1 or 2-to-6 train loading balancer

[SpeedDaemon]

I needed a 1 (or 2) to 6 balancer for loading trains, which means it couldn't be more than 7 tiles wide. I looked around and wasn't able to find anything like that, so I made one.

Uses an unfortunate amount of "vertical" space, but I can't see any obvious way to make it any smaller.

Blueprint

Code: Select all

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

[AndrewIRL]

Your loader looks quite good, similar to:


But Madzuri came up with a circuit build which is much more compact:
https://www.youtube.com/watch?v=4IjfnOSHZ1A

[Shokubai]

That's a decent layout. I've recently settled on a circuit based loading system. It's simple 1 belt in and the circuit balances the chests by average fill. It's clean and requires very little space.
http://imgur.com/wheyAuL

[Distelzombie]

That's a decent layout. I've recently settled on a circuit based loading system. It's simple 1 belt in and the circuit balances the chests by average fill. It's clean and requires very little space.
http://imgur.com/wheyAuL

How does this work?

[Shokubai]

In short terms. It takes the Average amount of an item in all chests. If a chest is less than average it loads...It has some fail-safe so it cant get stuck at some weird (not full) number. It will eat 1.5 blue belts per 6 inserters and load every chest equally.

[blizgerg]

The simplest circuit based balanced loader/unloader that I know of is to wire up all the chests together and input it to an arithmetic combinator set to divide by negative the number of chests. The way things are wired is just like the picture above. red wire connecting all the chests to the input of combinator, red wire connecting all the inserters to the output of combinator, green wire between each chest and inserter that is accessing it. Then you set the inserters that either load or unload to activate on anything > 0 or < 0 depending on if they are loading or unloading. This by it self will get stuck if contents of all chests is an exact multiple of the number of chests. So you add a constant combinator to the output of the arithmetic combinator that adds one or minuses one to the type of item you are loading/unloading.

Total additional parts over no circuit, 1 arithmetic combinator, 1 constant combinator, and lots of wire (wire free if using a blueprint)

anyone know of a simpler that can't get stuck?

[iceman_1212]

That's a decent layout. I've recently settled on a circuit based loading system. It's simple 1 belt in and the circuit balances the chests by average fill. It's clean and requires very little space.
http://imgur.com/wheyAuL

How does this work?

We can load evenly across any number of chests with with 1 arithmetic combinator. Here's an example of loading evenly into 24 chests. The benefit is that it's faster to load X items into a train if those items are distributed among multiple chests (as opposed to all X items in one chest). I have played around with circuit-balanced unloading but have found that it is unnecessary (for my bases, atleast) as long as I place one standard balancer downstream to ensure that items are pulled from all wagons.



Close-up of wiring:

Blueprint String

0eNrdnN1uozoQx9+F67AHG39W+yZHq4qmbGudhERAqq1WefcDibZNjP+sx1zRm12lSYZhxvOzmY/8zp52p/rYuqbPHn5nbntouuzh399Z516aajf+rX8/1tlD5vp6n22yptqPr9rK7bLzJnPNc/0re2DnzV+/Uv86tnXX5X1bNd3x0Pb5U73rb4Tw849NVje96119VeLy4v2xOe2f6na4yoesbhDiXl77/KLHJjseuuFbh2a8+CAp55vsffiPmfOomCeGf4rZV7tdvqv2x4CM4pv8I+WbDMkpb9Sptv/lrunqth/emZOlB1mb7Nm19fb6AR6QLD4ku/bQ5NvXuuunUn2hg++aq9Ru/AQb/2nr51tLuuEVK0Y7v7R13fhvqfOPc+hO5d+cONGN3+kWEKk+RP6sun7GdmzGdJd77tvD7vGpfq3e3KEdv7R17fbk+sfhvecPST9d2/WPkyX65tr+VO0+V+n1E3n9Vrfv/atrXrKrZQcXjyHCxhf7Y9VW/Xix7Ht2JlheQcsLYHn9odi+fnanfV7vhgu1bpsfD7t6zlpg0RpyDKmQGJuy9lXM2mcFcfEryuIXN1e8uIRDlzAUDYyl3LyMunlOvHlJjPz7u+cYBUyiuy8XwECFFyUTC2RKIFPSCSO/AGE8D0vsYY48nMBmtXbL+YZjM4YrkOE0Ga4iKMekAKaMAkwSuEWMaE4FtyCxi/srm0EHccQuzogqlhQV+QSvFquINnzOF6BQhFHIlyC7BDIFHRLl2iExLKx7D5f4+DCuz7CHEzYmsXrLiXi+csRXrsh85UE5mvoEyEAQJIGaR9HUElHFSahi/kLGjyYlOieUxQKscPBMzRbIBF4qOT3g+JcLODGDKos8XCY/fiJfCHIIF0E5MiXyipjIKxUx8gpS5Fk/8jROCsDI0zQVLUXD0j8KixkN0SZXppDRRrnHLiBEEV6VYgHJLBDJyNCxa2eO8M/oEucXxmUfXDkigdbF6i3H4g+WJcpLiZLK1mB6XIiE0DVRae0UaMclzInMJiXMB/55yxqfXSUCoiAi25A0VPGbikSna2HSIQiSzWIBq0HdQBZkPJgvRweJcx8CcVXSN6TVF1ykn5GROCMjYPGLU7EarJjIlHJhVNFApgA7qhYjJQ1aahG0JH4GlegJRRLBT6qXSD/do0pcOkVYlTqdgaBcIheQGlRLpCXTYfUpf+ljVeHjqkQpZUXfj1ZfZZL+M6yaOW4grCpyN0ewVqKo3RwgR65KGkhIlQHpo07PgARt4CoF81FFIbWgvwOZU6WLBBUHpcmRtvqCg/IPMBpH2rg5hReOSc3iIedaaugG0/C6SMnhxSxondKhEZWW19QGDQomBirce9vgLIBGXNVEkJHqBsrXUM9oiECmF3R7gLKBXtKhBkQmdEGsHTfaz3QYnOnQKDmt6ZxefbVFq/gTkUac1tTevGBlRCd0eETVRUwKruMk09oySPUW4+ekDU4BGMRUQ6Q+SUM90RBndwx6DDELWjxAUcSkYxpJpHc+FF+ODAZnOQxCqiHvRau3m/FzLxY/MhmUdDfUhrzgOdUk1DKjjqmG1uNBOksaP7lh8Y5k0Y5kC9rDNTid2fSmDnA4s+ldckjHkhpiqz/tmUnPazETYygDa0Xi0yVyhCTGbDAtZGl5YVIrqtUTu5mZZmEEdUurCJISTsbP0Fk1oyEM/wTwRSWcbMIZVcQNj6R3coC0EyvS6SWQSHKPw+obQMdV5kXNzOyNhbM3BZnTq08C2onlihnLaWg5auudCotJaOKIm30qFF1y5EgZDbWKNlRlJgsbH7UYY9A/hqQkcfJLEzYEPBlUpDd0oDEtlk5sNE3GyK0Oq69lssl8DJsb7isgJRh5a1r/0BxjBMBeDnvAdtQGPBMWk1Doi+q/Y4zWz0HqQRup56/AwR5/vH75w8wsIsQiS9gVdJw1aLuCXgZcNjdRCEdeGEvv9dAIj+kQB31+jJO7IPQXpMbMSOLlTBD2MCfvVqtvaGSTUckJK2YAjEfTObVVLzxcwm4GJqvW9a/7unfbfLjfJ9dc7ncKiYu42A7+T6GfvulSnFNtX+/dko9zmYdjPbjlokD2z/Clw6k/nshiSbEwumSTcehviHeeXmWwCEYyMSMFBVLnDsNjh4xT09FgPPxm7HAU0+RdfwilZcVViLpfkyIoktrSMf4syBChl5/4ebj5EaFNNgCju3xcCcGlsFwYdj7/D7i1P5U=

[Shokubai]

The one cimbinator method clogged for me randomly.

[Distelzombie]

BLUEPRIIIIIIIIIIINTSSSS AAAH

[iceman_1212]

BLUEPRIIIIIIIIIIINTSSSS AAAH

Added to my last post, my bad

The one cimbinator method clogged for me randomly.

Thanks, will keep an eye out for this behavior with my existing ones. I think changing the setting on inserter to <=1 in 0.15 (from <1) should prevent it from happening.

[torne]

If you divide by -N and then each inserter is active on <1, then it should be impossible for it to get stuck: this would imply that every chest has more than the rounded-down average number of items, which is impossible (they can't all be above average). If you use <0 then it will get stuck when the chests are all holding very close to the same amount (and also won't start at all when the chests are all empty).

[blizgerg]

Ooooo!! So I can remove the constant combinator (to keep it from getting stuck when with a multiple of the number of chests) by changing the <0 or >0 with <1 or >-1 !!!! Excellent!

[Shokubai]

If you divide by -N and then each inserter is active on <1, then it should be impossible for it to get stuck: this would imply that every chest has more than the rounded-down average number of items, which is impossible (they can't all be above average). If you use <0 then it will get stuck when the chests are all holding very close to the same amount (and also won't start at all when the chests are all empty).

This is possibly the reason mine stuck back then. Maybe I'll play with that again. I remember it would get stuck close to 1k in each chest.

Thinking this through
24 chests all with 800 ore going to give me a negative average off -800. 800 in the chest -800 is going to give me 0...would be stuck if <0 but not <1. Yay math.

[SpeedDaemon]

I use the circuit-based balancer sometimes myself, but there are a few gotchas. The inserters actually need to be enabled on "< max-stack-size" to prevent issues.
You can also run into problems if your trains arrive not-completely-empty, or any other external influence causes the chests to become wildly un-balanced.

Since this was going to go into a hybrid belt/bot ore transfer and distribution station that needs to sustain close to 32-express-belt throughput, I didn't want throughput dropping if bots came and added/removed ore to one chest or another.