1 to 18 even distribution

[Albravo]

I have spent a couple days trying to figure this out. How do you evenly distribute 1 fully loaded belt to 18 separate belts? Obviously for train loading.

[DaveMcW]

Split it 32 ways, then merge 14 of them back to the source.

[quinor]

You simply can't with joining/splitting belts only. I could give you even strict mathematical proof for that fact. You can split evenly only into the power of two.

Though if you don't need it to be precisely exact, split to 16 and then split 2 belts one more time what gives you 4 uneven belts or split to 64 lanes (it's easy though some space needed) and then join 10 fours and 8 threes of belts - it gives little more even distribution.

Another way (but only if amount of items is high) is using 18 inserters feeding 18 belts - if belt they feed from is full, they'll split items evenly.

[update] Dave's solution is better than mine, I just forgot about feedback.

[onebit]

To expand on DaveMcW's post, something ugly like this. The return routes could be joined by a splitter so it doesn't all go on one side of the belt.

[ssilk]

See aiso https://forums.factorio.com/wiki/inde ... _splitters

[GlassDeviant]

I would simply cascade down to 16 belts, with 5 loading each of the three rail cars and the 16th one feeding back into the beginning, to get equal balance. 6 inserters per rail car are not really necessary, I often use only two onloading and two offloading for oil barrels (empty/full). If you are finding the trains are not loading fast enough, either make them stay at the station a bit longer or place buffers on the feeder lines, either like the one on the top or like the ones on the bottom, on each individual inserter, or both if you feel the need. It's big, it's ugly, but it works flawlessly, at least so far, and doesn't need to be spread across both sides of the rail. Now if only I could figure out a better way to offload trains...



Note: If you're not mining fast enough, this won't load trains quickly enough, hence the double buffering. The chests load up in between trains so that when one arrives it gets loaded non-stop instead of having to wait for ore to arrive from the drills.

P.S.: Yes, I noticed I didn't line up the inserters very well for the rail cars, I'll fix it later...maybe.

P.P.S.: OK my OCD kicked in and I fixed it, so it's a bit more elegant now and I even threw in some express corners.

[ssilk]

Sorry for off-topic: as much I like to see those constructions, I need to ask, if it wouldn't be easier to going plain through to logistic bots research, and let the bots do that equal distribution of items.

Because you need to do this construction several times and every situation looks different, you cannot use blueprints most of the cases. With bots this afford is easier. And the point why I ask is, that newbies reading this (and many read it) might think, this is the only way to do it.

I won't discuss it, it is just a short "do you know, that...."

[GlassDeviant]

I would actually consider it on-topic, and bots is a better way to go, but not everyone wants to use them or wants to spend the materials required for the Roboports and bots. I just finished a game where I never used a single logistic bot, only consctruction bots and usually only installed temporarily for expanding my base quickly, until very late when I decided to blanket the entire base with Roboports (easier said than done, my base is huge, so big I had to stitch together screen shots of parts of the map to make one big map, and I still have barely one third covered with Roboports, if even that much).

[GlassDeviant]

So I did a "real world" test of the above distribution system and it works near perfectly, with a variance of only +/- 1 ore unit across the 15 buffer chests.

[GlassDeviant]

Expanded my materiel equalizer to 4 cargo wagons in length with 5 inserters per wagon.

Drained a small iron ore field into it, and each and every buffer chest contained 1180 or 1181 ore in it, which variance can be accounted for simply by the fact that the ore total in the field was not evenly divisible by 20.

On just one side of a train, 25 seconds was sufficient to load every car entirely as long as there was enough ore already in the chests. 20 seconds was not enough, thus if both sides were done and the chests could be kept adequately supplied, a 15 second stop would be sufficient to load the train but 10 would be too short. This is still superior to my usual trains that have to be stopped for 30 seconds at each loading site and still don't always get fully loaded each time.

http://home.cogeco.ca/~glassdeviant/exp ... alizer.jpg

For unloading I went with 6 chests per wagon and streamlined the unloading process onto balanced belts. Each of the merge points in the last two stages of merges have a two-tile lead-in per side because of some cluster lockups I was getting when they were shorter. Earlier stage merges don't have this problem because they aren't as heavily loaded. I have avoided using splitters to merge as even the express splitters cause slowdowns when merging instead of splitting. This can unload the same train of wagons in under 20 seconds, which will hopefully be enough to supply my block of 40 (20 per side) electric furnaces continuously.

http://home.cogeco.ca/~glassdeviant/exp ... loader.jpg

If someone has a better unloading system I'd love to see it, as I am not completely sure this will do what I need (hopefully it will, just have to implement it and see. With my artificial throughput test that eliminates backlog as a factor by using a long double sided buffer, the compression was never completely maxxed, though you can see binding at the merge points of the unloading system.

http://home.cogeco.ca/~glassdeviant/exp ... buffer.jpg

[Jackalope_Gaming]

A 1 to 18 splitter can be done with a 1 to 3 splitter and then 3 1 to 6 splitters. Here's the setup I use for my 3 wagon train loading stations at the moment.

[Jackalope_Gaming]

New post! I tinkered with the 6 splitter design and came up with this much more compact version for an 18: