000, but actually it's just 1000 SPM.

Additionally, if you say one per tick, that's 60 per second and 60*60=3600 per minute, so it's 3600 SPM you're aiming for. Not 43200 SPM.

60 per second/3600 SPM is one green belt per science pack being continuously consumed by the lab area. However, I don't know how quality science packs are being counted, since higher quality science packs have a probability to not being consumed by one lab cycle. I guess the community looks at the ingame production statistics, filter for "science pack", then look at the numbers shown.

[solublefish]

Yeah these are some of my questions.
Certainly I do "look at the ingame production statistics, filter for science pack, then look at the numbers shown". Of course there are multiple numbers and we want some kind of aggregate or summary. Yes, I'm doing about 3000 each and I think my build should support 3600 each (36K or 43K total) once I polish out some remaining bugs.
Regardless, it feels like it's not as simple as each vs total and I haven't seen much discussion of Space Age "mega".

• Tertius mentioned the quality wrinkle - of course different designs may use different levels or mixes. FWIW my understanding is that higher quality packs "drain" slower, rather than having a bonus chance. Admittedly I haven't played with them much and I could be confused. Regardless I don't know how they should be counted in a summary number. (I'm using only "normal" quality atm.)
• The agricultural (Gleba) science has less useful science per item than the others, since it's always at least somewhat spoiled by the time you use it. If you're producing, transporting, and consuming 1000 of each regular color, you might need anywhere from maybe 1200 to 1500 Gleba packs to come out even. In terms of logistics, that's real - you're physically shipping more items. But they drain faster so in terms of science value they're at a discount. (I'm using about about 40% "extra" atm.)
• Are we only considering "all-twelve" builds? I could see a lot of people just not being interested in scaling out Promethium. I know a lot would prefer to skip Agricultural. Personally I find EM (Fulgora) tedious at scale. (I'm doing them all atm.)
• I guess we're sharing whole games? There's kinda no such thing as an all-twelve "megabase" unless we're redefining "-base" to mean "5 planetary bases and some space platforms". (I have 5+18=23 surfaces atm.)
• I assume it'll be common to use Biolabs on Nauvis with their 50% "science pack drain". But it might be interesting to use regular labs on Gleba, Vulcanus, or in orbit. If your game does that, is your SPM number halved, doubled, or the same? (I'm using biolabs on Nauvis.)
• The game itself now reports the "science per minute" on the mouesover in the upper right. Surely that's the most obvious number people see, but it's an aggregate of some kind. I'm not actually even sure how it's calculated. I think it's the same as the "overall research output" in the stats graph. (Mine shows between 28K and 42K recently.)
• The "research productivity" infinite science increases well, your research productivity. I suppose we want to ignore that bonus since otherwise we'll end up mostly counting how long your game's been running? (I'm at +490% atm.)
• I guess it's not a new issue, but game settings seem important. Of course there are a zillion options and it may be impossible to compare between them. Maybe that's not even a goal. (I'm using railworld preset plus no enemies.)
• Similarly not a new thing, but presumably we're looking for some kind of sustainable average throughput? Hands off for 1 hour? 10? 100?

[Tertius]

The traditional (pre Space Age) SPM counting was actually the space science pack, because this was by far the most expensive one, and it boiled down to: "how many rockets can I launch per minute?" I don't know if you were around with 1.1, but the fastest launch cycle was 1 rocket per slightly more than 1 minute per rocket silo, and each rocket resulted in 1000 space science packs, so if someone says "I have a 3000 SPM base", he meant: "I'm able to do 3 rocket launches per minute". Military science was often left out while counting SPM, I guess because it required one of the bigger production lines and there was no infinite exponential research that included military science, so you had to manually add a new research every few minutes to not run out of researches.

In 1.1, the limiting factor was UPS. At some point, the engine was simply not fast enough to handle a base with even more SPM. So the optimizations went from "more SPM" to "more UPS" and many strange designs started to show up that primarily optimized UPS before throughput.

As far as I see, Space Age has not yet matured enough with the players to have a baseline of what is a good or a great number. But if you ask me, a megabase is only a full implementation of the game if it implements every game mechanics, so it has to include producing the promethium science pack, so the thing to research for benchmarking is research productivity, so a valid SPM number is what shows up in the stats screen for promethium science pack consumption (or any other pack) while researching research productivity. It makes no sense to pick just the easiest science packs - that's not competitive.

If you're unable to do research productivity, you have a growing but not yet finished (mega)base.

[solublefish]

Yeah I played ~500 hours of 1x but never really tried to participate in the online community. After a couple hundred hours in Space Age, and having had a good experience in the train intersection thread on this forum, I figured I'd look for the megabase group. But I guess it doesn't really exist here. The OP asked for "completely unreasonable" builds and I'm not even seeing "totally reasonable" endgame planets. They must exist - a quick look at the Galaxy of Fame shows plenty of big stuff. Of course those are mostly pre-promethium but some are from months ago. Is everyone on Reddit or Discord or something?

[charonme]

I'm interested in seeing a multi-k spm fulgora setup, does anyone please know where I can find some? I thought this would be the place

[eugenekay]

Great things take time to achieve.

I have been revisiting and redesigning my Blueprint Books and Production Setups continuously since Space Age released, but I have yet to “complete the game” by automating all Science Packs and scaling them Up - or Down? Quality and Recycling does peculiar things to Beacon maths. All of the Rail curves are different, the new Stack Inserters and Belts require change layout sizes, Power poles reach farther, more Circuit machine configurations to get lost in, the list goes on.

[solublefish]

I'm interested in seeing a multi-k spm fulgora setup.

Here's my Fulgora - 3600 EM science per minute.
It's neither pretty nor optimized, but it works.

02-18-2025, 14-18-12.png (793.16 KiB) Viewed 5352 times

02-18-2025, 14-18-24.png (7.62 MiB) Viewed 5352 times

02-18-2025, 14-18-34.png (7.26 MiB) Viewed 5352 times

Almost everything I've done on Fulgora is to save space, mostly for fun. Strictly speaking it probably would have been easier to just take more space. You can expand to more islands and use Foundation more aggressively and use lower-quality buildings and modules.

The actual production chain isn't too complicated. Mine scrap and recycle it. Cook the ice, stone and holmium ore into holmium solution for the electrolyte, superconductors, and science packs. About half the green circuits come from recycling blue circuits. I make some batteries to supplement the recycling. This isn't strictly necessary but it reduces the amount of scrap I have to deal with.

The main gimmick on Fulgora is, of course, all the extra stuff you have to deal with from recycling scrap. I think the easiest thing to do is set up a big bank of recyclers hooked to requester chests and some logic to recycle extras. I use a little circuitry and "set requests" but it can be done with just conditions on inserters. I have such a setup near the center of the base and really it's probably easier to use bots for everything.

But I also set up some mixed-cargo trains and "trash islands". I got annoyed by zillions of bots, plus in the long run it may be better for UPS. (Though the 2.0 engine appears to handle *lots* of bots.) Mostly this is just a pile of recyclers feeding themselves to empty the trains. I did also build a setup to improve the trashing speed by first making some things like centrifuges and landfill. This destroys more items per second in the same footprint, but isn't necessary. Note that re-re-recycling solid fuel is faster than burning it.

I'm using a fusion plant instead of stacks of accumulators, again just to save space. It's a ~5GW design based on one by dts: https://www.factorio.school/view/-OEifwlHmaB-_-GnTzVT Note that if you'd rather use accumulators, high quality ones are a good deal. No need to make high quality lightning rods.

[charonme]

Here's my Fulgora - 3600 EM science per minute.

Thanks, I'm currently testing this setup, it's supposed to be standalone and copied to multiple islands, so far I have 10 of them. They produce around 130EM science each from one turbo belt of recycled stuff.

fulgora_unit9_setup.png (3.38 MiB) Viewed 5274 times

[solublefish]

Thanks, I'm currently testing this setup,

Nice! I like the idea of using the same build on many islands. Especially the distributed rocket silo. And the rare quality buildings is a nice sweet spot. My hunch is you could fit 2 of these on most islands, depending on your map settings. I assume you have access to Foundation, otherwise it's a pain to make roboport coverage work. Not strictly required but very handy. I expect like me, your main challenge is efficiently recycling the extra junk and I see you're using some of the same tricks I stumbled on. If you have access to biter eggs, I think you'll find level 3 productivity modules valuable. Speed 3 less so, but they're easier to make.

[charonme]

I assume you have access to Foundation

I don't! I haven't even gone to Gleba yet and the roboports are local to that island only. Each of those island processing units is completely independent with its own silo, power, vault ruin mining site, and bot network.

[solublefish]

I don't! I haven't even gone to Gleba yet

Nice. I'd say you've done quite well on Fulgora and 1K EM SPM is rather overkill for now. But then, one of the great things about Factorio is how different players can dig down into different overlapping challenges.

[solublefish]

I've ironed out the last few hiccups and my first large Space Age base is running stably at its designed level. It's 3600 per minute (60 Hz) of each of 12 science packs, plus some partially spoiled Agricultural.

Here's 50 hours. That's not hands-off since that would be too boring. But over that time I've been been building out future-looking stuff - mostly chains for legendaries I'm not really using yet.

02-24-2025, 09-31-01.png (423.8 KiB) Viewed 4905 times

Here's a Galaxy of Fame link.

Everything is my own except:
- I'm using balancers compiled by raynquist
- I'm using a couple of fusion plants derived from dtds and www
- The 4L rail intersections are mine but they're based on Bocian's excellent 2L Celtic Knot.

I'll add posts about each surface rather than making this one huge. I can blueprint specific stuff if anyone is interested.
Nauvis
Vulcanus
Fulgora (posted earlier)
Gelba
Aquilo
"2B" space platforms
"Voyager" promethium ship

[solublefish]

I've ironed out the last few hiccups and my first large Space Age base is running stably at its designed level. It's 3600 per minute (60 Hz) of each of 12 science

Nauvis first, of course.

02-24-2025, 09-42-50.png (474.05 KiB) Viewed 4903 times

The first 6 sciences and biolabs live in the northernmost rail block. It takes in molten metals, coal, stone, and crude by rail. I hope to be able to clone this multiple times to scale out.
The builds are mostly straightforward - traditional belt and beacon strips. Normal quality machines, modules, inputs, and outputs. Most stuff is using stacked Express (blue) belts - 90 Hz per lane.
The raw inputs plus plastic are belted across the north, with circuits and science to the south. From west to east you have petrochemicals, yellow, purple, cyan, black, green, and red science. Then red circuits and green circuits. A few chunks are squeezed in sideways to save a little fooprint. No rocket supplies or construction materials are made here. The ratios are pretty tight and the input train stations are heavily utilized.
Notice that the molten metals are so much more efficient compared to furnaces that it's suprisingly easy for stone to become a bottleneck rather than iron or copper.

The lab array spans the south of this block and run 100% utilized. The 6 basic sciences come from the machines above, and the 6 "exosciences" come on mixed-content trains. All 12 are fed into the labs from the west. Agricultural (Gleba) science loops around the block - it's good to keep it moving so you can filter out spoilage as necessary. When running properly this isn't really a requirement, but it makes it a lot easier to deal with changes or if you research something that doesn't use Agricultural. Similarly the biolabs have spoilage output for that situation. That's the only use for bots in this block.

The next rail block contains the standby train depot and fuel station. (Plus it happens to have a crude oil field.) I'm using 2-4-0 trains on nuclear fuel. RHT, 4 lanes, 9 spacing. I've started making higher quality fuel but am not using it yet. Almost all the trains use the new interrupts to do on-demand routing and refueling. I'm sure there are better designs available now. In particular I think mine could be improved to better optimize the number that have to go back to the depot. I'm sure the rail signaling can be improved, and I can do better merges and lane changes. Probably that'll become necessary when I add science blocks.

The 3rd block is smelters. I'm still surprised how little space they take compared to 1.X. These aren't even high-quality builds. I'd like do central stone smelting here, though rails for purple science still need plain stone. As I scale out, if I get too much rail congestion I might move smelting near the mines, though that does involve dealing with calcite in a smarter way.


The 4th block contains... everything else. The "mall", rocket silos, cargo landing, fusion plant, and a few odds and ends. It's all a bit haphazard. Here I import the 6 exosciences, fusion fuel, calcite, and foreign building materials. Rocket supplies and most construction materials for export are made here.
The pile of circuits in the middle is my on-demand assembly contraption. I've seen the word "Automall" thrown around so I'm sure someone has a better version. There's a sizable area devoted to refining legendary biter eggs for productivity modules. That's mostly for future use - I'm currently only using them in a few targeted spots.
For scaling out, I'll probably need to handle the landing pad and imports more formally, but most of this won't need to grow much.

Outside the central rail blocks are just boring mines. I have a huge mining efficiency and railworld presets so I need very few. This is deliberate since I find building yet more mines pretty dull. There is a biter farm which is only slightly less boring since it has a special train to feed them bioflux and fetch the eggs.

[solublefish]

I've ironed out the last few hiccups and my first large Space Age base is running stably at its designed level. It's 3600 per minute (60 Hz) of each of 12 science

Vulcanus. Exports 60 Hz Metallurgic science, tungsten, turbo belts, Speed 3 modules, foundries.

02-24-2025, 09-49-45.png (298.44 KiB) Viewed 4898 times

The core of the base is the straighforward Metallurgic science build around a simple bus. There's a decent mall to the east, including Speed 3 and Turbo belt production.

A large chunk of the southwest is recently devoted to legendary tungsten and stone products. I'm not really using them yet, just preparing for the future. That area has sprawled a bit and will get a rebuild. Note: the Hazard Concrete and Hazard Refined Concrete recipes are very fast to build and recycle - very cheap extra quality chances.

Vulcanus is now fusion powered, though I went a long way on steam from acid neutralization. It's very easy to stamp down another 192MW whenever you need it.

No trains yet but I'll probably put one in when I have to build the next tungsten mine.

Note that this game has enemies off, which eliminates the devourers. I played a shorter game with enemies and I do like the devourers. The "territorial" design lets you build on Vulcanus and start to take advantage of the new tech, but puts up gates to serious expansion requiring work on other planets first. Of course like the biters, in the "endgame" they're more tedious than interesting so I expect most people to turn them off in large builds.

[solublefish]

I've ironed out the last few hiccups and my first large Space Age base is running stably at its designed level. It's 3600 per minute (60 Hz) of each of 12 science

Gleba. Exports ~100 Hz Agricultural science, carbon fiber.

02-24-2025, 09-53-10.png (208.48 KiB) Viewed 4895 times

Like many players (I think), It took me while to get used to building short lines and handling nutrients and spoilage at every step. For me one important key was treating excess inputs as a first-class problem needing planned logistics. Yes, it's useful to reduce wasted production as much as you can, but it's also valuable to let go of resources you're not going to use right away. Burn or recycle them if necessary. I am careful to process all the raw fruit first so I make plenty of seeds.

In the end I designed a pretty tight widget for science and stamped it down ~30 times. Carbon fiber and a few bits come from hacky variants of that. Pretty much all the nontrivial construction materials were imported. I'm still importing some LDS and blue circuits.

There's now a small chain for high quality materials for future use. It's fed only by "extras" for now.

No trains. The only outposts are the boilerplate farms. There are a bunch not shown in the screenshot.

Again I'm using fusion power now but most of the game ran on steam from burning pulp. You don't need much power since Biochambers use nutrients instead. Even now I'm only pulling about 500MW. That's mostly bots and the rest is rockets.

Note that this game has enemies off, which makes building out a Gleba base a lot easier. I did a shorter playthough with enemies on, and I do like the new pentapods. I did a pretty good job designing my build to avoid spoiling eggs, but I found I always still wanted turrets for backup. Once you have fusion and some military research, just plop down plenty of lasers. The 50% resistance barely matters at that point. I might like to see some pentapods with a flat resistance to encourage using the new tesla weapons.

[solublefish]

I've ironed out the last few hiccups and my first large Space Age base is running stably at its designed level. It's 3600 per minute (60 Hz) of each of 12 science

Aquilo. Exports 60 Hz Cryogenic science + Quantum Processors, lithium plates, fusion fuel, barrelled fluoroketone, fusion reactors and turbines, railguns for building Voyagers.

02-24-2025, 09-58-24.png (157.27 KiB) Viewed 4887 times

I'm least happy with my Aquilo base. It gets the job done with 2 cryogenic science strips, plus one for quantum processors. The main challenge was of course designing for the heat pipes. In the end that mostly came down to embracing tons of underground belts. I use bots for handling imported materials, but it doesn't take many so the penalty is manageable. One oddity is I'm making and burning a huge excess of fuel just to use up the ammonia that comes with the ice. I don't see a good way around that.

The ugly area in the middle is where I boostrapped averything and haven't finished cleaning up. There's a probably rebuild in my future anyway.

It's on fusion power now of course, but fission was critical to the build, both for power and heat. Even then I found it challenging to keep from stalling the water supply, and very hard to restart. If the roboports freeze or lose power, there's nothing you can do except take a trip to go fix it. Later I sent a couple of spidertrons out there to do my bidding. Low quality is fine - you just need the unfreezable fission power and portable roboport.

1 train to fetch lithium brine (railworld preset). It also carries solid fuel outbound to heat the outposts.