Water 8 Automation [SPOILERS]

[fleet pivot]

So I have finished the full AE water 8 automation that we talked about a few days back, @zenith ether , @odd ridge , @thorn crescent . I managed to cut even more corners than I thought I would be able to, so it's now 100% AE2 with ||one GT chest buffer||.

Picture below, will write a full solution in comments. Curious to see how it compares with yours!

IF YOU DO NOT WANT WATER 8 SPOILERS DO NOT READ THIS THREAD.

.
.
.
.
.
.
.
.
.
.
.

Labeled picture:

The core of this is a ME pattern that looks like this: https://i.imgur.com/hMkb5lq.png

That is, the entire sequence of 18 quark catalysts which contains every possible pair, making 2000 bary matter.

This pattern is being requested by the fluid export bus, in crafting mode, on the right side, pointing to an input hatch of the HILE. The interface with the pattern is in the middle, pointed to a HV chest buffer, which points to an input bus of the multiblock. So when AE requests the pattern, all the quarks get sent into this buffer, in the correct order, and then sent to the multiblock at a rate of one second per item (the output speed of the buffer).

The buffer also has a machine controller cover, which is toggled by a ME storage bus, which is in crafting mode and emits a signal while a drop of bary matter is being crafted.

So the whole process looks like this:

• AE requests the pattern to be crafted.
• Quarks get sent to the chest buffer.
• Level emitter turns on.
• Buffer sends quarks to the machine, one at a time.

At some point the correct combination of quarks is hit, and the machine spits out 2000 bary matter back into the system. This completes the ME craft. So the level emitter turns off, and no more quarks are sent to the machine, to save on infinity.

Now, at some later point, the waterline cycle finishes, and the multiblock sends the unmatched quarks, as well as two empty quark catalysts back to the system. So the system now has:

• All quarks that were not yet used (seen through a storage bus on the chest buffer)
• All used, but unmatched quarks (returned by the multiblock)
• Two empty catalysts (also returned by the multiblock).

The six interfaces on the HILE all contain a pattern of 1 empty catalyst -> 1 quark catalyst, and go to input buses with the appropriate circuit.

This means that at this moment the AE has exactly enough inputs to request the pattern again and start the next loop! All it needs to do is to use the two unaligned catalysts to craft the two missing quarks. The baryonic matter is already in the HILE, because it was put there by the fluid export bus from the previous cycle.

This also avoids the recipes having a loop. If you're paying close attention, if the quark catalysts were encoded "properly", the full crafting would look like:

empty catalysts + bary matter -> full catalysts
full catalysts + unused catalysts -> bary matter

AE wouldn't like this. But sending the matter from matching the quarks directly to the HILE avoids having to use it in the patterns, so this loop doesn't occur!

[zenith ether]

thats definitely a condensed setup that is for sure. certainly a lot more neater than my setup which consisted of like 20 counters 🙂

[thorn crescent]

@winter cedar is it the same or we had something different? i forgot

[winter cedar]

similar

I only had one interface tho on the hile

and used a fake pattern

[thorn crescent]

i don't remember the gt buffer tho

[winter cedar]

mine was just move a fake pattern to request the craft

with a two stage timer

[fleet pivot]

I love that the whole loop is controlled by "does AE have enough resources to start the craft". Originally I had an activity detector cover starting the craft when the waterline starts, but that wasn't working because redstone connection to AE parts is scuffed (I think).

[winter cedar]

in a closed network

[winter cedar]

it's fine if you know what you can do with it

I have an activity detector to trigger the timer

[fleet pivot]

Yea, using the baryonic matter from a correct match to succeed the craft and stop inserting catalysts was the first idea that inspired the whole desing.

[winter cedar]

which moves the fake item which triggers the quark patterns

which in turn will trigger the crafting of the consumed quark in the next cycle once it's repeated

since AE is actually smart

and you can pattern in the close loop all the quarks to be made with the washed up quark

(i forgot its name)

so that the fake pattern will end making the correct quarks

[fleet pivot]

Yea idk, I had a fluid export bus with a crafting and redstone cards, set to craft once per pulse, and no matter what I did it would only ever start the craft once. When the next cycle came around, even though it got a redstone pulse, it wouldn't do anything.

Breaking and replacing the redstone got it to craft again, which makes me think it's some kind of update issue.

[winter cedar]

we used an lv tank as a buffer

you can check it out btw in our saved up run

it should be findable in #french-general in the pins

[fleet pivot]

Well I need to head out now, can't wait to get back home and see how it breaks >.>

[winter cedar]

if you check out our save it'll probably be broken because the base is fked when you import it 😄

[fleet pivot]

I can kind of see a problem if the waterline ever completely stops, but even that should just fail one cycle and void some small amount of infinity because the quarks go in the wrong order.

The pattern will just be on hold because it never gets the bary matter, so it never completes.

[winter cedar]

the only time our automation will fail is if your server shits the bed for a very long that the timers get unsynced or if your hile/forming press power fails

[fleet pivot]

Yea, if I kicked it off with an ADC it would be resilient to that too. Idk, I had problems with redstone controlled AE buses too in the past. Never something I could reliably reproduce though.

[winter cedar]

don't do redstone controlled busses imo

😄

as you said it's quite shite

[fleet pivot]

tbh I wish they worked correctly ^^

But vanilla redstone connectiveness is such a sack of shit

And PR red alloy wires are a wholly different, incompatible sack of shit.

The whole way redstone signals work and propagate is tbh badly designed; especially when you start having microblock-like things like PR wires. But that's not something that can be fixed now.

[winter cedar]

[primal python]

Absolutely genius. Here I am with making a bunch of extra catalysts, and round-robin-ing them into the HILE's input busses and hoping that there's an even distribution between which catalysts get used

[zenith ether]

@fleet pivot this was my solution 🙂 counters incrementing by 1 which would just drop the appropriate quark into the multi, then using level maintainers to balance out quarks between runs.

[winter cedar]

AE2 Patterns are the most OP thing in the pack 🙂

[dark plinth]

holy hell thats tiny

gotta be the smallest t8 ive seen thats not oc

[odd ridge]

I thought there was some benefit to consuming 6 quarks at a time

like an infinity save

i have all of this pretty much but I also rename the quarks so they dont stack, and use the regulator to put in exactly 6 quarks at a time

[fleet pivot]

Well there is a benefit to putting in 6, as opposed to putting in 18. But no, infinity is consumed with every quark, so if you can stop the moment you find a match, you save the most.

[fleet pivot]

Ohh I love how simple this is 👍

[dark plinth]

Yeah this an optimization you can make for some more complexity

Whether that's worth it, 🤷