Introduction
A previous discussion yielded the point that a normal bus wastes a lot of space on unused capacity. I suggested a bus where, instead of a wide parallel bus with dedicated lanes, a narrower bus was used with time-shared lanes, and requested resources would be delivered in packets saturating the entire bus. Unsurprisingly, I was immediately challenged to make this happen, so here's my implementation.
Initially I tried to use a circuit network for signaling, but ran into the issue that signals are bidirectional. For this to work, we need signals to flow in one direction down a wire. This is doable by using combinators, but I didn't like how it was working out. Thus, I came up with the plan to use a single signal belt to handle requests, and a data belt to deliver them.
Thus, the Kanban bus was born (please forgive the mediocre video). It is named as such based on just-in-time logistics practice adored by industrial engineering professors, where a Kanban card is used to signal that more of something is needed.
Mechanism of operation
Conceptually, the Kanban bus is fairly simple. It consists of four parts:
- A main bus for carrying bulk materials
- A kanban belt which carries signal materials
- A set of source stations which releases a burst of items down the main bus, in response to a single signal item on the kanban belt
- A set of sink stations which release signal items when they run low on materials, and pull that item type off the main bus only when they need it.
The stations were a little bit tricky to make work though. The source stations are easiest; they are just a bunch of splitters that inject the items on demand. They use an integrator with a multiplier, so that each incoming item adds -960 to the count, and each outgoing item adds +1. The output belts are enabled when the count is less than 0, and pulse-count the items as they leave.
The sink stations were a bit harder. In the end, I only made them capable of 50% throughput -- because in practice I could double the number of belts without additional splitter hardware. This means that two material types can flow side by side without interfering. These stations require a bit more combinating. In general, the status is (-items desired) +(items in storage) + (items requested) - (items delivered). Items desired is specified by a constant combinator w
... keep reading on reddit ➡Hello everyone,
I recently started playing Factorio and I am have a lot of fun trying to com up with good blueprints. While there are already a lot of people that have made blueprints for megafactories full of trains, logistics robots, beacons and a massive main bus, I haven't seen a lot of is blueprints for the early-game. When resources are scarce, the main bus isn't yet operational and you only looking to make a handful of engines to make a car and two trains, it's not super important to balance assembler inputs and outputs, but you just want to automate some of your production - ideally with the least amount of materials invested.
Therefore I'd like to share with you you my Hand-fed Blueprint Compendium for automating production in the early-game in a hassle-free manner. These blueprints are not optimized for efficient production - the assembly line ratios are way off - but rather the aim is to make efficient use of player time and materials. The design should allow you to say 'hey, I need more of <X>', dump a bunch of materials in the right box and let the assemblers handle the rest.
To this end, the blueprints are mostly based around a cloverleaf pattern of assemblers that are primarily fed from a central box accepting basic ingredients (Iron Plates, Copper Plates, Steel, Stones). Intermediate products are made locally and the end products are output into the outer boxes. To keep set-up costs low, no belts are used and products are instead moved from box to assembler to box directly by inserters.
.
Example designs:
.
I hope this helps someone else!
Guys and girls! I'm finally ready to move on from my bus-base and level up production and I need a few hints and tips.
Usually I mine ore and send the ore via train to a "ore buffer base" where all ores get stored in a buffer-chain with chests storing thousands of ore. Then the ores gets shipped to iron plate manufacturing bases, steel manufacturing base and those products then gets sent to an own storage base where they then get shipped to bases where the goods are needed. And the chain goes on.
I have done this on my own earlier but I think I need to evolve my way of work.
- Is there a best practice design on a storage facility that many people use? Mine works but I feel that it's a bit ugly, and stupid.
- I'm using LTN (this is my first LTN factory) and I'm wondering if its possible in any way to place a blueprint which then tells a "local trainstationstation" which goods are needed to build the base and a train dispatches with everything needed. Roboports, constructionbots, materials and so on? Earlier I have just expanded my Roboport network on long distances which is... stupid. I want to do things a bit smarter now.
I'd love other ideas and pointers as well on how to do this in the most efficient way possible.
So a while ago I made myself an office chair from an old ugly neon bus seat. These seats came in pairs so I did the same to another one and gifted it to a teacher of mine, this resulted in people getting very interested in these seats. I began to purchase more from people parting out buses and began manufacturing and selling them to other teachers and friends. I was recently asked by a friend if I can legally do this, technically I haven’t done much to the original seats, usually only thoroughly cleaning them then bolting them to a modified office chair base. So this got me thinking, is this legal?
Here is my save file: https://drive.google.com/file/d/1LBbYKvUVQSVP_e6yghhsnEEIJCbRnt9c/view?usp=sharing
Thank you for your attention!
edit: Ideally I would also like to scale this using only signals to support one more train for each line.
