Page 1 of 1

Simplest full-speed blue belt ⇒ 4 chest loader?

Posted: Sat Oct 14, 2017 1:12 am
by quyxkh
I don't think it can be simpler than this, loading four chests from a blue belt is possible with four maxed-out stack inserters but it's not so easy to find.
snap=369x348-4.4765625+1.203125@z2T294861.png.jpg
snap=369x348-4.4765625+1.203125@z2T294861.png.jpg (9.51 KiB) Viewed 6590 times
snap=442x452-4.7265625+0.4140625@z2T208886.png.jpg
snap=442x452-4.7265625+0.4140625@z2T208886.png.jpg (13 KiB) Viewed 6590 times
I saw the one on reddit that's really compact and uses ug belts and circuitry and realized you can get the same 100% throughput without wires. I haven't seen this in use anywhere but I haven't seen much, if I've just discovered what everybody knows, then hey, at least I'm on the right track!

Pointing the splitter outputs at each other gets the inserters preferring different lanes, so they don't fight and load at full speed. This did a chest full of space science without a hitch (I measured), and I couldn't drive it to misbehavior no matter what I did to the input belt.
blueprint

Re: Simplest full-speed blue belt ⇒ 4 chest loader?

Posted: Sat Oct 14, 2017 3:13 am
by Jap2.0
Would the inverse of this work? It wouldn't have the curves.

Re: Simplest full-speed blue belt ⇒ 4 chest loader?

Posted: Sat Oct 14, 2017 5:30 am
by Lav
Sweet. 10/10, will use in my next train station.

Re: Simplest full-speed blue belt ⇒ 4 chest loader?

Posted: Sat Oct 14, 2017 1:31 pm
by quyxkh
Jap2.0 wrote:Would the inverse of this work? It wouldn't have the curves.
Four stack inserters have about 15% more capacity than they need for blue belts, if you want max-rate load/unload you can get a little more aggressive with balancing. This setup is close enough to balanced for the 85%ish load (full capacity for a stack inserter is 26 ticks/swing + 3 ticks/item for lol 11.612903225806451 repeating items/sec/inserter at full rate, so the four of them should be able to load at 46.451612903225806... items/sec but because of a slight left-lane preference this will only sustain about 43.8 items/sec without some added balancing.ef...

Hunh. "The two hardest things in programming are cache coherence and naming things properly. And off-by-one errors.". Properly fed, an inserter can unload or load its first item immediately so it only needs to wait for _11_ more, but on average it'lll have to wait for half an item or 1.5 more ticks if it's random (it won't be on a stuffed belt). Max inserter throughput on straight belt should by that math be 12*60/59 items/sec = 12.4137931034482758620689655172… = not 12*60/62. Measurement says unloading from straght belts _loaded_ by four stack inserters (the below setup rerouted to straight-belt unloaders) I can apparently get ~12.25 items per second each, and the setup above fed by any balancer that can keep the left lane stuffed can load 12 of that.
balancer-fed belt⇒chest loading
balancer-fed belt⇒chest loading
pic.jpg (126.95 KiB) Viewed 6549 times

Re: Simplest full-speed blue belt ⇒ 4 chest loader?

Posted: Sat Oct 14, 2017 5:30 pm
by Mehve
quyxkh wrote:Pointing the splitter outputs at each other gets the inserters preferring different lanes, so they don't fight and load at full speed.
I hadn't even considered that little detail, but you're right, they do tend to trip over each other if you just run the four final belt sections straight ahead.

Re: Simplest full-speed blue belt ⇒ 4 chest loader?

Posted: Fri Oct 20, 2017 12:08 pm
by olafthecat
Very nice for research faculties.
In fact, pretty good for anything.
Will use :)