Satisfactory: Manifold vs Load Balancing for Belts + Tutorial
Vložit
- čas přidán 1. 06. 2024
- Belts in Satisfactory can be hard, but there are two primary methods for getting items to where they need to go. Here is how they work and how you can use them in advanced scenarios.
Follow Satisfactory News on Twitter: / satisfactorycn
Check out my gaming channel: @masonzerogaming
0:00 Intro
0:26 Manifold basics
2:15 Load Balancing basics
3:42 It's math time, baby
5:41 Outro - Hry
![Eminem - Houdini [Official Music Video]](http://i.ytimg.com/vi/22tVWwmTie8/mqdefault.jpg)
If you're watching this, let it be known that I made a few mistakes which all of you helpful commentors have pointed out. Here is my follow-up video where I elaborate on the mistakes and important points that I missed: czcams.com/video/bNdn5aLkdh0/video.html
Manifold has one major advantage over load balancing - it's much easier to expand. Just add more machines onto the end of the manifold line.
That's a good point! Easily expandable!
You can also 'skip' one machine, and in this space add injector from another belt. Essentially expanding into two manifolds, when belt speed reaches limit.
yup, I also wondered who needs everything working immediately for the cost of so much space...
Indeed! And I use the fastest belt just for the main lane and slower belts from splitter into machines in that speed, what the required max input speed for item is. The main belt lane can be connected back to the beginning of itself, so it runs permantly in a circle. At the end of the main lane you can use an AI-Splitter to send your overflow to your storage for making stuff manually at the workbench or building more machines and structures. And if the storage is full, the rest goes automatically to the shredder for more shop-coupons.
If you plan ahead and maybe use some blueprints its actually quite easy to extend. I usually double or tripple my production along the beltspeed. So you only need to copy paste what you already had and add another top layer splitter.
If you expand to some strange number start with the first splitter and always use all exits. Any remaining splitter output goes into a manifold that refeeds at the top layer. Just repeat that process until only one splitter is left. Don't refeed the last splitter with a huge loadbalancer (>>27 machines) maybe avoid refeedind the second level as well. Or you will loose some beltspeed. With the blueprints it's actually really easy. Just copy past them a view times.
But it's not as easy as manifolds with additional feeders.
5:38 : "And then everything will be running smoothly."
In the background, the setup screeches to a halt at this precise moment.
Well, that is the ultimate satisfactory experience.
Yea his loopback was too early. You need to split it and feed 1 line to either side of the first splitter. You can't fit more than a full belt on a full belt lol so putting the loop into your full input belt defeats it's purpose a bit
"This is the one thing we didn't want to happen."
the smelters aren't running so of course it would stop?
@@dusk9910 Doesnt matter where it is. It stopped because the smelters aren't cooking.. lol
the idea of the loop back belt just gave me the biggest lightbulb moment for load balancing ever. i highly appreciate the way you explained it
Hey, a slight error at 1:18. The last 2 belts in manifold each output the same amount, in this case both 1/16.
Otherwise you would have 1/32 of the items phasing out of reality as 1/2 + 1/4 + 1/8 + 1/16 + 1/32 = 31/32
Great video for beginners though
That's true, I should have added "and so on..." to the graphic. I was trying to imply that every splitter will cut it in half again, no matter how you have. But yes, in the specific example here, you're totally right!
you can also use different belt speeds at your splitters to help load balance your manifold system to make them start up faster
Yep, if the machine itself only has a 30/min processing speed, then there's no point in using anything but a lvl 1 conveyor belt, and the difference in speed of the manifold vs the inserter belts will cause a back-up to occur much sooner
I have used this method a few times. Instead of just waiting for the machines back up, you create a controlled bottleneck that limits the throughput to each machine.
Only works to a certain extent though...
This is what I often do
I usually fill the last few manually if I want a constant work from the beginning.
great video man. i hadnt even considered manifold loading before but this was a great explanation of the two methods. the way you showed 5 way load balancing was really helpful too
I'm so glad it helped!
During the building process of a manifold I try to get resources moving early and each machine powered on accepting resources as soon as I can - this helps fill up the machines as I go creating the backup of resources that we want - it's also bizzarely helpful to make mistakes and end up with a surplus of items in your inventory to top up machines, which is helpful since I tend to make a lot of little mistakes
Same, when I am building a manifold I always start at the input and feed each machine as I setup the line to minimize the wait. Works great, by the time you are done almost everything is buffered unless you are making a longer manifold.
I had been doing load balancing intuitively since starting this game and it can get so complicated and had never even considered the manifold approach. It seems like I could create so much cleaner setups with it. Definitely going to start using it, glad I found this.
I've always referred to these as "serial" and "parallel". Works kind of like an electrical circuit. Serial mode (like a bunch of resistors in series) requires higher "voltage" (belt speed) to carry current to the end of the line. Just another way to think about it. Good video!
good video :) Another good thing to keep in mind is combining this logic with the intentional use if lower grade belts (closer to the input rate) to lessen the time it takes to flow throughout the system. you can even use the difference (aka spare capacity) of these belts, split them off and rejoin them in order to expand production while remaining efficient..
I usually use the manifold system. But I sometimes apply power to the first set of machine to prime them while I'm building the second set of machines that take the output of the first set of machines. So they're usually filled up already before the second set comes online.
I usually don't have too many issues with the manifold. I'll typically use an MK4 belt to feed the splitters and I'll use an mk1 belt to feed the machines. The rate at which the MK1 belts pulls the items off the MK4 belt can not keep up with the rate at which the MK4 belts moves the materials. I typically do the same on the outputs except I may use an MK2 or MK3 belt feeding into an MK4 belt on the merger side. The builds are typically much easier and cleaner looking and faster to set up. There are some cases where using the balancing method is the better option but more cases than not, the manifold is just fine. Also with early stage productions such as smelting and or construction of basic materials or parts, I'll usually add a storage buffer both before and after each stage of the production line. This helps to minimize shortages when starting to build more complex pieces that involves assemblers and manufacturers. There will typically be an ample supply of materials that are ready to be used instead of having to wait for the production lines to catch up. It does help to reduce bottlenecks.
I'm just going to point out that there's a third option here - A Hybrid of the two. For example:
Suppose you're feeding an array of coal generators via a mk4 belt. 480 coal per minute / 15 = 32 generators.
You *could* do a manifold and wait for it to stabilize.
You *could* do a big complicated 1 to 32 balancer.
You could also do a simple 1 to 8 balancer and feed 4 generators off of each of the 8 lines.
The pro's and con's should be fairly obvious but just in case: This sort of thinking will let your array stabilize much faster than a straight manifold while still being way easier and more compact to build than a 1 to 32 balancer.
this was seriously the best way i have ever seen this explained.... it finally clicked.
It's funny how all this information was out there, and these methods were what I just figured out through brute force of playing before realising they existed on the wiki 🤦♂ Very good and clean guide though, I think you captured each perfectly!
For me personally, I always use the belt balancing system for coal power, and always use the manifold (with oil pipes) for fuel generators. When it comes to production of items I usually get my available input in ingots, then use my big manual ingot:item ratio list to find out how many of the item I want that I can make, then will use balancers in that situation to get that amount. The great thing about this is that the efficiency is fantastic. The bad thing about this is that it's not expandable, so if you ever want to start producing more you need to either redo your system, or essentially get enough input to "double" your system capacity, which is very expensive in both resources and space.
Based on this, I think for me the best system to use is balancers when you know you are using your full capacity and don't expect to need to expand for a while e.g. I use a bunch of balancer systems for my Iron Ingot -> Heavy Modular Frame plant. However if you know you want to constantly be expanding a certain production (for example Fuel Generators, or say you have a factory that exclusively produces Screws to be shipped around) then the manifold is the best bet. Chances are in this situation you have no reason to get up and running instantly and are alright to let it run a bit before full efficiency. However it also makes it super easy to expand, so for example Screws are needed in massive quantities for a ton of different items, and so it's a good candidate for regular expansion. Using the manifold here means all you need to do is add on the extra constructors and input as necessary, and also lets you overclock at will since efficiency isn't so much of an issue as long as you know you have "enough" being output.
I somehow already learned this, but boy, this was a great video. I would really like to see more tutorials like this. I’m good enough at this game to get through the basics. But the intermediate stuff really slows me down. I just built my first fully automated modular frame assembly line. It was a monster. I had to use external resources to know how to balance the constructors. I really would like to be able to do that sort of thing on my own. But I have no idea how to do the math, even in a spreadsheet, to figure out how many constructors to place at which points, etc. Anyway, enough about that, this was a great video!
Thanks so much! This intermediate stuff can be hard for me as well. I didn't fully understand load balancing until I started writing this video, and the research process helped me understand it. This game is a constant learning experience!
Pen and paper has really helped me. I usually figure out how many items I want of a particular resource per minute then work backwards to the total number of raw resources. (I realize you probably already kind of do this, but my hope is that it helps someone else with their design thought process :D )
For example on Modular Frames, let's say I want 10 Modular Frames per minute. I would break it down like this (assuming all standard recipes):
-2 Frames/min * 5 Assemblers = 10/min
-----12 Iron Rods/min * 5 = ****60 Rods/min**** = 4 Constructors, 2 Smelters
-----3 Reinforced Plates/min * 5 Assemblers = 15/min
---------15 R.P. ÷ 5 R.P./min = 3 R.P. Assemblers
-------------60 Screws/min * 3 R.P. Assemblers = 180 Screws/min
-----------------180 Screws ÷ 40 Screws/min = 4.5 Constructors (round to 5)
--------------------5 Constructors * 10 Iron Rods/min = ****50 Iron Rods/min**** = = 4 Constructors = 2 Smelters
-------------30 Iron Plates/min * 3 R.P. Assemblers = ****90 Iron Plates/min**** = 5 constructors, 5 Smelters
From this we can see that we need a total of:
9 Smelters of Iron Ingots
5 Iron Plate Constructors
8 Iron Rod Constructors
----13 Total
3 Reinforced Iron Plate Assemblers
5 Modular Frame Assemblers
This is just math I've done in this comment section, and I'm sure I may have fudged up a few numbers. It's much easier to write down thoughts on paper and make annotations as you go along. I give specific shapes to my stuff when I write down my actual design but that comes down to personal style.
i do this in a google Docs or sheet document, really helps with not messing up your factorys, the pain begins when figure out that you will need 5 Iron nodes etc and start to bring in thousands of Items@@stealthninja9999
Ah, an age-old question with a simple answer. I hadn't thought of it that way, but I'll make sure to remember! Thanks!
when doing a Manifold you can use smart splitters to make it so it fills up the first machine before going to the second machine. it's very handy
I like to have the manifold input in the middle. Helps reduce the logarithmic fallout.
Oh.. You covered it as I was typing. Manifold injection.
Your rock man. You made this really easy to understand thank you!
For manifold efficiency, use your fastest belt for the core line, then slower belts into each of the machines. If the machine only needs 30 items a minute, you can use a 60 item per minute belt for example. This way the core line will back up much faster.
This is not true. The first half of machines will startup faster, but because of that you have less excess to fill all the inputs, and you need to have all inputs filled except the last two to be 100% efficient. So, it actually takes longer if you do this. The only way this is faster is if each machine needs exactly 60 or 120 items/min.
If you want your machines to startup the fastest with a manifold, use smart splitters to overflow items to the next machine when one is full.
@@jordanhoman0212 Never thought about this but you're totally right. Might actually use smart splitters for my next manifold blueprints!
Thank you for explaining the load balancing method since I wasn't completely clear on how it worked.
Honestly it was pretty confusing to me until I started working on this video. It's simple, but it LOOKS complex.
A manafold man myself but I'll admit, load balancing is a MUST for train loading buffers. Very helpful 101 video, thanks!
with a manifold I power each one up after the last one in line is full. I also manually fill each machine sometimes. As long as your input is equal to or slightly higher than consumption you are fine.
Same. Usually when i'm building, the "input" resource is already backup up. I just manually fill the excess resources into the "last" machines
Saw a comment about manifolds on reddit searched up the term found this video left pretty satisfied on what I learned
Very clear and informative tutorial, thank you.
This was perfectly explained thank you
I use the balanced manifold method ^^ - splitting everthing into 4 equal belts ,then go full manifold from there ,also using the different belt tiers helps with the balancing part ,@least where you need ressources in higher numbers
That sounds like a good strategy! Using the best parts of both.
I also use manifold most of the time but when you get to aluminum scrap for large setups the load balance makes life easier because the mk 5 belts cant handle the amount of items on them , this is when i start load balancing my lines , great video and info keep up the good work with this great game 👍
Thanks! I love the different ways to accomplish things in this game, and I hope some people learned something with this video!
Sometimes the best solution to a large system is to build it as a set of modular units. For my aluminum I build units consisting of: 3 refineries with sloppy alumina recipe. (1 using water from collectors and the other 2 using ONLY the by product water from the whole unit. ) Each feeds the alumina solution to 1 refinery making aluminum scrap by the base recipe, and in turn feeds aluminum scrap to 6 smelters making pure aluminum ingots. Then feed the output of each set of smelters into a shared belt to where I'm using the ingots. Need more: I just add more units. (Not 100% efficient, but I've never been able to keep the system running if the fresh and recycled water pipes have any connection at all.)
@@richardreid6377 At first I had a lot of trouble with connecting fresh water and recycled water as well. Than I used a few valves and a water tank. You must prevent back flow at all cost and (in my case) create a buffer with the tank. The tank can only fill with recycled water. You can split the recycled water pipe to create not one but a few merges with the freshwater pipe. So you priorize the recycled water when filling up the freshwater pipe.
My aluminium production runs extremly good (needed a lot for conveyers and other things) my steel stuff not so much because I was to lazy too optimise because I already had more than I could use xD
Both systems work reliably for almost every situation in my experience. I almost exclusively set up everything with a manifold. Be it 2 machines or 20. Manifold takes more time to equalize, balancing uses more real estate and brain power. I would say that it comes down to preference and how organized/complex you want your factory to look really.
You balancing method is better then my method. I normally just create a complex chain of splitters and mergers that produce the same result but costs more time resources and space thanks for the help
You're welcome! I hope it helps you be a more efficient FICSIT employee!
For a coal power plant or something similar, I like to use a "smart manifold". I use smart splitters to fill up the first, or first few, power plant(s) completely before allowing the coal to overflow to the next few machines. This makes a nice smooth ramp up in my power graph and prevents sputtering of the last few machines that are fed inadequately by the other two methods.
Thanks Mason, helps a lot!
I used to use load balancing when I can, but when I need 11.456 heat sinks to go into 7 assemblers manifolds just make it so much easier
I always knew the Manifold method as the Overflow Method and is the only way I set up my factories. With Lane mathatics and utilizing belts of different speeds, you can fill your line of machines to the max.
I had been doing load balancers to try to maintain efficiency from start-up, not realizing that manifold eventually balance out. I recently starting converting most of my basic production to manifolds, and I gotta say that opens up so much more space in my factories.
I set up manifolds for my coal power and didn't even realize it. My buddy didn't think it would work, but I told him to just turn it on and come back later. All 6 in 1 line stayed full. Even explained the math to how I knew we could over clock each one without running out of materials. Fun stuff. Still haven't built a proper factory yet though.
Fantastic explanations! Thank you. :)
I use both depending on given situations. I tend to use balancing on the output from smelters when you produce 3-5 or more items per cycle. That tends to load up belts and backup machines quickly.
I don't understand what you mean.
@@justintime5021 If you have for example - 10 foundries with an efficient combination of coal and iron ingots that is spitting out a high number of steel ingots quickly on a row; you need to figure out a way to unload all the items efficiently so that the foundries don't get filled up. If that happens it is a sign that you need to balance things out somewhere along the production line. Getting items offloaded and loaded into storage units quickly, usually requires some kind of belt balancing. As long as you keep items flowing on belts you keep the foundries working on shipping out items. Having the belts tightly packed at the same time requires some tinkering. It is the same principal for every type of manufacturing not just smelters, but smelters is always a good example because if there is something that you should always pay most attention to. It is having ingots flowing in your system.
Really no major downside to manifold method if you simply don't connect the machine output until they are all fully capped on both raw material and end product.
Secondly the best way to deal with load balance is the Modular Load Balancer mod. At least until they add a function to programmable splitters to set desired item/min in vanilla.
Definitely true, that's why I use it most of the time. The only time I used load balancing was in a system that had 5 nuclear reactors, but also letting it back up would have worked just fine.
Good call on the mod, although I just limit this channel to the vanilla game. Great options out there if you want to use them.
I'd say the only downside is when redioactive materials are involved. When these back up in the machinery they create a large hazard zone and manifold method guarantees they will back up. With load balancing you can avoid exactly that as long as you produce parts at the exact same speed as they are consumed and split them equally.
I use giant, loonnng manifolds.
@@anthonystahl8996 I think the most important thing about load-balancing is how incredibly satisfying it looks in action when instead of completely overloaded belt you get a separate items flowing in at fixed intervals and everything still works at 100% with next to no items in the internal storage.
or just simply run the line as you build the system, every thing should be topped off by the time you are finished
the day i found out about the 5/1 splitter system, this game became a million times more fun for me and the reason i tend to prefer load balancing.
I was struggling with understanding load balancing outside of 2's and 3's but this makes it sound super simple
I'm glad it helped! I honestly had trouble understanding it until I started writing for this video. Writing it down helped a lot.
@@SatisfactoryNews writing it down helps for sure lol. I've got my own pen and paper beside me doing maths while playing so it's important lol. This also makes it seem a bit easier than what I felt it was.
The only advantage of balancing vs using a manifold is that the production can start at 100% speed immediately. For a manifold system (especially if you do not optimize belt speeds) you have to wait for the whole production setup to reach 100% speed, at which point it operates exactly the same as if you perfectly balanced the belts. That's it, I wouldn't even call it an advantage personally since it doesn't really give you anything in a long run. Meanwhile manifold is easier to build, takes up less space and is SIGNIFICANTLY easier to scale up.
Outside of some niche applications with particular recipes that you want to ensure perfect rations for, manifold is THE way to go.
I knew how to use both methods but I didn't know taht loan blanacing would be balacend using loop. Now I know, thank you.
There could be other methods depending on the scenario but this is the basic way to do it
While Manifold is certainly space efficient. I have a hard time adjusting my thinking/play style off load balancing and I actually find the challenge of designing the system and shaping it with the world's surroundings an additional fun quirk to the design.
I usually build factories from resource node to end product, planning the numbers beforehand, but building 1 layer at a time, this way, by the time I get to the end, every line will be long saturated. In my first game I mostly did load balancing, but I progressively switched over entirely to manifold method for scalability and upgradability.
But with you mentioning nuclear waste I realise that it was probably this, why I kept having problems with waste buildup in my first nuclear power plant. Next time I make one I think I'll load balance them and make every particle accelerator and mixer a little overclocked to ensure no waste pile-up. Yeah. Putting both the mixer and accelerator on the same line was too sloppy in the end.
For sure. My issue was having several nuclear plants and the control rods were being built at the exact same speed as they were being consumed. So when using manifold, the first nuclear plant was getting most of the fuel rods and the last plant was getting basically none. Rather than waiting for hours for the rods to back up, I just set up a load balancing system so that the rods would be evenly distributed. Makes a big difference in this specific case, but usually it doesn't matter.
@@SatisfactoryNews I did the same thing, but I DID wait hours for the rods to build up. As I said, my problem was putting the mixers and particle accelerators handling the waste on the same line, which, looking back on it, could've only seemed like a good idea after the many many many many many other hours I spent setting up the rest of the power plant. In short, the whole system got clogged up. HARD. On top of that, there's the bug with Mk.2. Pipes that they don't carry the full 600 m3 water they advertise so generators just keep turning on and off, making my power output fluctuate and further screwing up the math on waste.
Something I do to help my Manifolds populate quicker is have the output belt go the opposite direction. So if I have the input start on machine 1, I'll have the primary output go from say machine 9. Early game this is much more viable because you are more limited by belt speed. Similarly, if you use an injection method where it feeds to the beginning and end of a row of machines, it can help greatly in 'warm up speed' of the line.
I also like to use a manifold for a long line, but insert midway instead of at one end.
That would be a hybrid where the injection method supplies left and right
There’s actually an alternative use for manifold by using the smart splitter and the overflow function. With this the first machines fill up faster and push the items more effective throughout the system. Using this backwards, means last machines first I noticed the system fills up quite good. Splitter > any undefined > splitter & overflow > machines.
Dude besides the video, I liked how you give geniuine replies on the comment section. So great... (I did not need the video content, yet liked and subscribed because of what i mentioned)
Thank you! This community loves to discuss the game and have intelligent conversations. It's the least I can do to reply to as many comments as I can.
Manifold mostly. But given the slowness of early game belt speed. I use the reverse of injection. Rather than trying to speed up the end, I speed up output, by adding an extra splitter in between the third and fourth splitter. This is then fed to the injection system further down the line, where delivered its seem to take forever to reach. You can tell when this is needed. The factories will be filling up with the item being made, and the items are backing up along the belts. As a bonus of doing it this way, you can easily remove these extra splitters and belts, when you get faster belts. This will make the entire thing more pleasing to the eye!
I use smart manifolds for all my setups. I also overload on machines so I will have more consumption than input available. the smart manifold makes sure only 1 machine is in that half on half off state. everything else will be running 100% or 0% and I can see which easy by looking at the lights on top. (as long as all the material is being used or consumed on the output)
I have been doing manifolds for a very long time, I tend to be a bit more interested in a clean setup then absolute efficiency. You can add a little bit more efficiency to the manifold by feeding your input into the center splitter in the lineup.
The manifold method is good for designing _very_ compact smelting foundries and constructor factories that you can plop down with one click of your mouse, hook it to the grid, connect the input/output belts, and be done with it.
Take the dimensions of the Smelter and Constructor, for example. In the blueprint designer you can lay a 4x4 foundation, then upon that foundation you can squeeze in a two-story microfactory with any combination of SIXTEEN smelters and/or Constructors with all of the necessary belts, splitters, lifters, and power poles with virtually no clipping save a for a few yellow splitters on the output side of the buildings.
And the best part? The blueprint preprograms the buildings to the input resource, output resource, and clock speed they were set at when you saved the blueprint. Want an iron foundry with an i/o rate equal to a mk 4 belt? Sixteen smelters at default clock speed. Did you unlock mk 5 belts? Upload the existing blueprint, add some power shards, upgrade the i/o belts, and reconstruct your new blueprint over the old one. Don't want to use shards? Just upgrade the belts on the existing module and drop a new foundry on top of the first one, hooking the input belts via a lift. Manifolds are great for easily building upwards, and the tutorial straight up tells you to prioritize building UP and not OUT.
Wait, you wanted a copper foundry? Just load your iron foundry into the blueprint designer, set all the smelters to copper, and save the new arrangement. Boom, you now have a 480/min copper foundry, all within a tidy 4x4 footprint.
With this design philosophy you only have to build a specific input/output factory once, on the blueprint platform; the rest is just fiddling with belts, clocking, and chaining.
Does it take time to fill up? Yes, but you are plopping down a compact microfactory with a _tiny_ footprint and connecting it to the supply chain in less than twenty seconds. I know this can be done with balancers, but feeding an entire supply chain a with a single input belt and cramming it with splitters and buildings is so much more space efficient.
Multi-input buildings fare better with balancing, but I'm an insufferable contrarian and have vowed to manifold *everything*. I will not rest until my entire factory runs purely on sushi belts.
I think that the manifold is better, it occupies less space and if you have stable productions and a regular amount of material arrives it is better, you just have to balance the line with different types of conveyor belts for balance, before starting production they have to be full It is more difficult to achieve this in productions that depend on gas or liquids, but for me it is a better system.
If i need to split into 5, i build 6 machines and reduce the clockspeed to 0.83, works always without any problems.
Sometimes that's just easier!
Wouldn't that use more power than five at 100%? Depending on where you're at in the game, power can be a limiting factor. I just unlocked oil in my first playthrough and am trying to future proof my build but math isn't my strong suit
@@Miss_Trillium I think you got it backwards. 6 machines at 0.833 uses less power than 5 machines at 1. At 0.833 speed, the power usage is 0.747 (multiplied by the building's base rate). So all combined, the 6 underclocked machines would use 4.48 power instead of the 5 power used in the 5 building setup. Underclocking always makes things more power efficient just like overclocking makes them less power efficient. This is precisely why people run twice as many constructors at 50% speed in early game to save on biofuel.
You can also start your manifold in the middle, like where you had the line coming up. would save space in the set up you had as well.
For sure, you should do whatever makes the most sense for your factory and how the rest of it is set up. I tried going for the most straightforward examples in this video.
I pretty much only use manifold, but I will prime the belts with materials first before starting everything up and it doesn’t take much time at all.
Damn, this was really helpful!
I'm glad it helped!
Great video!
I feel like next tips should be Factory set up, mini or standard versus mega where all the materials are pulled in
That's a great idea,, there is a lot to say about both!
Hybrid systems are also viable. Especially when dealing with high-ratio production like steel screws. When the 260 screws leave the constructor, it is better to balance them into the next stage (e.g. reinforced iron plates) because using manifold for the output of 20 steel screw constructors will require a ridiculous amount of mk5 belts
That's a great point! Both systems should be used, whenever they make the most sense.
Fairly new player. I note I have gravitated toward Load Balancing after experiencing the slow to start manifold method. My current setup has an injected manifold (Manifold Loading?) for Coal Power. But it just makes me nervous. If something goes wrong it's a dicey setup to get back up to stable, safe, full power. I tend to let the whole system back-fill before actually turning on the machines.
Sounds like you've already adopted some good habits! One thing you can do (although it takes up more space) is to include a storage bin as a buffer, that way if something goes wrong you have a surplus of coal to fill the coal generators with rather than waiting for your miners to catch up.
I pretty much exclusively use mmanifolds. I'm originally a Factorio player, so it more closely matches what I'm used to (inserters pulling resources off of a single line until it's empty or until the output is satisfied).
That's interesting, i will add my humble contribute, Wich is put the machines with the more part needs in the first part of the manifold to speed the full regime. Then, the load balancing and how to deal with numbers outside base 2 or 3 is also very interesting.
Manifold method with smart splitters is the best of both worlds. It starts up relatively quickly and is infinitely more scalable. Imagine having to tear down and rebuild all your production lines whenever you upgrade conveyer belts. With malifold you just add more machines.
Manifold is good for high throughput / low stack size factories. It doesn't matter if it takes a few minutes to start up. It's when you get to the end game and you have very low throughput, complex factories that load balancing becomes essential. Otherwise, most of your machines won't be running for a LONG time.
Something to help mitigate the slow start up of a manifold system is to hook up power while your still building, let your smelters build up while your building constructors, let your constructors fill up while building assemblers and etc.
using priority splitters in a manifold that prioritize the earlier buildings and just overflow into the next buildings is a very factorio-esque way of doing it, but I don't know if there's any change in efficiency from regular manifold splitters
Ah, load balancing, some of my favorite servers . . . oh wait, we're talking about the game. Good stuff.
Worth to note that load balancers with loopbacks don't work right on full and near full incoming belt since create a choke point where additional resources merge back.
As for example, you have 1 to 5 and 780 belt filled to 700 (140 into 5 directions). This won't work right because 1/6 will follow back and choke incoming flow since reach belt speed limit, back up and cause uneven split. At this point may as well use a manifold.
Furthermore, final speed of a loopback is actually higher than 1/6 of the incoming flow since flow is now faster due to loopback.
I often use the load balancing for nearly everything. math is pretty easy for them plus I like the consistency but the couple of times I have done the manifold method I have been pained by having to wait for the long start up which when combined to making modular factories. I want that thing working before I leave because if there is any issue I could have an important process I though I had up and running then I come back to it like an hour or two later to find out I had forgotten to have the recipes set on a couple of machines making it work at only partial efficiency. (I almost always load balance for any power solution because I like a reliable power grid that will never hiccup like I have had some in the past do.)
On that note: I believe super factories would benefit by the manifold system best while modular factories would benefit more from balancing overall due to the need of making the most out of your space on the super factories. The easy expansion of the manifold works better for the super factories while with modular factories you want to know it is working and at that efficiently before you leave it so the balanced method may be favorable. I would really be curious if speed runners would have a category created for tutorial-stage 4 of the space elevator. (Yes, they would be long speed runs but it really interests me to see how optimized they could get and one of the biggest things would be the manifold versus balanced conveyors, heck at some point I may be the first submission into such a category just because that is how interested with wanting to see others optimize the process. The most fun part would be to compare their runs to what I predict it would look like from thought experiments.)
You can just move a bit of material in by hand to get everything running faster at the beginning. I think this method is the more efficient. You can build just one assembly line and than expand as necesarry, you safe space and you potentialy safe time because you need less planning ahead.
You wrote that you like a stably power grid. If you have a LOT of batterys a few up and downs in your power consumtion (and generation mostly do to gysiers) don't matter xD
I really wish that they added another better battery. There are small and large tanks for liquids, why no large batterys? :(
I made a "playthrough" with nearly no foundations, really little optimisation, a lot of clipping, stacking stuff (you can build conveyers not only onto each other but even on power poles!) and it was not only a total mess but made a lot of fun. Can only recommend xD
Load balancing is (usually) a waste of space, time and resources... And ficsit does not waste
Great Video! For me the only way is Manifolds because I build my factories in steps for example: Iron Ore to Ingots then to screws and so on and while I am building the screw Part the Ingots can already fill up. At the end I just wait until every last machine is full and then connect to a storage or Train Station
That's pretty much what I do as well.
That cool, and interesting, not certain how useful it is though. Manifold is more space efficient and after it gets running fully(very short time if primed) works just as well. And late game with 50+ machines is pretty well a must.
I’m working on a 200+ coal generator build right now and have about 8 rows of manifolds. To load balance then would be nice but completely infeasible (actually impossible given space constraints)
There are no space constraints in the sky....
I like to use Load Balancing whenever it is possible. And not only for Inputs, but also for Material Outputs. It can save Building Material because you need less Splitters or Merger.
For Example: If i want to feed 6 Machines in a efficient Way, i only need 3 Splitters. One to Split the Incoming Main Belt into two and the other two to Split these into 2x 3 Belts.
I don't know if I'm right about that, but I think, each splitter and merger that you place on the map must and will probably be calculated individually. The fewer of them I place, the longer Satisfactory stays smooth. So I always try to place the amount of splitters and mergers as low as possible.
For 6 melting furnaces with a manifold I would have to place 6 splitters at the entrance and 6 mergers at the exit. Here I prefer to choose load balancing, placing only 3 splitters at the input and 3 mergers at the output. As your own factory grows, the difference will certainly be felt over time.
I actually use a combination of load balancing and manifolds. Split original resources into a few equal parts at the entrance to the factory, then create separate manifold systems for each. That way, startup time isn’t too long and I don’t have a complicated system to load balance at the beginning.
I don't think it was mentioned that another advantage to manifold method is how easy it is to expand a production line later on. since it all works on a straight line and a single "main" belt, it's super easy to add 1 more splitter/merger for each additional building you tack on, and since you can snake your production line quite easily, you don't have to worry about "future proofing" during your initial build. expanding a load balancing system however gets more complicated the larger it becomes since it branches out like a tree, adding an additional "branch" can take many more splitters and rearranging belts, and if you didn't give yourself plenty of space for future expansion you might end up needing to redo a ton of work.
Great point! Manifold is going to be the easier method for just about every reason imaginable, and that's one of them. I think that matters more in the early game though. If you're an advanced player and near the endgame, you can plan your factories a lot better and create load balancing systems that are designed to be permanent.
@@SatisfactoryNews That is the key part for me. If I plan on keeping a setup for a long time then I take the time to build the load balance. but if I am building a slap dash factory to get a specific part for an upgrade and not sure how important that part is going to be later in the trees I'm not putting the extra effort. The only exception to this is a coal power or fuel power setup where a load balance system is basically wasted effort and space.
Sadly, all builds reach a point where you have to start a new line. Load Balance just happens far faster, due to changing the setup requires far more space. Best tip I can give is don't worry with efficient until you reach Phase 3, then make a true "Starter Base" I refer to this when I mean basic material efficiency setup. All base Iron, Copper and Limestone along with Re-enforced plates, Rotors and Modular frames. This small step will greatly make transitioning into and beyond Milestone 4 much easier. This will also have you primed for transitioning quickly into stable power and stop your Greenpeace rallies at your HUB. This is by far the ting I see most often with newer players.
Secondary tip is early into Phase 3, don't try to automate anything beyond the basics I listed above. Just setup Dump fill mini lines. I normally setup the small storages with required machines to just produce those Phase 3 items while I'm off getting HD's and doing MAM research. I mean, your time is you only true resource that you waste. Everything else is opportunity for growing and learning.
One last point...
Don't forget that Starter Base area is only just till you are able to branch out. It will stay that way the whole time. So just keep flowing.
You can speed up startup of a manifold by grabbing the excess resources from the first smelter/constructor/etc and using those to prime the rest of the system
Another way to make manifolds easier is "system flooding." By having the lines running before the machines are even turned on and also manually inputting the resources into the machine to get the system started up the manifold should run flawlessly as the belts catch up with you manual loading.
There are two options for this: Perfect Manifold, we need splitters with system that determines how much IPM for a certain side. Working like valve, this'll evenly split the manifolds like perfect split or load balancing.
Second option is to get the mod "Buildable Resouece Nodes" and feed all machines with their own resource node.
So in a manifold system, do you have to make more of secondary buildings in order to keep the first ones from bottlenecking? I'm using the manifold system but I've come to a point where only the first few smelters are producing and the rest in line are not contributing anymore. I don't know if this makes sense.
Yeah that's the problem with manifold, but it should be fixable. Are your belts giving your smelters the proper amount of materials? Like if the smelters can process 120 ore per minute, are they receiving 120 ore per minute? If yes, then you can temporarily pause whatever machines come after the smelters, so that the smelters have time to fill with materials and back up. Once the input belt is fully backed up and stopped you can turn everything back on, and the system should work forever. As long as the math is correct.
Where applicable, I use manifolds with different speeds of belts. If the destination inputs are
When I first started playing Satisfactory, I used the load balancing method, because I didn't even know about the manifold method. It didn't even occur to me that you could use splitters and mergers any other way until I saw someone build a manifold on CZcams. I understood immediately what I was seeing and had a "Duh, I'm an idiot!" moment.
I kinda mix and match myself. Sometimes I'll use manifolds, sometimes balancing, depending on how I feel about it. For instance for smelters and foundries I tend to go with balancing, and for coal generators I tend to use manifolds. Not for any specific reason. I just...feel like I want to build it that way. :P That does cause the issue of the last ones in the line taking a long time to fill up and run properly, but eh the game is relaxed anyway. I don't mind watching youtube for like 10~15 minutes to wait for stuff to fill up. :P
I can understand how balancing helps a lot for nuclear waste though. You generally want to get rid of that as quickly and efficiently as you can.
late to the party here, but the main downside to manifold is eliminated by setting up your first-phase production before setting up the rest of the factory.
your factory won't produce anything until it's finished anyway, so letting the first-phase production fill up their buffer inventory as everything else is being built ENTIRELY does away with the slow startup speeds.
even if you don't do that, using your fastest tier of belts between splitters (and only the speed needed as the input into your machines) solves the problem. a low-speed belt can only hold a fraction of what a high-speed belt can, so the drawback of slow-startup is mitigated.
this solution is probably the one that requires the least change in playstyle, so personally, i think it's the best option.
to add to that, you should really only use load balancers on raw item input and output. as in... ores onto belts, ores into ingots, ingots into belts, and ingots into factory input.
only balance your load where the load needs to be balanced. think of a belt splitter more like factorio's inserter rather than its splitter. yes, it does still perform the function of a factorio splitter, but you aren't placing splitters EVERYWHERE in factorio.
I do the manifold with smart splitters and set it to overflow on each output. It takes a minute to fill up but each machine will only grab what it needs when it needs it.
I think you can do that with regular splitters because when one of the outputs of a splitter is filled, it just ignores it and starts balancing between the two other outputs.
i used manifold because it looked aesthetically pleasing. I realized it's flaws as I saw how materials would need to wait but didn't know what load balancing was.
Good explanation!
Thank you!
I always go for the "tree structure" of splitters as I usually try to use the right amount of everything, raw mats to final products. It's also interesting for splitting 240 to 2x120, it's not always enough to use lvl 120max belts, i.e. the math in the game is not at all perfect, stuff can lag, over or under produce, etc.
Also, I was waiting for smart splitter in hope it's about load balancing - but it was a huge disappointment.
I wish there was programmability in the game as well but that would lead to serious performance issues I guess.
Further down the Caterium tree is the Programmable Splitter, which keeps the filtering properties of the Smart Splitter but also elts you manually set split ratios. So with that you could have a single programmable splitter split at a 2/1/2 ratio, then have each of those two larger outputs split 1/1 with a normal splitter, resulting in 5 equally split lines
Im now 100 hours into sastifactory and i honestly never knew manifolds existed , but im gonna try them out and hopefully my brain doesnt melt
I usually use Manifolds for things like ore processing, while Balancers for Concrete,
load balancing is sometimes required in the early game where the material you need to move to be most efficient is significantly higher than the speed of your belts. You could have a manifold, but you'll essentially be wasting sometime 50% of your ore. Also I default to load balancing because it's more challenging and makes every factory its own unique web of spaghetti
manifolds can be balanced a little bit better if you input into the middle of the manifold and split out to two different manifold sections. won't be a big deal for 3-5 machines, but for large numbers it will be helpful.
I guessed the manifold by myself quite quickly while working on satisfactory setups, it's really easy to scale up (even vertically since you have the "back" slot always available for a vertical convoyer), is clean, simple to put in place etc. The main downside for me isn't the balancing but the fact that the main convoyer belt item rate is nerfing the system, so it's hard before mk4 to do big setups :x Obviously the injector is the solution, but I use it only when really needed.
The load balancing is fine, but a lot harder to scale and to make it look clean, and there is in my opinion almost no benefit from it, maybe for the nuclear waste but since it doesn't solve the problem of item rate it's not so good :/
Anyway thanks for this video, it was hard to explain to my friends how the manifold is morking without showing it ^^"
That's awesome! And I completely agree. Manifold does the job 99% of the time.
I use belt speed limitation as a motivation to use verticality a bit more :-)
Note that the loopback method can cause issues if you're not careful of belt speed. For example, you want to feed 5 machines 12 things/min, for a total of 60. 60 in gets split into 6, 10 goes into each machine, and the 6th line loops back to eventually send the remaining 2 into each machine. But, and here is where things can go wrong for you, now you have 60+10 on the belt between the merger and the first splitter. If you did the whole system with Mk1 belts, because your total input was only 60, right, now you've got a bottleneck that's going to clog up, and the machines are only going to get 10 each.
Yep, I made a whole follow up video about this oversight! (Looks like you found it though)
@5:03 so effectively you decrease the max size of a system by running some % of your items back through your core source feed belt (effectively adding a multiplier to it's throughout limiting total throughput). So you become belt volume limited before you would with a manifold system. Practically speaking manifold is easier from a design perspective and the inefficiency is only in starting the system, which is negligible. The inefficiency at start of a manifold system also allows for a stockpile of the resource in the machine otherwise known as a buffer, something valuable should there be an interruption in supply source. Manifold is better for probably 90% of players but the math nerds that love the math and or logistical side of the 'idea' of perfect efficiency will go splitting. Manifolds can be started instantly while under construction as well, split systems must wait until the full belt design is assembled.
Perfectly said. The load balancer is for nerds. Sometimes it's fun being a nerd about it. But manifold gets the job done almost every time.
Great vid
It's worth pointing out that splitting takes time. Not much, but if you split a 60ppm belt into two machines taking 30ppm each then your main line will back up even if there is no backup on the feed lines
Slight correction on the Manifold at 1:20, the last two belts in the manifold will always receive the same throughput as each other. So in this example, the furthest two belts are both receiving 1/16 of the total throughput.
Once you get to the last splitter, all that remains is simply being split in half - so they are getting equal amounts.
If you want to think about it mathematically, the fractions at 1:20 don't add up to 1.
1/2 + 1/4 + 1/8 + 1/16 + 1/32 = 0.96875
1/2 + 1/4 + 1/8 + 1/16 + 1/16 = 4/8 + 2/8 + 1/8 + 1/8 = 1
Of course, this all matters very little on any sufficiently large manifold :)
Yeah, I think my intention was to say "these are the fractions and also it continues as such if you make the manifold larger" but based on the comments, that clearly did not get communicated very well.