Factorio - automate ALL the things!

[KnightExemplar]

https://forums.factorio.com/viewtopic.php?f=3&t=46452

•Changed production science pack recipe to require assembling machine 1 instead of pumpjack.
•Changed science pack 3 to require electric mining drill instead of assembling machine 1.

Fuck. This would have made my life easier. But I'm well into High-Tech science now, and it'd be a pain to change my "starter base" to use these easier formulas. Fortunately, I'm in the middle of building my main base, and I haven't done my main-stage science yet. But my "Starter Base" is cranking away. I don't feel like redesigning my "starter base", but I'm many hours from getting a usable "main base".

I'll go for the new update only once my main-base is ready for science. But for now, I don't want to halt my science advancement due to the update.

The PA is finished and working quite fine.

You keep saying PA, and I hoped that I'd understand by now... does "PA" mean "Primary Area" ?? Or... something else? Very elegant base btw. It looks like you have a very well designed rail network, and I'm not seeing any obvious issues with your rail network design either.

[raudorn]

•Changed production science pack recipe to require assembling machine 1 instead of pumpjack.
•Changed science pack 3 to require electric mining drill instead of assembling machine 1.

Fuck.

Ha, yeah, that change came kinda out of left field. They filed it under balancing, so I guess they wanted to streamline the progression for new players even more and rebalance the iron/steel requirements.

You keep saying PA, and I hoped that I'd understand by now... does "PA" mean "Primary Area" ?? Or... something else? Very elegant base btw. It looks like you have a very well designed rail network, and I'm not seeing any obvious issues with your rail network design either.

Thanks, it's a rail design that I think I saw on some stream, but can't remember where. It does have some flaws that make it a poor choice for a true RPM+ megabase, but for everything smaller I guess it's fine.

PA means production area. As far as I know it's not a term usually used in the factorio community, but I use it for the area where all the item get produced, that aren't meant for consumption in the further production of science or rocket parts. I guess "the shop" would be fitting too, since you walk in, order a bunch of stuff and bots bring it to you.

[KnightExemplar]

•Changed production science pack recipe to require assembling machine 1 instead of pumpjack.
•Changed science pack 3 to require electric mining drill instead of assembling machine 1.

Fuck.

Ha, yeah, that change came kinda out of left field. They filed it under balancing, so I guess they wanted to streamline the progression for new players even more and rebalance the iron/steel requirements.

I agree with the changes. To get to my current point in my current map, I mined out ~4 different Iron Ore patches but only ~1 Copper patch. The new changes greatly reduce the Iron Ore consumption in the mid-game. Hmm, maybe I should update for my 0.15 run. The changes definitely make those late-game science packs cheaper overall.

You keep saying PA, and I hoped that I'd understand by now... does "PA" mean "Primary Area" ?? Or... something else? Very elegant base btw. It looks like you have a very well designed rail network, and I'm not seeing any obvious issues with your rail network design either.

Thanks, it's a rail design that I think I saw on some stream, but can't remember where. It does have some flaws that make it a poor choice for a true RPM+ megabase, but for everything smaller I guess it's fine.

PA means production area. As far as I know it's not a term usually used in the factorio community, but I use it for the area where all the item get produced, that aren't meant for consumption in the further production of science or rocket parts. I guess "the shop" would be fitting too, since you walk in, order a bunch of stuff and bots bring it to you.

Yeah, I think more people call it "the store" or "the shop". I personally call it the store.

-------------

I've been trying to look at everyone's rail designs. I want to hit 1-Rocket per Minute (especially now that rockets generate science, and that infinite-science gives a real goal to march towards). But my experiments with railroads demonstrates that bottlenecks are very common, and that the 2-rail network with 1-2 trains has maybe ~3500 Ore per second limitation... at least... a 1-1 Train has a ~2000 Ore per Second limitation and I'm assuming 1-2 trains can handle slightly less than double.

I've tried to wrap my head around a 4-rail network (2-rails in both directions), but I don't quite understand the balancing or how the bottlenecks would evolve.

My current design is to have to separate rails networks. I've named the two networks "Outposts" and "Smithy". These are connected by a Logistic Network. I know that at Megabase size, Logistic Networks are bottlenecked by the charging stations of Roboports, but ~75 Logistic Bots per Roboport... plus a large scattering of Roboports + thousands of bots should be able to handle the load (famous last words anybody??).

I'm hoping that resolves the bottleneck issue when the base grows to megabase size. But I'm many hours away from being able to accomplish this and test the idea...

[CorruptUser]

Got back into it since leaving a while back. So... much... to do...

[SuicideJunkie]

Something I set up just last week, which is quite nice; Train based bus service.

1) At various places around the map, I've got a pair of adjacent stops.
a) One stop is named "Bus_Pickup"
b) The other stop is named "Bus_{locationDescription}"
2) The bus_pickup is wired to a wall segment and gate sensor, which enables the stop whenever you are standing nearby.
3) There are two 1-0-1 trains on the grid (single-track) acting as buses.
a) Bus one has a schedule of I) Visit Bus_Pickup, stay until Gate=0, II) Visit Bus_DepotInbound
b) Bus two has a schedule of I) Visit Bus_{locationDescription}, stay 15s II) Visit Bus_DepotOutbound, stay forever. Then it repeats (I) and (II) for all destinations.

So, the way this works, is:
You walk up to a bus stop and the gate triggers the inbound bus, which shows up within a few seconds.
When you board the bus, the gate closes, and the train immediately takes off for the depot.
Get out at the depot, and board the outbound bus.
Choose a destination to go-to on the outbound bus. Logistics bots will swarm you with mall deliveries while you decide.
Hop out when you arrive and run off to do stuff.
You don't have to worry about leaving the trains parked in inconvenient locations, and they never get lost like the car does.
No matter which stop you got off at, or how far you wander on foot, the nearest bus stop is always the correct one
And they're filled with rocket fuel, so they're super quick.

I'd put in nuclear fuel for stupidly-fast bus service, but the 1% energy efficiency is so horrifyingly bad that I can't justify it to myself despite having a ~6 million uranium ore stockpile. (1.21 GJ each, is the equivalent of just over 100 GJ if the ore is processed into fuel cells in the power plant instead)


Previous games, I had an Uber train, but having to call it in manually was sub-optimal and I was often just going back to the mall anyways.

[Quizatzhaderac]

Just started playing a few days ago. Just rebuilt my factory to run off a memory sushi belt configured to produce whatever is in short supply. I'm glad I read the wiki, because I was about to attempt that before researching circuits (only filter feeders) which is a more involved solution.

I have yet to see any of the local mega-fauna. Default world generation.

My next big thing will be to go out and find an oilfield. Hopefully I'll be attacked in the process.

[raudorn]

I'm surprised you haven't had a run-in with the biters yet. Usually they start attacking soon once you got a smelting and a power setup, though the starting area might simply be pollution-friendly, e.g. with lots of trees or a big lake.

As for the sushi belt, that's quite an advanced technique and while it's not usually used for more than proof-of-concept, it certainly beats the bowl of spaghetti everyone cooks with their first base.

[CorruptUser]

I set up my games with a very large starting area and no enemy expansion and rare but large resources (a custom version of rail world), so I don't see attacks until around the time I've got laser turrets. I do tend to exterminate a few nests around the time I set up oil production, but my factory needs to expand...

[Quizatzhaderac]

There are a lot of trees around, maybe an abnormal amount. I did find enemies while looking for oil. There's a base of nine nests (and I've destroyed a few) right by the nearest oil field, with a worm that seems immune to shotgun damage.

I think I may have made a mistake setting up radar as earlier as I did. As I understand it, scanning a chunk causes it to be generated which starts the clock on base reproduction. My radar has detected a many bases within it's range.

A picked the sushi belt because it scales well with complexity, thus preparing me somewhat for the production of things I haven't learned about yet.
I was thinking in analogous terms of vertebrate circulatory systems where everything <-> blood, which is (in a way) simpler. My method of achieving feedback without a circuit network (which I didn't use) was based off of kidneys.

The system I currently have now has a green network doing hold read across my whole main belt.
My red network starts with constant combinators broadcasting the amounts of everything I want in the negative (if I want 40 iron bars on the belt, I broadcast -40).
Belts segments coming in from raw materials are connected to both networks, but don't read. They are activated if the amount for the resource I want is less than zero (the combined signal for both networks is howManyThereAre - howManyIWant).
Inserters feeding my assemblers work on that same logic. Each assembler is single purpose.
I have lights next to most assemblers and raw resource feeders on the same logic to show when work is going on. If the light is on a lot while the assembler isn't working I know I have a bottleneck.
I have a goodies chest/system designed to put certain numbers of certain things in a box, which I personally take whenever I want and get automatically replaced.

[CorruptUser]

IIRC, the aliens start colonizing and such regardless of whether or not you scan the area.

[SuicideJunkie]

I think the map operates a fixed distance out into the blackness. Scanning out further would activate more unscanned border tiles in that case, but overall the point remains that just because you can't see them, doesn't mean they're not there.

I started a Bobs + Angels game, and after a decent start I've slogged right down.
My little starter base is operating OK up to blue science, but my big efficiency expansion is killing me. I've got a big spaced out ore sorting & processing facility, that sends a train off to whatever ore patches it needs to pull from, and grinds out pure tier 1 & 2 ores to a huge line of train stations. (Iron through to silicon)
But those ores need a big smelting operation each to become plates efficiently, and I've only got Iron/Steel/ half copper out there.

I think I'm trying to reach too far on the long-process ore efficiency schemes. Haven't even got close to circuits on the outer base, and morale has fallen .

[Quizatzhaderac]

I quickly learned that the sushi belt it's terrible for some high throughput items (like simple circuits). Now that I know everything I might produce, I'd probably limit myself to a sushi belt for the labs.

Launched my rocket, figure I'm going to try and aim for having all the finite technologies researched. Still haven't done rails yet, limited to one belt on each good. Want to add a rail system for the sake of it. I will probably also redo my main factory.

I plan switch to nuclear power; I have 500+ panels now. My map is relatively coal poor so I've been trying to avoid using coal as fuel. I started collected uranium well before I could use it. I'll soon be have researched Kovarex enrichment and I have the 235 needed to use it. I also have the circuit network skill needed to make sure my nuclear plant is only on when I need it.

On the grounds that the circumference/ area ratio decreases with size, I've decided to build a 'great wall" surrounding my interests, producing an area I'm sure won't have enemies (they feel dirty). Mostly undefended, but I'll know whenever anything attacks. I might add artillery at points around the perimeter on the grounds for general bad-ass-itude.

[SuicideJunkie]

A laser fence may be easier than a stone wall. More expensive at first, but lower maintenance.
Just put packs of laser turrets around a large power pole, and string them out at the edge of your artillery range so the biters can't colonize your territory and make your artillery waste ammo.
If you carefully target arty at manual ranges to kill the spawners but not the worms, then they won't have space to recolonize outside of your fence either, and the attacks will tail off since worms can't spawn things.
Also useful is to put a line of yellow belts facing inwards at the edge of your laser range. While it does pull the enemies in towards you, they don't like to run on belts so overall they move slower. Having the belts push outwards leads to a pileup and causes the biters in the back to attack the belts.

For the nuclear plant, I'd suggest wiring up a system to monitor power draw via steam flow rates. You can then set off an alarm speaker if your draw gets too close to your plant's design capacity, so you know when to expand.
There's lots of room to get fancy with memory logic for a rolling average, and airlocking the steam so you can add one tank worth (25k) at a time and factor that in to your wattage logic array.

When you do trains, consider using rocket fuel with productivity and beacons to get more energy out than you pay in solid fuel + electricity.
With enough speed beacons it is an over-unity process, such that you can roll 100GJ of solid fuel in to a self-powered black box, and get 108GJ of rocket fuel out. The numbers are even better if you're feeding it solar (or nuclear) electricity instead of powering it with its own excess rocket fuel in those 50% efficient boilers.

[Quizatzhaderac]

I'm already placing ghosts of laser turrets around the wall. Once the wall it stable (I have clearing out convex indentations were I'll need to clear out an area to smooth out) and have my new power source I'll probably make them real. I'd have to seriously expand my power network to support that many lasers and I don't want to do that with a legacy system.

That's a good idea to monitor my power draw. The logic for my peak consumption should be pretty simple, but the logic for my average consumption swill be more interesting.

I also have some logic worked out for smelting outposts I'm looking forward to trying. Essentially it's a power switch that turns the whole thing off when idle (preventing draw).

I may also set up a grand network of roboports and radars so I don't actually have to go places to construct/deconstruct. I'm not too sure how efficient that would actually be though without putting the unplaced buildings near the construction site first.

[SuicideJunkie]

I've never bothered with a stone wall. The lasers do the job against any biters without blocking your own stuff, and spitters don't care about walls anyways.

Check if you can fit all the types of widget you use into a train, even if you go to just one stack of each.
If so, you can have a supply train drive around to the separate roboport networks in your base and deliver stuff. Use a constant combinator and a comparison combinator to control a filter inserter so it only grabs the things you're short on from the train (and only enables the robo-restock train stop when you're short on stuff).
The train can then head back to the mall and restock before heading out to the next zone that needs resupply.
Drills, assemblers, electric furnaces, inserters, belts, splitters, undergrounds, power poles, robots, roboports, turrets, boxes, rails, signals, stops, combinators... Might need a 2-car train to fit one stack of everything.

Not having bots try to fly a thousand tiles across your base's monolithic logistics network really improves their response times and helps them not cross enemy territory to get there.

[Quizatzhaderac]

One thing the wiki is seems confused on:

When a nuclear reactor overproduces, does it reach 900C and explode, or does it reach 1000C and waste output?

[raudorn]

It will sit at 999°C and waste all heat output not used by the heat exchangers. When the feature was developed the devs thought about making the reactor explode if too much overproduction exists, but then they decided against it. Of course, there is a mod for that.

[SuicideJunkie]

Also, the adjacency bonuses are huge.
A single reactor is not worth it unless you're desperate. A four-pack of reactors can squeeze 336GJ from every unit of U-235 you find (1.4x productivity on fuel cell production).
Definitely wire yourself up a control system to only insert fuel when your steam reserves are low and your water supply is adequate. Even if that's a manual operation at first, 24 GJ (4 cells x 8GJ x 300% efficiency due to adjacency) will last you a long time.

On the other hand, you've probably got millions of uranium ore units easily accessible, and if you run nuclear efficiently you'll never even exhaust your first mining drill. So you can be wasteful if you like.
Depending on richness, a single drill's mining area should give something on the order of an entire PJ after enrichment and reprocessing and productivity and adjacency bonus before being exhausted. (10k ore *1.2 * (1.2/10) to U * ~2 reprocessing * 14 to cells * 8GJ cells * 300% adj = 968 TJ)

[Quizatzhaderac]

My first setup is going to be 1 reactor, 4 exchangers, 4 steam tanks and 4 turbines. Fuel triggered when steam is below 4,000.

My second setup will be a 2X2 grid of reactors, 48 heat exchanges, 48 steam tanks, and 32 turbines. All four reactors input triggered by steam below 180,000.

I definitely have enough uranium to be wasteful (closest deposit was 666K), but I'll optimize for the sake of it.

Oh, and I figured out that the meltdown is if the reactor is too hot AND it is destroyed by external causes.

[CorruptUser]

Later on you get that method of turning 238 into 235. Takes quite a lot of time before you get enough 235 for that to happen but when it does...

Anyway, I only bother with uranium when I feel like setting up megacomplexes for circuits and such. Until then, solar megafarms are more than enough. It's a shame, really, I'm missing out on all the high end weaponry, which finally gives a reason to not just rely on lasers.

[Quizatzhaderac]

I think Kovarex enrichment may have actually been a bad design choice. For my planned goal of 160 MW usage, two uranium miners and without reprocessing or enrichment is enough. But I'm going to be tempted to be cheap and refine much of my 238, even though I have literal tons of it sitting around.

[SuicideJunkie]

I think the design is more about not having slag/waste products in the stock game.
Without the enrichment, you'll end up with bucketloads of waste that you can't get rid of, but still have to produce in order to keep your power plant running.

It would be absurd if you had to rig up a rube goldberg scheme with bots filling chests with uranium which, when full, trigger smoggy burners to attract biters, which destroy the chests and their contents in their rage, which turns off the burners, which then allows the bots to rebuild and refill the chests with more waste.

[CorruptUser]

It was designed that way for people who wanted to have all the weaponry. Before uranium bullets, there was very little incentive to not use laser defenses as lasers completely ignored armor. I mean sure, you had 3 (used to be 4) research paths for bullets instead of 2 for lasers so at the very end piercing bullets would end up doing more damage than lasers, but that was not only a long way off but it still required a complex logistics system to get a belt from your munitions factory to your walls (that or logistics robots). But with DU weapons, your uranium bullets already do more damage than lasers before getting an extra upgrade path for super-death. Before getting into the infinite research pathways, the multipliers are 2.2*2.2*2.5 or 12.1 for bullets, and 3.3*3.2 or 10.6 for lasers. Oh, and the gun turret already has 10/s shooting, or 3x the base shooting speed of lasers, so your uranium bullet with its damage of 24 is cranking out 24*12.1*10 or 2900 DPS vs the laser's 20*3.3*10.6 or 700. That's more than 4x the damage! Even with a Behemoth Biter's 12/10% physical damage, it will only block 1/3 of the bullet damage, so you are still getting close to 2k DPS with uranium bullets.

[Quizatzhaderac]

I finished my great wall. It's relaxing to actually have the area between my concerns were I'm sure there aren't nests. I spent a long time optimizing around the wrong things (which I'll insist is half the point of the game) so there's a very large density outside of my walls, and I get regular colony attacks in places (some places I've cleared out the nests and left only worms).

Nuclear power is up.

I'm rebuilding my base to a larger scale with train based transportation.

Planned locations:
Iron, copper, and coal mines.

Uranium mines: H2SO4 in by train; ore out.

Iron forge (well, smelting really), in and out stations.

Copper forge, in and out stations.

Steel forge: out station only, connected by belts to iron forge.

Oil ocean. In: Crude (pipes), water (pump), coal (train), iron (train), copper (train). Out: Plastic, explosives, rocket fuel, batteries, H2SO4 (train), lube (train).
I'll also have a spot where I can manually add steel and manually pick up flame thrower ammo.

Circuit city. Iron, copper, plastic, and H2SO4 in. Circuits, advanced circuits, processing units, speed modules and rocket control units out. Productivity three modules outputted to box for manual pickup. One station for leaving copper, one for other inputs, and one for outputs.

Carnival Capé. In: rocket fuel, Rocket control units, batteries, steel, copper, plastic, circuits. Output: Space science packs.

Academia. In: Circuits, processing units, speed modules, Iron, copper, lubricant, coal, steel, stone brick, batteries. Output: Science!

Pripyat. In: Uranium ore, iron. Out: supreme power.

I think the design is more about not having slag/waste products in the stock game.

I was thinking about that in regards to sulfur. In real life, sulfur is a by product of processing crude oil and much more is produced than can ever be used.

Mechanics-wise, realism would be terrible. There's the issue of creating a dump system. And there's also the fact the oil processing is complicated enough as is, without adding a 4th output.

[SuicideJunkie]

I don't know if it is an issue for you, but I like to set up my module assemblers as a cluster around a single chest, with circuits and low level modules going in.
That way, I can auto-trash level 1 & 2 modules into the stream when I'm upgrading to 3s.