Main bus problem

[WHOISATLAS]

Hello,

I am trying for the first time to make a main bus using KOS guide, but im having issues with some materials pulled out from the bus, mainly iron plates and green circuits.
I dont know why so few are getting to the assembly line.

https://steamcommunity.com/profiles/76561197979164139/screenshots/
https://steamcommunity.com/profiles/76561197979164139/screenshots/

Any advices?

Thank you

[Hedning]

You don't have enough smelters

[Fel]

(it's the url from "share" that you want to paste here)
https://steamcommunity.com/sharedfiles/filedetails/?id=2247970948
https://steamcommunity.com/sharedfiles/filedetails/?id=2247971010

You are making a rookie's mistake, which is understandable since it's your first time with a bus.

For starters, your smelters are not even making a full belt per group for some reason, your iron only has 12 steel furnaces per side (48 in total) and 17 on each side for copper (34 in total), that is not enough to populate 4 belts.

12 steel furnaces per side is enough to fill a yellow belt but since you are using red belts you are producing a full belt worth of iron plates (and less than that of copper).
You then divide that between 4 belts, so each belts carries 1/4 of a belt of iron plates.

When branching it out, you are using a splitter, which means that in theory your machines are getting 1/8 of a red belt worth of iron plates.

This is of course assuming that all of the belts are moving and none are backed up but that's pretty much the reaon why your iron plates are only trickling to the green circuits for example.


Basically you need to increase your mining/smelting in order to produce much more plates.

[WHOISATLAS]

Thanks for your response, im gonna add more smelters
I tend to upgrade yellow belts to red ASAP but is keeping yellows somewhat a good technique? Since its easier to fill

Thanks again

[Fel]

Upgrading to red doubles your throughput so it's not necessarily bad but you typically don't need red everywhere and if you don't have a factory that uses that throughput it's a resource sink that doesn't really bring much.

[Ragnaman]

You might want to build bus that has 4 belts of iron plates, 4 belts of copper plates and 4 belts of green circuits (with their own dedicated supply of plates).
This kind of setup at least guarantees no issues for a vanilla game.

You dont have to build all the belt bus from 0 to 100 right away, just reserve enough space for such creation and smelters that would feed it. Then, as time goes complete the bus/smelters.

Basic underground belts go perfectly underneath the 4 wide belt bus. So you would build:
https://steamcommunity.com/sharedfiles/filedetails/?id=2248186608
Of course, reserve plenty of space for other materials, and a pathway to run fast :p
You can go crazy with even wider belt buses, but at such sizes, personally, I change to rail systems.

Always have 2 spaces between your 4belts and you should have no issues leeching from the bus with splitters. 4belt buses also allow to use the neat/compact belt balancer, larger than 4 usually get quite big.

[Hedning]

If you just keep one belt type it is easier to build because you don't have to think about what to bring. I would say as soon as the next type isn't prohibitively expensive then use only that. Having extra capacity on the belts is not a problem, but if it doesn't fill up and stop you should look to increase the production of the item it carries. The best of course is a belt that is both full and always moving, but if you spend too much time worrying about that then that is time you are not spending worrying about more important stuff.

Also I know some big youtubers like main bus designs, but it really isn't very efficient.

[Ragnaman]

Originally posted by Hedning:

but it really isn't very efficient.

Yes, correct, the main benefit of belt bus is simplified logistics and less strain on thinking/worrying about such things, just like you said.

Belt bus is perfect for science/starter base.

[Premu]

In my experience 4 lanes of iron, copper and green circuits is quite an overkill for early bases (aiming at 45 SPM with assemblers 2/ 75 SPM with assemblers 3).

For such bases I have two lanes of iron plates, three for copper, one for steel (which isn't even close to be fully saturated), one for green circuits, one for plastic, one for stone bricks, one for red circuits and a half belt for sulfur and blue circuits. My bus is "dynamic" - if enough is used up I'll remove a lane, and replace it with new input. (like replacing one copper belt with the green circuits belt)

If you make your bus unnecessarily larger you're wasting ressources and create an unneeded buffer. In the worst case you make yourself problems like you experienced.

[RiO]

Originally posted by Premu:

In my experience 4 lanes of iron, copper and green circuits is quite an overkill for early bases (aiming at 45 SPM with assemblers 2/ 75 SPM with assemblers 3).

4 lanes of green circuits are a minimal must for a full lane of red and a tiny bit of blue circuits.
Both of those directly and indirectly feed into the yellow utility science and purple production science.

Next to being murder on your red circuits, production science typically is also a major drain on steel. Unless you plan and build specifically to avoid it, you will almost always find yourself constrained by either a lack of red or a lack of steel once production science starts kicking into gear.


From my own experience, the minimal belts to start into the infinite sciences and launching rockets without draining a bus-based base dry, are 8 belts of copper and iron, 6 of each dedicated to green circuits. 1 full belt of red circuits, which needs approx. two belts of plastic and two belts of green circuits in the process. And a bit of blue circuits on a single belt, which should be fed by 2 more green circuit belts and half your red belt. (Ideally; you'll actually scale up to about 1.5 full red belts instead of having just 1 full belt. Red is needed for more than just blue circuits.)

For steel, you should be at around 1.5 belts worth to cover all its various needs and then dump what's left into the massive drain of electric furnaces, which you need as components for purple science.

And even this is just marginally OK.

[Premu]

Originally posted by RiO:

Originally posted by Premu:

In my experience 4 lanes of iron, copper and green circuits is quite an overkill for early bases (aiming at 45 SPM with assemblers 2/ 75 SPM with assemblers 3).

4 lanes of green circuits are a minimal must for a full lane of red and a tiny bit of blue circuits.
Both of those directly and indirectly feed into the yellow utility science and purple production science.

Next to being murder on your red circuits, production science typically is also a major drain on steel. Unless you plan and build specifically to avoid it, you will almost always find yourself constrained by either a lack of red or a lack of steel once production science starts kicking into gear.


From my own experience, the minimal belts to start into the infinite sciences and launching rockets without draining a bus-based base dry, are 8 belts of copper and iron, 6 of each dedicated to green circuits. 1 full belt of red circuits, which needs approx. two belts of plastic and two belts of green circuits in the process. And a bit of blue circuits on a single belt, which should be fed by 2 more green circuit belts and half your red belt. (Ideally; you'll actually scale up to about 1.5 full red belts instead of having just 1 full belt. Red is needed for more than just blue circuits.)

For steel, you should be at around 1.5 belts worth to cover all its various needs and then dump what's left into the massive drain of electric furnaces, which you need as components for purple science.

And even this is just marginally OK.

I talked about the pre-infinte phase and did the math for it (no beacons or productivity modules are involved). To achieve 75 SPM with assemblers 3 (all six packs) you need 15 assemblers producing green circuits, 37 producing red and 7 producing blue. You might add some extra as buffer. To load everything of this on the bus you only need 1 blue belt or 2 red belts for green circuits. For red and blue circuits half a belt is actually sufficient.

For steel you need overall 125 furnaces - for which half a blue belt or a single red belt would be enough. To actually feed 2 whole blue belts you'd need to quadruple the steel production.

Besides that, you'd need 2 blue belts/ 3 red belts of iron plates (AFTER the iron plates needed for the steel production have been consumed), 3 blue belts/4 red belts of copper and half a blue belt/one red belt of plastic.

This is calculated without starting the rocket, though - if your target is to just launch a rocket you can just stop all research and you can use such a bus easil to feed a rocket launch site. If you're going into infinite research you most likely want to go towards a mega base for which a single main bus is not practical anymore.

Of course having some extra production doesn't hurt so that your mall can produce stuff in parallel to your research. I have serious trouble to produce enough stuff in the mall to permanently consume two blue belts of green circuits and a whole belt of steel, though. ;)

So - always ask yoursef the question - what do you want to achieve? If you're just aiming at rather moderate production don't overbuild your bus needlessly.

[Hedning]

Originally posted by Ragnaman:

Belt bus is perfect for science/starter base.

Since a bus is a compromise that trades efficiency for a structure that some people may find easier to memorize I wouldn't say it is perfect for anything. I don't think they are very good for new players as they don't have experience of how much space things take. They don't have a memory to draw on.

I have watched quite a lot of youtuber "nilaus" recently, more specifically his deathworld let's play. I think he could be called a master of the bus design as it seems to be his preferred way to structure the base and he's very experienced with the game. Yet when I watch him play he is struggling with the same things I struggle with, ie he still has to move things around as his base grows (not less than me) to make room for more production and he still has to redraw belts, which takes time and can cause errors.

Comparing my base and his the main difference is just how far he has to go to get around his bus, how much space is just the belt infrastructure, and how much more resources he has to spend on belts, undergrounds and balancers. Constantly rebalancing the belts so that he ends up with 4*1/4 full belts instead of 1 full belt. For blue circuits he eventually "cheated" and didn't put the greens on the bus, rather taking the belts behind the bus and directly to the blue the same way I do, but with the difference that his belts going to blue circuits has to travel very far because blue is still placed along the bus, way down relative to green.

[RiO]

Originally posted by Hedning:

Yet when I watch him play he is struggling with the same things I struggle with, ie he still has to move things around as his base grows, to make room for more production and he still has to redraw belts

That's because something is always bottle-necking something. It's what you get if you build out by just-in-time satisfying demands with supply. The alternative is oversupplying and that's simply not efficient. Trains alievate the visibility of oversupply and lines stalling, because oversupply ends up being merged away into the trains' (un)loading chests. Eventually it'll still surface when your trains themselves end up stalling while unloading.

It's no better or worse. Just different.

Originally posted by Hedning:

For blue circuits he eventually "cheated" and didn't put the greens on the bus, rather taking the belts behind the bus and directly to the blue the same way I do, but with the difference that his belts going to blue circuits has to travel very far because blue is still placed along the bus, way down relative to green.

The bypass he did is no different than building a set of 4 dedicated bus branch-offs; i.e. don't split the lanes, just take them off the bus in their entirety. The difference is a bypass doesn't take the additional overhead of managing to put it onto the bus, only to take it off the bus 1 or 2 city blocks down. Also; in the grand scheme of things distance means nothing. The belts will fill up and compact. It's just initial latency of a few seconds; a minute at best.

Originally posted by Premu:

Of course having some extra production doesn't hurt so that your mall can produce stuff in parallel to your research. I have serious trouble to produce enough stuff in the mall to permanently consume two blue belts of green circuits and a whole belt of steel, though. ;)

It's not so much the player's production hub that's the issue wrt to green circuits, but ancillary stuff. You directly or indirectly need green circuits in bulk for e.g. laser turrets; solar power; modules; etc. as well. It's not just the production hub and the sciences' lines.

The three listed examples are also those that receive additional stress exactly when scaling out and up.

If you want to do anything outside of just launching the first rocket and 'winning the game', you need to take those into account. Just basing your calculations for supply and demand on purely <x> SPM is a lie at that point.

[Hedning]

It's not just different, it is worse. Let's say you start with one of each digits 1-4 and want to eventually have 4 of each. Well in a bus design your start will look like this:

1 * * * 2 * * * 3 * * * 4 * * * 

where * means space left open for future digits. In a "spaghetti" base it will look like this:

* * * *
* 1 2 *
* 3 4 *
* * * *

Clearly in the latter you have to spend a lot less time and resources in the beginning building infrastructure, and it is not harder to expand.

And yes, the point of the blue to green is that it is one case where there is so much saving on what you call "overhead" that even the most experienced bus player cannot justify it being on the bus. The same argument can be made with all other resources that you put on thebus. Taking the detour to the bus and constantly rebalancing them is a price you pay just to have the structure. The only benefit to that structure is that it is easier to remember for some. The price is there whether the structure helps you or not.

[Premu]

Originally posted by RiO:

Originally posted by Premu:

Of course having some extra production doesn't hurt so that your mall can produce stuff in parallel to your research. I have serious trouble to produce enough stuff in the mall to permanently consume two blue belts of green circuits and a whole belt of steel, though. ;)

It's not so much the player's production hub that's the issue wrt to green circuits, but ancillary stuff. You directly or indirectly need green circuits in bulk for e.g. laser turrets; solar power; modules; etc. as well. It's not just the production hub and the sciences' lines.

The three listed examples are also those that receive additional stress exactly when scaling out and up.

If you want to do anything outside of just launching the first rocket and 'winning the game', you need to take those into account. Just basing your calculations for supply and demand on purely <x> SPM is a lie at that point.

You have to take my post in the correct context. I wrote: using that many belts on a bus is not needed if you're aiming for 75 SPM.

And depending on the phase of the game and your overall goal this can be more than enough.

I'm currently constructing a 2000 SPM base. To research everything before the rocket I started a main bus-design balanced to 75 SPM. After everything was researched I tore down the whole science pack production, while the stuff needed for the mall remained.

For 2000 SPM this is obviously not enough - but currently I'm in the process to replace my main-bus design with dedicted city blocks for smelting, chemical products, electronic cirucits, etc, and substitute the existing parts in my base factory. For example - after my dedicated iron and steel smelting facility was ready to provide the plates and steel via train, I tore down my original smelters and deliverd those in larger quantity by train. Once my electronics factory is ready I can finally speed up my building process significantly as my main bottleneck for the modules will be gone.

My point is - do not just copy a design, understand it and adapt it to your needs. In my case I did not want to build up a massive production early on those construction would actually delay my research significantly, just to tear it down later. That seemed to be a waste of time for me.

And you definitively should avoid building a bus with four lanes for iron plates if you only have production capabilities to barely fill two of those. It just means you have more issues balancing your belts. You can reserve free spots to expand your bus if needed - but don't fill it out with belts until you actually need those for your throughput.

[RiO]

Originally posted by Premu:

And you definitively should avoid building a bus with four lanes for iron plates if you only have production capabilities to barely fill two of those. It just means you have more issues balancing your belts. You can reserve free spots to expand your bus if needed - but don't fill it out with belts until you actually need those for your throughput.

That, I think is the crux of using a main bus design effectively while scaling out from a starter base. A main bus doesn't have to be noticeably inefficient if you just take care to not over-engineer it. Just like everything else, it should scale with your needs.

Way too many people who say spaghetti bases work better because of less overhead managing belts, somehow fail to understand that part.

Originally posted by Hedning:

It's not just different, it is worse. Let's say you start with one of each digits 1-4 and want to eventually have 4 of each. Well in a bus design your start will look like this:

1 * * * 2 * * * 3 * * * 4 * * * 

where * means space left open for future digits. In a "spaghetti" base it will look like this:

* * * *
* 1 2 *
* 3 4 *
* * * *

Clearly in the latter you have to spend a lot less time and resources in the beginning building infrastructure, and it is not harder to expand.

There's more to efficiency than raw space usage; up-front investment and items 'wasting away' on belts.

The reality for a spaghetti base isn't that it's a neat and permanent build-once-and-never-look-back arrangement either. Your production and consumption of various intermediate products will fluctuate just as well over time. The difference in a bus design is that you have the space to amend the existing sub-factories to boost their output; draw in another line of raw resources to compensate for increased input; and output an additional belt if you end up needing to output above the capacity of a single belt.

In a spaghetti base, you can't do that. You either tear everything down and redo it; or you start a duplicate copy somewhere 3 to 4 screens further down in a totally unrelated space. Tracking down what your current bottle-neck product is, is also a lot harder in a spaghetti base than it is in a bus base where you can easily visually inspect the state of the mainline belts on the bus to see which ones are under-supplying.

A main bus is a paradigm that gives you easy expansion options; traceability for bottlenecks and good, easily navigable topology. Sure; it wastes some raw free space, but that is probably the cheapest resource in Factorio, save for if you're playing death-world. And sure; you need a few more belts to lay it down; but those are some of the cheapest items to produce in case of yellow belts. And sure; it takes a bit of planning up front, like maybe 10 minutes to plan out your stuff in map mode with a few markers, and plonking down some skeletal work for the bus's sets of 4 bonded belts. But at the end of it, you win all of that back easily compared to managing the muck of a spaghetti base.

There's a reason it works like that in the real world as well, you know. ;)