Miners - saturated belt - furnaces - saturated belt

[zOldBulldog]

I easily watched a dozen videos on the topic. I got some nice ideas but there seems to be a lot of conflicting information about the correct ratios to achieve a fully saturated belt on the back end.

What I *think* I got from all the videos for Electric Miners,YELLOW Belt, Stone Furnaces, Yellow Inserters is:

MINERS:
-- Most tutorials do an array of 2x7 for iron/copper/coal. Presumably 2x5 for stone but I'm not sure.
-- A belt balancer on its output, to ensure both lanes are equally loaded even when some miners exhaust their resources.
-- This should result on a fully loaded belt on its output side, but... I've seen math that shows that I really need twice as many. I can't explain the difference between the math videos and most tutorials tutorials.

FURNACES:
-- An array of 2x24 for iron/copper/coal/stone/steel.
-- Steel takes 5 times longer but takes 5 times the input. So the same array (perhaps with Blue Inserters on input to compensate for furnace load time) seems like it should work.
-- Stone Brick takes the same time, so the same array should work... but it takes twice the input. Here my brain tells me that I should have two lines of stone for input instead of 1, but my gut tells me that perhaps the furnaces are the limiting factor in the chain, so perhaps it would still produce a full compressed line of stone bricks on exit with just one input stone line.
-- What really puzzles me is that I've seen tutorials recommend/show 2x16 or 2x12 arrays for lines that feed "presumably fully compressed belts" to the main bus.

RED BELT (double the throughput):
-- Use Blue Belts for corners.
-- Double the miners (in line is fine).
-- Use Steel or Electric furnaces.
-- Upgrade Yellow to Blue inserters, use two Blue inserters in the inputs where there used to be one.

I am absolutely certain that I must have misinterpreted some stuff. Can you guys point out any errors on the above?


-----

On a related topic: When I start running out of raw resources in "belt range" and switch to trains for transport, is transporting raw ore to my smelters good enough for early/mid game? In other words, is a train per ore deposit fast enough to keep my smelters constantly running? Or does it make more sense then to dismantle my smelting arrays, add a smelter per ore deposit (or for a set of nearby ore deposits) and load the smelted products onto the train, which then would get piped directly into the bus?

[Zool]

Some of your information is simply outdated:
- For example, with the actual version of factorio, the corners should work without reducing the throughput of a belt. So no need to use a red corner in a yellow belt.
- The usage of underground belts might no longer influence the compression - at least thats what they wrote in a friday fact lately

[Zool]

Regarding trains:
Well, in general a train system works great for incomming ores. depending on the size of the deposit, you simply need more or bigger trains, or more efficient stations - telling a specific number of trains is nonsense anyways, bevause it differs with the distance, the deposit size and so on.

[zOldBulldog]

Some of your information is simply outdated:
- For example, with the actual version of factorio, the corners should work without reducing the throughput of a belt. So no need to use a red corner in a yellow belt.
- The usage of underground belts might no longer influence the compression - at least thats what they wrote in a friday fact lately

This is good news. I just upgraded to 0.16. I'm going to edit the original post to remove those things.

On the train issue, I think I was not very clear but you answered my question: Carry straight ore at first, then as the deposits get farther use bigger trains, then eventually smelt and process at least steel, stone bricks and other heavily used good closer to the deposits so that I carry less back to the bus.

[Zool]

Smelting at the ore site is rather inefficient appart from VERY early game.
Miner -> smelter -> factory

In mid-game, you carry the raw ore to the smelter setup, simply because its too much effort to build both smelting and mining at every new ore-field. You rather build a big smelting setup and carry the ore there from different directions.
miner -> train -> smelter -> factory

In lategame, it might even be an option to have the smelters completely seperated.
miner -> train -> smelter -> train -> factory

Why is this change?
In earlygame, everything is very temporary, and is just meant to help until a first basic setup is finished.
In midgame, the setups become increasingly bigger and more efficient, while your first mines slowly get depleted. So you need the trainlines to transport the ore to the smeltery.
In the endgame, everything is too small. You suddenly want 1 rocket per minute which requires huge amounts of resources.

[eradicator]

@zOldBulldog: This website has really good up-to-date info on all basic ratios... but honestly i don't recommend caring too much about ratios as a beginner. I think it's a much better learning experience to just fiddle around yourself and enjoy the "aha" moments =). Enjoy the time you can still naively "just play" the game, before the inevitable life-sucking phase of dreaming about optimization patterns beginns :P.
-----------------------

Smelting at the ore site is rather inefficient appart from VERY early game.
Miner -> smelter -> factory

You can't really make a universal statement for that. On-site smelting has some benefits even in post-endgame:

• It's easier to build and supply multiple smaller furnace arrays directly from mines. (More space, easier to have more belts)
• You can reduce the amount of inserter operations needed to get ore -> plate -> assembler which saves UPS.
• You can ship plates directly to outposts that need them, requiring fewer load/unload cycles as a centralized furnace and also requiring fewer trains, making the train network easier.

Centralized smelting works well if your whole factory is centralized and you don't need to load it back into trains. It often needs fewer furnaces, but tends to be less expandable due to space constraints.

TL;DR:
There's no "optimal" furnace strategy.

[zOldBulldog]

Update, after some testing and upgrading to 0.I6;

- Two 7x2 electric miner arrays seem to properly saturate a yellow belt. Since Red is twice the speed, it should take 4 such sets to saturate it.

- 48 (2x24) stone furnaces seem to properly handle a saturated incoming yellow belt of iron/copper ore and produce a saturated yellow belt of plates. Thus replacing the furnaces with steel/electric and red belt should work the same (twice the ore/plates of course.

[gsezz]

I generally don't care about the belts on ore fields. It's just too much effort. I stencil down my blueprint as often as possible. Some red belts have 50+ miners, some only 10. I squeeze all belts through a 4 to 4 or 8 to 8 balancer, that feeds into my train station. Done. As long as the belt to the station are full, it's fine. The miners aren't balanced, but why should I care? The throughput of the ore field will decay earlier, but I better spend my time building additional outposts. This doesn't only ensure future throughput, but also sheer quantity.

For the furnaces, I go with 24 electric on a yellow belt, merging two yellow into one red. (I don't like too long rows, I prefer a more square layout)
For the most time I build train to train smelting setups:
1x4-wagon train -> 4x red belt -> 8x yellow belt -> 8x24 furnaces -> 8x yellow belt -> 4x red belt -> 4x1-wagon train / 2x2-wagon train / as needed -> directly to production lines. Need more: build more of these.

[Aeternus]

Smelting at the ore site is rather inefficient appart from VERY early game.
Miner -> smelter -> factory

Depends on your method of handling ore/smelted products. If your train network is becoming overloaded, smelting at the ore site is preferable since you only require half the amount of trains - smelted copper/iron compresses into stacks of 100, ore in stacks of 50. A heavy central smelter however can make more efficient use of beaconing plus production modules to squeeze out more resources at the cost of power.

My own preference is to simply stamp down a square miner pattern on an ore field with one or two blue belts snaking through the field (4 - 8 if the field is huge) and dumping that straight into the station. Ore production will fluctuate as mines tap out anyway, so perfectly balancing mines is pointless, just send it to the station and balance/buffer it there. Problem solved.
At endgame, I'll have one or two 4-12-0 trains servicing each mine - long enough to plow through behemoth biters, short enough to be managable on a shared railgrid.

Calulating how many mines per belt is also tricky, as you start to invest in Mining Productivity research, the amount of ore produced per mine increases

[JimBarracus]

perfectly balancing mines is pointless

Ore fields also don't have an equal spread of ores per tile
In the middle there is a way higher ore density.

Just cover the complete ore field with miners and balance it at the station.