Factorio Forums

Saunis wrote: ↑

Sun Sep 23, 2018 4:30 pm

What would happen if I drove many wagon unload cycles through those things?
Would unbalanced wagon unload start to slowdown whole cycle because 1-3 chests were full too soon?

All designs slow unloading to the compressed belt output speed if the supply is constant - it's the output belts, not the buffer chests, that are the ultimate limiters. Unbalanced buffer chest designs just get closer to that limit faster.

What's important to note is not so much whether or not the buffer chests are unloaded in a balanced way, but is there enough time for a train to be replaced with another (cycle time) without starving the compressed output belts.

This cycle time is where unbalanced buffer chest unloading shows its brittleness. If one chest empties before the next train arrives, it can starve the output lines even though the others still have lots of items built up. Essentially, unbalanced buffer chests can give you less time to cycle trains.

If you look at my (unbalanced) design in a production environment, there is barely enough time to cycle trains with nuclear fuel. But there is enough time. So, given a constant supply of trains the unbalanced chests aren't relevant.

If you can't get trains in fast enough for any design then it doesn't much matter what design you use to unload as all of their output belts will be starved by insufficient supply. What they look like when they're starved, in my opinion, is irrelevant. If they're starved in some manner (whatever manner), the problem is that you need more supply not that the unloader isn't working well.

If there's a more efficient 4-lane per wagon unloader than the one I posted above, I haven't seen it (and would love to).. but there's no denying that it's tough to meet its cycle time requirements. I do wonder if some of its brittleness can be corrected with lane preferences set in the splitters, but I haven't played in a while to test it. And some of the other designs with a similar footprint that are less brittle (more balanced) aren't really that much less efficient that they should be set aside.

To give you an idea of how I think stations should function to meet such tight cycle times, here's my solution to zOldBulldog's orphan train problem. I don't think I described it very well, but if you drop the blueprint into a creative mode game and have a look it's not that complicated.

Durentis wrote: ↑

Sat Oct 06, 2018 4:52 am

All designs slow unloading to the compressed belt output speed if the supply is constant - it's the output belts, not the buffer chests, that are the ultimate limiters.

Not all designs, there's several in this topic that unload the buffer chests faster than it's possible to load them from a train.

quyxkh wrote: ↑

Sat Oct 06, 2018 10:52 am

Durentis wrote: ↑

Sat Oct 06, 2018 4:52 am

All designs slow unloading to the compressed belt output speed if the supply is constant - it's the output belts, not the buffer chests, that are the ultimate limiters.

Not all designs, there's several in this topic that unload the buffer chests faster than it's possible to load them from a train.

No. Unloading buffer chests faster than they can be filled (onto an excess of belts) implies unsustained uncompressed output. If you want sustained compressed output belts, which I think is a focus in this thread, the train unloading speed is ultimately constrained by (slowed to) the maximum output speed of the belts. The compressed belts slow the emptying of the buffer chests, which in turn slow the unloading of the train into those chests.

DaveMcW hit four per wagon target in the second reply to this topic, seems to me everything since has been about "how fast can you unload?". The current record is nearly 8 belts per wagon and the limiting factor on that is not the output speed of the belts.

@quyxkh:
I assumed that max unload of a wagon would be (27.7 i/s *12)/40 -> 8.31 blue belts per second?

disentius wrote: ↑

Mon Oct 08, 2018 4:07 pm

@quyxkh:
I assumed that max unload of a wagon would be (27.7 i/s *12)/40 -> 8.31 blue belts per second?

Yah, but since chest⇒belts can be made faster than wagon⇒chest, the limiting factor is how fast you can get the next train into the station. Best I've gotten is 117 idle ticks with circuit-controlled 1-1 nuc trains, say for green chips you get 40 slots @ 200 chips to 12 inserters @ 60/26*12/inserter, (40×200)/12/(60/26×12) is 24.07sec to unload that, (117/60)/(24.07+117/60) is 7.5% idle time, 8.31/1.075 is ~7.73, so I think the blue belts/wagon limit for sustained delivery over time is 7.73 belts for chips, ~7.29 for plates, ~6.67 for ore.

For space science you could deliver enough to a stop in one wagon that you'd need more than eight belts to stay ahead of the game, you'd only need like 20KSPM...

edit: thinking about it, the train can leave as soon as the inserter picks up the last handful, so unloading gets one free swing and the actual inserter-idle interval is only 91 ticks + however long it takes the train to actually leave. My benchmarking thingy says the max-rate unload design sustained 18799.9/min unloads for five minutes, a close match for the 94-tick delay math.

Damn. totally forgot the unload to splitter thingy. thanks

disentius wrote: ↑

Tue Oct 09, 2018 5:52 pm

Damn. totally forgot the unload to splitter thingy. thanks

Oh, God, somebody got a stack inserter plus a longhand feeding a car, and since you can put cars on belts . . .

Wholly cow!
And nobody came up with this here?
Much more to know about about this game. Much to learn. Awesome

So to get back to the topic of this post:
This: Satisfies all requirements of OP.
Sadly, no bonus points

quyxkh wrote: ↑

Tue Oct 09, 2018 6:54 pm

disentius wrote: ↑

Tue Oct 09, 2018 5:52 pm

Damn. totally forgot the unload to splitter thingy. thanks

Oh, God, somebody got a stack inserter plus a longhand feeding a car, and since you can put cars on belts . . .

So that raises the theoretical maximum to 373.8 i/s (9.345 blue belts...)

That's really compact and energy efficient.

It doesn't allow enough time to cycle trains even if they're filled with 200-item stacks though, which means that the compression isn't sustainable. Because you only have enough surface area on the tank to unload the wagon with three inserters, the the wagon takes twice as long to unload and leaves the tank pretty much empty when the train finally pulls away.

Not a big fan on abusing cars/tanks with inserters in the first place and I don't think we should be able to insert into them. Very neat what can be done with them though.

disentius wrote: ↑

Thu Oct 11, 2018 9:21 pm

So to get back to the topic of this post:
This:
2 belt per side unload.gif

Satisfies all requirements of OP.
Sadly, no bonus points :)

No it doesn't. It breaks tilability because it uses 7 tiles per wagon. Might be fixable with a splitter. Also you can side-load cars with up to 4 inserters. It also breaks the "must tunnel one unrelated belt to the other side" rule. It further doesn't satisfy the two bonus rules, but hey...they're bonus. So you get 1 of 3 points. Which is definetly not a passing grade :p.

Durentis wrote: ↑

Thu Oct 11, 2018 10:08 pm

It doesn't allow enough time to cycle trains even if they're filled with 200-item stacks though,[...]

I'm pretty sure the thread concluded a few pages ago that fast enough cycling is generally impossible. At least not for anything but pointless 1-1 trains (pointless because if 1-1 trains were the target tilability wouldn't be required in the first place). The target is still trains with at least 8 wagons.

OOps! You are right, I got rather carried away after seeing that video.
Here is a better one.
Edit 3:
4 inserters sideloading:: Tank, yes, Car, no.
And there is no way you can get enough trains in but: EXPERIMENT!
(unloading train to car: 83 i/s) EDIT:
forgot to mention the need to adjust stacksize for gapless output:
Left and right symmetrical. From wagon up: 12, 7, 7, 12, 7

Experiment, yea. But at only 3 unloaders you're only 3i/s above the output, a bit too close for my taste.

disentius wrote: ↑

Fri Oct 12, 2018 9:10 am

4 inserters sideloading:: Tank, yes, Car, no.

Well, then why are you not using tanks!

EDIT:
Because I am an idiot who doesnt recognize a good solution staring him in the face:)
The top one is optimized for everything. very bloody beautiful!
Now we can finally close this topic,and I can date again instead of factoriopitmizing.

Nice tank solution! Now if we could only include them in the blueprints... it would be perfect. (At least for vanilla. I saw Eradicator's review on Miniloaders and they were the best option I have seen to date. That is what I now use in non-vanilla.)

zOldBulldog wrote: ↑

Fri Oct 12, 2018 11:39 am

Nice tank solution! Now if we could only include them in the blueprints... it would be perfect. (At least for vanilla. I saw Eradicator's review on Miniloaders and they were the best option I have seen to date. That is what I now use in non-vanilla.)

You saw my what? I don't remember doing any reviews. If i wanted a modded solution for the problem i could just write one myself. But the whole problem is trivial if you allow infinite-speed inserters (ia.k.a. "loaders").

eradicator wrote: ↑

Fri Oct 12, 2018 12:56 pm

zOldBulldog wrote: ↑

Fri Oct 12, 2018 11:39 am

Nice tank solution! Now if we could only include them in the blueprints... it would be perfect. (At least for vanilla. I saw Eradicator's review on Miniloaders and they were the best option I have seen to date. That is what I now use in non-vanilla.)

You saw my what? I don't remember doing any reviews. If i wanted a modded solution for the problem i could just write one myself. But the whole problem is trivial if you allow infinite-speed inserters (ia.k.a. "loaders").

LOL, similar names. Turns out I always confused you and Xterminator, so I was thinking it was your review.

This is the review I was talking about: https://www.youtube.com/watch?v=6aGD3fh2P5A It is really slick and UPS friendly, but it *is* a mod even if QOL. So it is a tool I now like for modded gameplay.

Of course it is useless for this challenge since you are going for vanilla. Which is why I liked the tank-based design. Really slick. I only mentioned miniloaders because the tank design is the closest to be as compact and effective as the mod. Doing that in vanilla is quite an achievement.

eradicator wrote: ↑

Fri Oct 12, 2018 7:39 am

disentius wrote: ↑

Thu Oct 11, 2018 9:21 pm

So to get back to the topic of this post:
This:
2 belt per side unload.gif

Satisfies all requirements of OP.
Sadly, no bonus points

No it doesn't. It breaks tilability because it uses 7 tiles per wagon. Might be fixable with a splitter. Also you can side-load cars with up to 4 inserters. It also breaks the "must tunnel one unrelated belt to the other side" rule. It further doesn't satisfy the two bonus rules, but hey...they're bonus. So you get 1 of 3 points. Which is definetly not a passing grade :p.

7 tiles per wagon is tileable - you have to count the space between wagons, so as long as there isn't an adverse interaction (which there doesn't appear to be in this case) you can use 7 spaces.