Factorio Forums

First of all sorry, that I need to spam, but I can upload only up to 3 pics per post and I won't stop this now.

Kovarex wrote:

The point is, that the flow on the transport belt is limited. One fast belt can barely cover one bigger mine.

So my thought was this: How far can we go with belts? What are the limits? When are belts more efficient than trains?

I describe some techniques to come to the limits of the belts. First some simple knowledge:

To get over the long distances, belts are needed. There are basic-, fast- and express-transport-belts.
The basic belt costs a wheel and a iron plate, so in sum 3 iron plates (one wheel needs two plates). This is really not so expensive compared to the tracks, which costs more. We must take also in account, that belts are easier to lay than tracks.
The fast belt is double as fast as the basic and is about 4 times expensive (11.5 vs. 3 plates), while the express-belt is "only" 3 times faster, but costs in sum 41.5 plates, which hurts and here rails begin to be more useful.

At full capcity a basic belt can deliver 3.6 items per second (both sides must be used). That's 216 tiles per minute. This can transport the output of about 20 electrical miners or feed about 12 basic furnaces. A fast belt is simply double as fast, so it transports about the double amount, but not when it has edges. See (1).

The task is: Transport as many stuff per second from a mine-field to your "central smelting area" with the lowest costs and the lowest afford that's possible.
- Basically iron-ore and copper is target of this. Coal is normally not needed in such amounts.
- The task is not "transport as fast as possible", the task is "transport as much as possible" so it's doesn't matter so much, how long it takes from the start to the end. At the end a continuous item-stream is needed.


Tipps
---------

1. Use faster belts at the edges. For every edge of a belt, the inner side is slower. So with some edges in your belt, the belt has not the full capacity! (about 2 items instead of 3.6) With a faster belt at those edges (for basic belt the fast and for fast belt the express belt) the inner edge of a belt keeps the speed and the afford to built this is not too high. 2. Use belt compression at the beginning.

When you use a faster belt before a slower, the items are "pressured to it's minimum length on the belt". This is important in conjunction with splitters and joining stuff. It makes it also easier to fill in some more stuff, because the gaps between the compressed stuff are then bigger. It makes also easier to see, if your belt is really at full capcity. 3. Use splitter as merger instead of feeding from side.
When you feed a belt from the side, the throughput is very limited (feeding from side is only a good use, if you want to make a queue for the mines). The splitter is not only useable for splitting belts, it's the tool to join two item-streams. There are some more tricks:
3.1 Don't just make two belts into one, put the remaining side also into the target belt.
3.2 Use faster splitter if you want to feed a faster belt, because otherwise your splitter is the bottleneck.

4. Use splitter to equal the left and right side.
Sometimes the left and the right side of a belt moves differently due to some antisymmetry (e. g. the miners are only on one side of the belt). The one side stucks, and the other is nearly empty. Use a splitter to splitt into two belts rejoin the one side into the empty side of the belt. Also this trick can be useful: Use underground belts to split both sides of a belt and rejoin them into both sides of a belt.

5. Use 2 lines of belts. As long as you are outside of your factory you normally have enough place to just use 2 basic belt lines. Doubles the throughput. This is much cheaper, than to upgrade the belts.

Simple calculation here: a double row of fast belts is 4 times faster than a basic belt. That's fast! We use 23 iron plates per tile-distance for that. This hurts a bit. But it is about the half of an express-belt!

6. Use faster belts.
Replace basic belt with fast belt: Don't forget the edges: You need then of course express-belts! The experience showed, that it is more or less fast enough to handle the edges.

Replacing fast-belt with express-belt: It depends. Remember: Express is not 4 times faster than basic belt, it's only 3 times faster! So while you can join 2 basic belts into one fast, you cannot join 2 fast into one express!
And the afford to produce such amount of belts is high! I use express-belts normally only at the "central-region", where you don't have enough space left.
And if your belt has edges it doesn't make much sense, because there is no "hyper-express-belt", which you can place in the edges. But If you want to make that afford: replace the edges with two lines of belts. Use 2 express-splitters (one for splitting before the edge, the second for joining after the edge). As said: The afford is high and when you are here the usage of the train begins to make sense.

There is also another usage of this: You don't need to replace all belts, sometimes only parts of a belt need faster belts.

7. Precompress.
Assuming, that you have a combination of coal and iron-ore in close quarters. Then the pre-production of steel makes much sense, because 1 steel needs 5 iron, so the iron-ore is compressed by factor 5! Nothing can't be faster and you only need furnaces and inserters. This can also be done with other items, but no wheels, basic belts, copper wire, because they don't compress or even inflate (one copper plate is two copper wires!)! The afford is of course higher and you have not such a control about it as when you produce the stuff in your central factory, because the ways are long and you very soon have the problem to produce too much or too less and the reaction is slow. But good candidates are simple circuits, basic inserters, amunition and such stuff, things, which you also need in extreme masses.

8. Load a storage. For experts.
Sometimes you don't need so much resources. This means, that the mines stand still, the items on your belts don't move. You have this expensive infrastructure, but it is not used! This is very bad, because in the moment, you really need it, then you don't have enough to fulfill the need and you need to expand the bottleneck first. It's even bader: Mines, which stand still need about 3 watts of power. With 100 mines this is 300 watts standby. Thats a half steam engine.
So, the idea behind this is, that you need only one belt per bigger resource-field, even if the field is big enough for 50 miners. The belts are a bottleneck - 200 items per minute is slow, but what counts is not only the throughput, but the total delivered stuff, because the game time is long! And in 1 hour even a basic belt can deliver more than 12000 items! So with the time every mine-field is transported - even with basic belts.
Assume we find 3 big resource fields and for every field we lay a belt. So if you have 3 belt lines, that's about 600 items per minute or 150 mines and that's currently more than enough.

But 150 mines needs much power! You want to clean a resource-field as fast as possible to use the miners somewhere else (while the storage delivers you meanwhile). So instead of removing a bottleneck, you elongate the usage-time of it, because the items gets transported in some time without much cost and you don't need to use faster belts.
This is a useful technique, because this has several useful side-effects: The needed amount of energy is reduced (miners need 3 watts in standby), you don't need so much miners and you have a reserve, which is important to get time for intensive building, planning, exploration and other stuff.

Two places are useful for building a storage:
- directly at the mine-field or
- before your central smelting area.
Both have their usages.

By building it directly at the mine-field you can harvest really big fields in relative short time. You can store each "row" of miners in one or more chests (a fast inserter can handle about 4 mines, look into the info of the mines to estimate the needed chest capacity!). The output of the chests needs not to be fast, but constant. 4-6 basic inserters can fill one side of a basic belt (the pic lies, the left 2 chests are all empty). They also need not so much energy. Use the above techniques to keep the stream as constant as possible. It needs about 1/2 hour to fill one chest and 1 hour to empty it.

Pro-tip: Later in the game you can replace the chests with provider-chests and let your enhanced logistic robots move that stuff to your central.


By building it directly the storage in your central region (you can built it before or after the furnaces, not much difference) you can guarantee a constant stream of raw-material. If you use storage-chests for that, you see also, how much is currently left. To fill that chests use a cascade of up to 3 splitters. This cascade has the use to let only 1/8 of the item stream through to the storage, while you need the stuff in the factory, but let the full stream to the storage, when you don't need it. This has the purpose that you use only as less inserters as possible (which reduces the needed power).
The items are then pulled into a cascade of chests. For a basic belt about 8 basic inserters are needed. You can use of course much more, more than 10 iron chests seldom makes sense.
After the storage the items are rejoined. The orientation of the splitter is here important, because the splitters prefers always the left side to join. So the picture above is the wrong side! The left side of the splitter should not be the storage-side!


9. Use car to built
You can built inside the car - don't ask me, how this works, but it does. This speeds the built process obviously up. But you need to plan to hit the target!



I think with this techniques the usage of belts becomes a high range: up to 500 fields! I think this is the area, where belts are easier to handle than trains.



PS: This can be used in the wiki!

Optimal usage of belts is a topic of great interest to me. I've spent a lot of time measuring throughput and designing "hacks" similar to what you've described here. I've also encountered some issues which I hope the developers will address.

Your numbers are way off. A basic belt can transport 12 items per second. You'd need about 40 stone furnaces to fully load one.
(I think you've mixed it up with the speed at which items move on the belt, which is 1.8 tiles per second. But there are multiple items in each tile and as you said yourself, speed isn't important, capacity is.)

As noted, fast belts for turns and mergers are required in order to maintain this capacity.

A fast belt is advertised as being twice the speed of a basic belt, but in fact according to my measurements its allowed density is lower than that of a basic belt. So instead of 24 items per second you get about 19 items per second. If you have turns (with express belts) this goes down to about 17.

I still need to do some measurements to figure out the efficiency of splitters in turning/merging. But in any case, upgrading from basic to fast only increases throughput by 50%. Not really worth it until express belts (needed for reinforcements) become pocket change.

Great write-up ssilk!
Many good points and ideas, with good pictures to go with them (though the larger pics are difficult to view in the embedded scrolling view, like your Clipboard06/in-line storage pic)
Having some hard numbers for capacity/speed definitely help, and with Holy's correction, further justify my reliance on Basic Belts

n9103 wrote:though the larger pics are difficult to view in the embedded scrolling view, like your Clipboard06/in-line storage pic

Note that the images expand if you click on them, at least if you have javascript enabled. Usually less effort than scrolling, although I'd prefer to do away with the scroll bars in the first place.

Nice essay!

I would like to see a performance-comparison trains vs belts as well, because I am to lazy to test it by myself

@Holy-Fire:
Yes, I mixed tiles and items up. Thanks! It was 2 o'clock in the morning when I thought there must be a mistake, but didn't found it. My measurements said other things than the theory. And then I thought, someone else will find it.

And when I see your numbers: I made some calculations. 12 items per second means, that means a basic belt can handle 48 electrical miners (A miner depletes every 4 secs one item, so 12 itemspersec / 0.25 itemspersec = 48 mines). Can't be true!
I think 24 is right. You see, too much things I don't really know. But I will make some more measurements (I think to a test-drive) for that and will print the results here.


@n9103: When I'm sure I have the right numbers I will edit the above and also replace the pics with smaller versions.


@wrtlprnft:
This is not that easy to say, because logistic of trains is very different to a logistic of belts. I think they can't be compared directly.

Belts are a constant stream of items. Mathematical they are very easy to handle, because it works like a fluid in a pipe (volume or amount per second is not much difference).
Trains are very different. You need to fill them as fast as possible, because when a train stands, it doesn't transport anything. When it transports, it is very, very fast and then you need to unload it - also as fast as possible and then the train must travels back - most times empty or not completely emptied. A belt doesn't need to travel back (ok, the underside travels back). And you have the problem, that at the start or end of a train there are always belts included, which are then may become a bottleneck again.

Then think to the needed resources. For efficient transport you need two tracks in the two directions. You need stones, not only iron. You need to built some more different parts: signals, curves, stations. You need an armada of smart-inserters (which needs not so less power!), chests and all of that must be placed more or less exactly, so that the train stops at the right point to fill/empty it.

Then you also have this storage-problem. With trains you need a big storage at the start-station near your mine and a second storage near the central-station. And that must be really, really big, when you begin to handle many trains, because it is very stupid, when the wagons cannot be fully unloaded and the trains travels back not completely emptied. And normally you cannot built the central station near you central, because there is no place. So you need also longer belts to connect the station with your funrnaces or you move the furnace-area to the station... which doesn't solve the bottlenecks and you think: Why should I build that ...?

You need built the locomotives and wagons, put locs and wagons together, program the route. You need to make sure, the locs have coal and are refilled. And not to forget: You need to protect the train stations against the creepers... that's not needed with belts!

So you have a lot of steps, which must be done and which costs time, which you then don't have for other stuff. And sometimes you need to control, if everything works fine.

So, I can say only for me, that the range of about 300 tiles (or even more?), where belts are better than trains, is my personal feeling, because it can't be calculated, because there are to much different ways to built it. But that area, which this range covers is then so big, that it would be a wonder, if you won't found enough resources within, so currently (!) this is not a question for me.
Beware: we are talking here only about transporting raw-material to the factory, the calculations will look also very different, if the train is used to build a track to the ... starting-area for the rockets - just for example.

And for transport of raw-material the calculations will also look very different, if there are some suggestions from the forum implemented.

Just because I'm in this mode:
- automatic building of tracks (you plan the track for example on the map and the track is build for example by the logistic robots)
- train-stations, which have big storages and fill the wagons very fast. It doesn't matter so much, how expensive they are.
- automatic train production (ready to go trains, eventually with a control, which tells the train, which is his schedule)
- train-stations can have weapons.

And the situation will look completly differently, if we have an industry, which has more raw-materials and items, which have a high-compression factor and the resources are at very different locations. Then it is much easier to make it like in OpenTTD, drive to oil rig, get oil, drive to refinery, get different sorts of products, drive to ... that is something, which is then at some point more efficient with trains, because you can use a belt only for one kind of item efficiently.

ssilk wrote: And when I see your numbers: I made some calculations. 12 items per second means, that means a basic belt can handle 48 electrical miners (A miner depletes every 4 secs one item, so 12 itemspersec / 0.25 itemspersec = 48 mines). Can't be true!
I think 24 is right. You see, too much things I don't really know. But I will make some more measurements (I think to a test-drive) for that and will print the results here.

An electric drill takes about 2.2 seconds to mine an item, not 4 seconds. So 12 items per second require about 26 drills.


And I must correct you on another matter - the recipe for a basic transport belt calls for an iron plate and gear per 2 belts. This means that the cost of a single belt is 1.5 iron. So a fast belt is almost 8 times as expensive as a basic belt.

In 0.5.x I used a second faster belt just after corner. In some cases it wasn't enough to just have faster belt in the corner. Haven't tested with 0.6.x.

Also you can use feeding from side to get full capacity just fine if you use faster belt for both feeder and target belt.

Phantasm wrote:In 0.5.x I used a second faster belt just after corner. In some cases it wasn't enough to just have faster belt in the corner. Haven't tested with 0.6.x.

Also you can use feeding from side to get full capacity just fine if you use faster belt for both feeder and target belt.

To the first: Currently I'm build a test drive, which enables me to test such things, but it is much more complicated than thought to build it so, that the results are reproducible.

To the second: with feeding from side is meant, that you feed then only one of the both sides of a belt. With a joiner this is avoided.

ssilk wrote:@Holy-Fire:
to the ... starting-area for the rockets - just for example.

How do I build the starting area for the rockets?

DRBLN wrote:

ssilk wrote:@Holy-Fire:
to the ... starting-area for the rockets - just for example.

How do I build the starting area for the rockets?

it's an example for a reason, where trains are more useful than belts!
See this thread to understand, why I choose it https://forums.factorio.com/forum/vie ... f=6&t=1114

I wasn't simple! Hard work until I build a reliable test-track.

First results for the max throughput of belt-constellations follows. All tests have been done between 3 and 5 times. A systematic measurement-error cannot be excluded yet! All numbers in tiles per minute if not other said!


Simple straight belt

Basic: 716-719 (that's 11.93-11.98 tiles per second, which is nearly the theoretical maximum)
Fast: 1182-1184 (this is NOT double as fast as basic! it's only 1,65 times faster)
Express: 1774-1775 (this about 2.5 times faster than basic)

Belt with 4 bends (2 left, 2 right)

Basic: 477-486
Fast: 786-792
Express: 1038-1042

Belt with 4 bends (2 left, 2 right) and faster belt at the edges:

Basic with fast edge-belts: 674-683
Fast with express edge-belts: 1058-1061 (The fast belt with express edge seems to be faster, than pure express belt!)


Some more stuff:
- Splitters have sometimes problems! They are some times slower than the appropriate belt.
- It is really difficult to measure the express-belts, because the throughput is so high.
- side-inserts, even for some items only slow the speed definitely down.


TODO

- rework my testdrive, because it is hard to make so much measurements and wait all the time
- more systematic protocol of the measurements
- documentation of the test-track, so that everybody can repeat the measurements.
- Rework the above documentation (pics and new text)

PS: Does anybody have a good idea, how I can bring the train to run when a test is ready and a new train drives in? Like the smart ìnserter? I ask, because I haven't build much with tracks yet, but I found the wagons extremley practicable for counting the transported items, because I can fill it with 6 inserters at the same time, which is not possible for chests - and as you can imagine, it is hard to look into 16 chests and add the inserted number of items, and then you need to empty the chests before you can make a new test.

Hm. I would need smart chests, which can be used like a deterninistic finite automata. Would be really cool, but with red and green wire only not possible yet.

You can only upload 3 attachments per post, but if you use an external website like imageshack u can do whatever u want.

I think that is like using the car driving before you to brake. I don't understand this limitation of the forum.

It helps limit the damage a spambot can do.
It also help police normal users from spamming threads with pictures, and from using up too much server space too quickly.
Splitting your posts to make more pictures shouldn't be a problem if they're legit, though I do agree that 3 seems a bit low.
5-7 would probably suffice for most cases here. Then again, not many have run into the 3 pic limit.

ficolas wrote:You can only upload 3 attachments per post, but if you use an external website like imageshack u can do whatever u want.

FOR THE LOVE OF GOD PLEASE DON'T USE IMAGESHACK!

Sorry for caps, and now why:

From what I understand, if you get too many image views, the pics will stop showing up because you went over your limit. I could be wrong, but I KNOW that imgur doesn't do any of that, so, for future reference, please use imgur.

On topic: Looks very cool, but I would never use trains anyway.

Off topic: Another reason not to use other image-hosters is, that they can just delete the picture. Most of them have also ugly ads or draw your attention off to sex sites or - even more bad - other games.

On topic: With the modules and bigger factories and usage of logistic bots to feed all that stuff (see this https://forums.factorio.com/forum/vie ... 1205#p8434 ) the need for resources gets so big, that I don't know how to build that all without trains, because when playing with nearly normal settings the resources are so far away, that it takes some minutes to flow on the belt to the center, and the throughput is just too low, even with fast belts. Just to compare: A fast belt can transport between 1000 and 1200 items/minute. A long train can do that in 10 seconds. And you can use many of them.

ssilk wrote:On topic: With the modules and bigger factories and usage of logistic bots to feed all that stuff (see this https://forums.factorio.com/forum/vie ... 1205#p8434 ) the need for resources gets so big, that I don't know how to build that all without trains, because when playing with nearly normal settings the resources are so far away, that it takes some minutes to flow on the belt to the center, and the throughput is just too low, even with fast belts. Just to compare: A fast belt can transport between 1000 and 1200 items/minute. A long train can do that in 10 seconds. And you can use many of them.

On the contrary, with productivity modules, iron and copper are completely not an issue since you get so much out of each ore unit. Power is an issue, specifically the time it takes to physically place the panels and accumulators.

Of course, if you use coal instead of solar you'll need to bring in lots of coal, but I think this is inferior to solar long-term.

I haven't played around with trains much; I'll take your word for their capacity, but I assume the numbers apply to a long straight line taking resources from point A to point B. Setting up your tracks and routes to bring in resources from all over the map must take lots of time, which would have been better spent placing solar panels. For the resources you do need, belts are much easier to set up and extend.

PS I'm focusing on productivity modules for a reason, I calculated the benefits of each type and productivity wins by far.