So I set out to explore Arbor Acorns farms, and in the process I discovered (or rediscovered) some very interesting details that I want to share here.
Disclaimer: since the method used triggers frequent room rebuilds (each time a branch respawn), building Arbor Acorns farm will have a non-negligible impact on performance.
But first, why? Once you get enough Acorns, you are done, so why? Well, Pacus can eat seeds now, so a self-sustained Arbor Acorn farm allows feeding Pacus.
But, I hear you say, why feed Pacus when you can have unfed Pacus forever? Well, fed Pacus produce an excess of eggs. Those eggs can be turned into Omelette without Dupe labor, and they are the highest calorie/kg food you can produce without Dupe labor.
That’s important because that’s what you want to maximize Experiment 52B. A good Arbor Acorn farm is the foundation to that.
This idea was introduced by @ghkbrew here:
Now that that’s out of the way, let’s focus on those Arbor Acorns farms. As a warning, no matter the actual design, it’s very hard to have an accurate number for the number of acorns produced by cycle, because there is a huge variance. Take any number with a grain of salt, and aim to overproduce to ensure variance doesn’t bite your ass.
All designs are based on this old discovery that you can break branches with liquids to boost Arbor Acorns production:
Here is the main picture of arbor acorn farming for reference:
Let’s get back to the basics, because that link doesn’t explain the details. A Tree trunk takes 2 cells, and it can grow branches in 7 cells, shown with drywalls. The left-most example shows those 7 cells.
To get more Acorns, you need to break branches so that new branches grow again: acorns have 5% chance of being generated when a new branch starts growing.
The only way to break branches fast (kinda) and automatically is to have “full” liquid cells over the branches, either by liquid stacking or by using a waterfall, as shown above. When drowning them, new branches are generated about every 80s (about 75s of drowning followed by a few seconds before the new branch replaces it), and a tree can only have 5 growing branches at a time.
As you cannot easily have a full liquid cell in the cell above the tree trunk, it’s better to prevent branches from growing there in the first place by placing any building over it after the tree is planted: ladder, crown moulding, or even a useful building like an auto-sweeper or conveyor loader.
That leaves 2 main choices for tree spacing: either one every 2 cells (middle example in the screenshot above), or one every 3 cells (right example). It isn’t quite easy to choose, as both have pros and cons.
When doing one tree every 2 cells (note: pip planting will require a 2 cells gap every 3 trees), a tree shares branch spots with its 2 neighbors. This means a tree with 1 neighbor will have 3 guaranteed spots and 3 non-guaranteed spots, averaging 4.5 branches spots, slightly under the 5 spots maximum. A tree with 2 neighbors will have 6 non-guaranteed spots and average 3 branches. Since in both options, trees will always use all available spots, you might as well have designs with a single neighbor.
When doing one tree every 3 cells, every tree will have 6 guaranteed spots and will use 5 out of them: this means one of those cells can be used for something else.
Overall, even with 1 cell being used by something else, one tree every 3 cells will have a better branch density for the same width, as there are less tree trunks, and since you can fit machinery in those unused spots, it will take less height, so it will be better from a purely space-efficiency standpoint. Here is an example of this, which is nearly perfect, and has a perfect rectangular shape:
That spacing has a huge flaw though (thanks to zivvel for pointing it out to me): Pips cannot cross those liquids. This means there must be at least a Pip per tree.
To allow Pips to cross, you need to have hop liquid locks, which are only possible with one cell of spacing. Unfortunately, that means more height clearance: you will need to lower the floor to allow Pips to hop (they cannot cross liquids if it starts at the same level as trees), and you will need to fit auto-sweepers and loaders. Here are 2 designs that do that. The right one was designed by @tuxii, and while it has a bulge, it is symmetrical, and replacing a tree with a sweeper is actually a good idea anyway, since all trees on that design have a single neighbor.
Allowing pips to cross also has another side-effect: there can easily be more pips accessing each tree, which means more chances to rummage a tree before the next branch producing an acorn overrides the previous one in the trunk, since the trunk can only store one acorn.
After lengthy tests, over long periods, having a single pip per tree lowers the probability to actually get an acorn out of a branch. At most, it drops from 5% to 4%, though in most cases it will be around 4.5% +/- 0.25%. It’s also resilient, which is why I would advise going for one tree every 2 cells spacing.
To allow tame pips, you need to feed them. This means leaving a few branches uncovered, such as in @ghkbrew design:
Obviously, this will lower yield, as those branches are not producing acorns.
The main question you might be asking is: how many acorns can you expect?
With long running experiments, the average is around 0.345 acorn/cycle for each branch spot with single stuck pips and 0.378 acorn/cycle for each branch spot with several roaming pips.
That means the one tree every 3 spots design (left) with 15 spots should average slightly above 5 acorns per cycle, the compact one tree every 2 spots (middle) with 15 spots should average slightly above 5.5 acorns per cycle, and the larger one (right) with 18 spots should average slightly above 6.5 acorns per cycle. To be conservative, you should count on one less acorn per cycle for all variants, as explained above.
There you have it, all there is to know about Arbor Acorn farms, and a few example designs to get started.
