Late-game marathon Ore to circuits

[vanatteveldt]

background

I am working on a "university", a modular science plant which is attached to my rail network that converts raw ore into science, aiming at 1 science per second of all types (including space). So, if I want 10 science per second, I would just put down 10 universities (and make sure there are enough outposts). Ore will be distributed through warehouses, i.e. each university will have trains from the warehouses, and each outposts will have trains to the warehouses.

The general setup of each university will be [station] -> [smelters] -> [plants]. I will use belts for the "bulk" transport of ore and plates, and bots for most other things.

The most resource consuming part of each uni by far is green circuit production. Total ore consumption per minute is 10.6k iron and 7.7k copper (according to Helmod), of which almost all copper (6.4k) and a third of the iron (3.5k) go to producing 3k circuits (by 3.2 plants!). The remainder of the iron goes to gears (2k iron to produce 700 gears) and steel (4.4k iron to produce 650 steel).

Inspired by PacifyerGrey's ore-2-bus designs (viewtopic.php?f=202&t=49452) I made a setup to convert 2 belts of copper ore and a belt of iron ore into green circuits. Ore and plates are moved by belt, cables by bot. It is possible to squeeze more circuits out of the setup (circuit plants are running at 82% capacity) by adding some more cables, but for my need this is enough and the overflow copper/iron is used elsewhere.

Compared to PG's designs mine is 3 rows instead of a single row, mostly because I use marathon, which means you need 2.5 times as many copper smelters and cable plants. Also, I've not found a way to move the cable from 3.5 cable plants to the green plant by belt or chest insertion, so I needed bots anyway and figured it was nicer to have 3x17 smelters/plants than 1x51. Since the cable plants are pretty close to the circuit plants on both sides, the distance covered by the bots is quite small (around 100 bots seem to suffice, depending on distances)


(full size image: https://i.imgur.com/INsMaKq.jpg)

blueprint

Code: Select all

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

calculations

See: https://docs.google.com/spreadsheets/d/ ... sp=sharing

The current setup produces ~1.5k green circuits from 2 belts of copper ore and 1 belt of iron ore. This requires 1.6 green circuit plants fed by 5.8 cable plants producing 11k cable. I've found it impossible to move that much cable any other way than by bot, even with direct insertion via chests I just can't get the cable moved around quickly enough (viewtopic.php?t=46989&p=271136#p271136). So, I decided to use belts to move ore and plates, and bots for the cable.

The copper goes through two identical setups: 10 smelters produce 1933 plates onto a belt, and 3 cable plants convert the plates from the belts into 5412 cables loaded into chests. 3 more smelters smelt (most of) the remaining ore to a provider chest or use elsewhere (red circuits, etc), which should be enough (total plate production is 10k out of 9.3k needed for 1/s science)

For the iron, 12 smelters produce 2320 plates on a belt. 2 circuits plants run on .82% capacity to consume 2160 iron plates and 10824 cable, producing 1515 circuits. The remaining iron is smelted into provider chests, producing 560 plates. Total iron consumption for 1/s science excepting green circuits is 3330 plates, so with the 560 overflow a single blue belt of iron ore can produce the needed extra 2770 plates (2400*1.2=2880).

[Engimage]

I have seen designs where people did move copper wires by belts but that took 2 to 3 underground belts (with stack inserter per belt) to move wires from one wire assembler to green circuit assembler. So it IS possible but it is far from being beautiful.

I have also found that making columns with 2 copper wire assemblers on sides of one green wire assemblers with direct insertion via 2-3 stack inserters (where green circuit assy is lined up with beacons to make wire ones shifted) work the best. So you can create a column having 2 of such groups. This makes green circuit assemblers idle much but in the end overall efficiency is the biggest.

[leitk]

I've had good results by moving wires from assembler -> chest -> chest -> assembler with the assemblers touching. This allows alternating lines of beacons to have their maximum effect, and still allows the throughput of chest to chest inserters. The limiting factor here is inserters, a 12 stack inserter has a throughput of a bit over 27 items a second. Bots don't improve that unless they allow you to fit more inserters on the side of an assembler.

[vanatteveldt]

I've had good results by moving wires from assembler -> chest -> chest -> assembler with the assemblers touching. This allows alternating lines of beacons to have their maximum effect, and still allows the throughput of chest to chest inserters. The limiting factor here is inserters, a 12 stack inserter has a throughput of a bit over 27 items a second. Bots don't improve that unless they allow you to fit more inserters on the side of an assembler.

Yeah, this works fine for non-marathon ... but on marathon you need about 3.5 cable plants per circuit plant, which means that you get in trouble placing the chests and also getting the inputs and outputs in place. I've tried in the past to get chest based insertion working but couldn't get enough throughput to get the plant working at capacity, and found it cheaper to just do 1-on-2 and accept lower output per circuit plant (viewtopic.php?t=46989&p=271136#p271136). If my analysis is correct, the problem is not inserting enough cable into the circuit plant (you can chain cables to four chests, giving you 27.7*4>110 cable per second), but extracting enough cable from each cable plant: as each cable plant needs to produce 110/3.5(>27.7) cable for perfect ratio, you need more than one inserter extracting cable out of each plant. Combined with needing to insert copper as well I'm not sure how to set this up, and never got to 100% capacity of the green plant.

I have seen designs where people did move copper wires by belts but that took 2 to 3 underground belts (with stack inserter per belt) to move wires from one wire assembler to green circuit assembler. So it IS possible but it is far from being beautiful.

Hmm, I dismissed underground belts as belt insertion is always slower than chest insertion, and frankly I don't see how it could work.

A circuit plant working at full capacity (4 prod3 + 8 sp3 beacons) has crafting speed 5.5, so it needs 2*5.5*10=110 cable per second (and 22 iron). According to the wiki, a stack inserter can move up to 13.58 items from a (backed up) belt to a container. To move 110 cable per second you would need 8 inserters, so I really don't see how that could work, as you would then require at least 10 inserters per plan? Do you have a blueprint/link, or was it for non-marathon?

In my current setup, I move iron on a belt, which would also be problematic (as 22 > 13.58), so I 'borrow' two tiles from the neighbouring plat to insert into my cable requesting chest. This does not allow full capacity as the 4 cable inserters are more or less 100% needed, but I only needed 80% capacity so that was fine. It's possible that a chest insertion solution with more circuit plants and fewer cable plants would ultimately be cheaper (in terms of modules and/or bots), but it would also complicate the plate logistics as now I can use separate columns for iron and copper.

I've tried to use a solution with two cable plants in separate columns that use an UG belt to pass the cable underneath the beacons, but can't get it to work:



this only produces around 410 cable per minute, or 44% of capacity (2* 5.5 * 60 * 1.4=924 max output of a circuit plant). This corresponds to the maximum inserter throughput of 12.2 for container->belt (4*12.2/110=44%).

So... I'm not sure how to avoid using bots for both cable and iron transport and still get max productivity, As said, the solution posted here uses belts for the iron transport but that reduces max capacity. And ideas?

[leitk]

I am playing marathon.

I have 4 cable assemblers feeding each circuit factory. It is more than needed, and the cable assemblers are idle sometimes.
It would work to have 3.5 cable assemblers, but that would need some creative juggling or bots. It would also need 2 inserters taking from each cable assembler as they can make over 30 a second and one inserter can't take that all out.
I am using bots to deliver the copper and iron, and remove the circuits, so it would make some sense to go to a 7:2 ratio instead of 4:1 (or 25:7 for a perfect ratio).

The circuit assembler needs 4 input cables, 1 input iron and one export circuit, so if it has other assemblers on each side, no room for anything else, if you're going with a 5.5 speed assembler you need all inserters to be going to/from chests.

[vanatteveldt]

I am playing marathon.

I have 4 cable assemblers feeding each circuit factory. It is more than needed, and the cable assemblers are idle sometimes.
It would work to have 3.5 cable assemblers, but that would need some creative juggling or bots. It would also need 2 inserters taking from each cable assembler as they can make over 30 a second and one inserter can't take that all out.
I am using bots to deliver the copper and iron, and remove the circuits, so it would make some sense to go to a 7:2 ratio instead of 4:1 (or 25:7 for a perfect ratio).

The circuit assembler needs 4 input cables, 1 input iron and one export circuit, so if it has other assemblers on each side, no room for anything else, if you're going with a 5.5 speed assembler you need all inserters to be going to/from chests.

Yeah, exactly... 4-1 is doable, but only marginally outperforms 2-1 on a throughput per module basis, and 2-1 at least allows you to use a belt for the copper plate.