My take on an automated Puft ranching design

[Ayreon98]

I never really used Pufts in past playthroughs. With the introduction of Geotuners and geysers that required bleach stone, I wanted to try to come up with my own modular design. The goals were for it to be automated without using critter drop-offs; self-sustaining, given enough gas for the desired population; and easy to blueprint for as many ranches as materials allow.

It uses a few known concepts with managing the Puft Princes from other designs I've seen, and it's been stable for about 1000 sandbox cycles. I thought it was a fun challenge to work around the limitations of sensors and critter drop-offs, which can't detect what kind of egg or critter is present in the room.

General Layout, with Two Ranches Side-by-Side (Decor items optional):

Automation:

Conveyor System:

Initial Setup:

1. I set up two ranches outside of these: one for Pufts, and one for Puft Princes. The Pufts will mostly lay Puft Prince eggs, and the Puft Princes will mostly lay Puft eggs. That part of the population is easy to manage. I don't think these need to be of any particular design or geometry, so long as there's enough polluted oxygen for the Pufts, and the critter count in each room allows them to breed. By mid-game, I usually aim to have a surplus of water - and therefore oxygen - so I usually supply the dedicated Puft Prince ranch with plain oxygen.
2. To start breeding the other morphs, I use one "block" from the attached images. These don't require critter drop-offs to get started if using critter sensors and automation as described.
3. This block houses Pufts with one Puft Prince, which will start to breed the other Puft morphs.
4. Then I make another copy of the design next to the starting block for Dense or Squeaky Pufts.

The Block:

Each stable is 96 tiles, and they fit in a 4-tile height. The space underneath each block is just a transit way for duplicants and conveyors/solid filters. I'm using late-game liquid locks in sandbox, but any kind of liquid lock at the entrances will do.

A few blobs of liquid need to be placed in three key places:

1. At the bottom-left corner where the conveyor chute is, which is where eggs get dropped off.
2. Between the Puft Prince's section and the rest of the room.
3. If desired, at the storage bin to stop the resources produced from off-gassing.

Naphtha works well due to its viscosity, but so long as there's enough to form a full, stable tile of liquid between two other solid tiles, the flying critters won't cross it. Hatchlings will find a path away from it into the open room.

The placement of the auto-sweepers is such that they have full coverage of each room partition. The conveyor loaders can accept all materials. The grooming station keeps the morphs happy, but won't keep the Puft Prince groomed. If I understand the mechanics of the "Glum" de-buff, this is actually a good thing. The Puft Prince doesn't produce a favorable ratio of resources from the gas they consume, and "Glum" will reduce their metabolism by 80%.

Logic and Process:

Note: The Signal Switches aren't required, and were just for testing.

The automation is driven by the Puft Prince's lifecycle. As known from other players' designs, the meat dropped by a Puft Prince is a reliable way to detect that the ranch needs a new one. I use a Conveyor Rail Element Sensor for this. An OR gate with a critter sensor set to `below 1` can be used to get the ranch started without manually transporting a Puft Prince into the room.

The element sensor starts a series of Memory Toggles for each phase or "state" of the Puft Prince replenishment process. The logic I came up with is as follows:

1. The stable has no Puft Prince, OR the Puft Prince has dropped meat.
   1. The stable needs a new Puft Prince egg.
   2. Other morph eggs should NOT be allowed in until the Puft Prince egg arrives, so as to avoid having to wrangle surplus.
2. The weight plate must NOT be above 0.4kg. This makes certain that the next thing that falls on it is an egg, which weighs 500g.
3. Once the weight plate is clear, the auto-sweeper should NOT try to pick anything up until the egg hatches and the weight plate resets.
4. A Signal Counter is set to `1` in `Advanced Mode`. This generates one, and only one, green pulse to reset a Conveyor Meter. The Conveyor Meter is combined with a Conveyor Shutoff to allow one, and only one, Puft Prince egg. The inputs to each conveyor component overflow back to the main Conveyor Rail to supply other ranches or an evolution chamber.
5. Conveyor Rails from the breeder ranch of Pufts transport a Puft Prince egg. A Solid Filter allows the use of a single conveyor rail for any eggs or resources, and filter out only Puft Prince eggs when needed.
6. The egg drops down from a chute and sets the weight plate. The weight plate is an input into a Memory Toggle so that it can detect being unset when the egg hatches. It also allows the other morph eggs to be delivered again via conveyor chute if needed.
7. Once the egg hatches, it generates the signal used to reset the all the Memory Toggles. It waits until meat is detected again.

Automation and Filter Settings:

What I like about the design:

• It's 4-tile room height. For some reason, that's very important to me.
• It appears to be stable. After 1000 cycles, the population didn't seem to suffer any crashes or failures. No unusual behavior between saves/loads as well.
  • If the population does crash, it's self-recovering so long as the other dedicated Puft + Puft Prince ranches are reproducing (the resources for which are more straightforward to sustain).
  • The ~6-18 cycles it takes to replace a Puft Prince seems unavoidable, but hasn't caused issues, and specialty puff egg counts usually remain positive.
• A critter sensor can be used to fine-tune a target number of Puft morphs (N - 1, for the Puft Prince).
• Multiple ranches can chain together off a single conveyor rail. Solid Filters redirect the different egg morphs as needed, and all excess can go to a centralized evolution chamber.
  • A few auto-sweepers and a conveyor loader above the evolution chamber can run on automation signals to load eggs when needed by the ranches, but I only used this for Puft Prince eggs to minimize the time that a ranch goes without one.
  • This also makes it easier to expand if another ranch is desired. Just extend the rail system over to a new block.
• It's (relatively) simple. There are only a few automation settings to configure once it's built.

Other notes:

Mirroring the design - that is, having the partitions and room entrances on the opposite side - makes it much more difficult to route the automation wiring. It's possible, but requires more bridges and is confusing to follow.

Although the pictures show a pool of polluted water underneath one of the ranches, I'd be more likely to just pay the power cost and pump gases in using the gas vents in each stable. I decided not to include Morbs in the sandbox play, as I was more interested in chlorine -> bleach stone production. Maintaining atmospheric pressure also helps reduce a bit of material off-gassing, though that's more of a concern if transporting the materials to other map locations, rather than the short amount of time they spend on the ground in the ranch before being auto-swept.

I haven't used this in a non-sandbox game yet. Since I don't normally ranch Pufts, I don't have a good figure for resource production vs. Puft consumption, especially for elements like chlorine. That said, chlorine vent output can be boosted with salt, and the population can be easily controlled, so the sustainable numbers would just take some math to figure out. I'd be curious to test if boosting a Salt Water Geyser with 2+ Geotuners produces enough salt to boost one or more chlorine vents, and if that ends up being a power/resource-positive process - but that's an experiment for another time.

[myxal]

Yaay, another ranch design for the under-appreciated mascot critter of the game!

14 hours ago, Ayreon98 said:

easy to blueprint for as many ranches as materials allow.

As someone who has gone through this exercise 2 weeks ago, my advice is: don't. I've come to the conclusion that with pufts, it's just too costly/inefficient/complicated to have just 1 design for the mixed ranch and keep repeating it - rather, I consider what resources I need pufts to produce and in what quantities, and together with the mutual egg sustainability, build multiple different ranches, which provide eggs for each other. Most of the ranches end up being trivial.

TL;DR conclusion is that a single full-size "complicated" mixed ranch is able to sustain ~11 "trivial" ranches* , with only worker morphs. Dense and squeaky pufts will only replace their own morphs, slime pufts in a mixed ranch will replace both squeaky and dense, but not necessarily their own morph. If if all you need is slime, 2 trivial ranches with princes and slime pufts work just fine. Scale the prince ranch (and/or disable grooming) according to the number of slime pufts to avoid wasting resources and that's pretty much it.

* Theoretically. In practice, prince downtime will bring this number down, and is a reason I aim to replenish the mixed ranch with an adult prince, or at least a prince puftlet. I'm working on a "puft prince dispenser", that can reliably release 1 puft prince into the ranch upon request.

14 hours ago, Ayreon98 said:

As known from other players' designs, the meat dropped by a Puft Prince is a reliable way to detect that the ranch needs a new one. I use a Conveyor Rail Element Sensor for this.

I had this idea too, and my design broke immediately because the idling dupes in the colony waltzed in and carried the meat themselves to the manual edible drop-off loader. Good job walling the prince in, but I'd put the liquid "window" somewhere higher up, so dupes definitely cannot reach the meat dropped from the prince. 

As for the build itself - If I understand correctly, the prince egg is coming in freshly swept from this or other ranch, right? So it's at low incubation - without incubator it'll take 15 cycles to hatch, and then another 5 cycles before the baby turns into adult so the "prince influence" is restored. That seems like quite the downtime, I'd love to see how many worker eggs this ranch can produce over, say 100 cycles. But then again, if you are repeating this design for all puft ranches, then you don't need to care, I suppose.

 

 

[Ayreon98]

1 hour ago, myxal said:

I've come to the conclusion that with pufts, it's just too costly/inefficient/complicated to have just 1 design for the mixed ranch and keep repeating it

I appreciate the feedback, and I'll keep that in mind. I was watching the egg counts as mine was running, and I'm inclined to agree that this wouldn't need to be stamped out for every specialty Puft ranch. The Dense Puft egg counts would peak at around 20 as-built in the screenshots, which seems more than enough to sustain a few ranches that don't have the Puft Prince.

1 hour ago, myxal said:

Good job walling the prince in, but I'd put the liquid "window" somewhere higher up, so dupes definitely cannot reach the meat dropped from the prince. 

I tested this - the inner room floor is out-of-range of manual dupes. I dropped in materials on each floor tile with the auto-sweeper off and tried to manually sweep them, and the dupes wouldn't pick them up.

I sandboxed it with the liquid tile another cell up, but I like the one-tile high placement because there's no need to build extra tiles to get a bottle emptier to fill the tile. If something in the inaccessible room breaks, it's a little easier to refill if the liquid if it spills out. Just break the tile above the liquid, and replace it when finished.

1 hour ago, myxal said:

So it's at low incubation - without incubator it'll take 15 cycles to hatch, and then another 5 cycles before the baby turns into adult so the "prince influence" is restored.

I had mine sweeping in from an evolution chamber, so there's a chance that some eggs are partially incubated. You're correct that it's still ~20 cycles at worst.

I tried all kinds of things to speed that up with an incubator, a drop-off, and automated doors to let Puflets into the room, and then shut the door behind them. I kept having issues where the critters could get trapped in the door rather easily.

I thought about trying to work around it with waterfalls or more Memory Toggles, but it seemed to just add a lot of complications. A waiting room with a drop-off would have to unlock when a Puftlet is needed, and then lock and remain locked when the duplicant leaves. Then the doors into the ranch would either need Memory Toggles that could "retry" if counts didn't normalize (the counts of which aren't guaranteed to remain because of the shared ranch), or there would have to be automation to start flooding the room to force the Puft past the door, and then drain it again on reset.

Alternatively, using layers of multiple liquids so that the critter drop-off doesn't register as being flooded would probably work, but I think it would be awkward to set up. The incubation and maturity times ended up being a trade-off I was willing to accept.

[myxal]

6 hours ago, Ayreon98 said:

I tested this - the inner room floor is out-of-range of manual dupes. I dropped in materials on each floor tile with the auto-sweeper off and tried to manually sweep them, and the dupes wouldn't pick them up.

Thanks, good to know. I avoid making things inaccessible, I wish dupes would get their own range visualisation

6 hours ago, Ayreon98 said:

Alternatively, using layers of multiple liquids so that the critter drop-off doesn't register as being flooded would probably work, but I think it would be awkward to set up.

This is exactly what I did in my semi-automatic design. It's doable

I know of 3 ways of stacking liquids: spill & mop, spill & drop, and metered spilling. I used metered spilling - the piping required is a hassle and you want to postpone adding any ranch automation where the vents/meter goes, but it makes the building process SO much cleaner. With the non-metered methods I'm not sure if the drop-off might become flooded, but IMHO that can be fixed by de- and re-constructing it after the stacked liquids are set up and at low mass.

7 hours ago, Ayreon98 said:

I thought about trying to work around it with waterfalls

Yep, currently wrestling with this as well. Door liquid pumps inexplicably spill liquid out the top, and waterfall stops being a wall of water in an inconvenient place.

[degr]

Here is my design for puft ranch:

all eggs are distributed into chambers, in each chamber I have critter drop off which is set to max critters 0, auto-wrangle on. I have gas vents which brings clorine and polluted oxygen into chambers. No automation, only some conveyor rails. Works somehow during 1000+ cycles. From time to time I kill wrong pufts manually.

[degr]

I think a little bit, it could be improved with timer sensor, counter and critter sensor. It should just count 75 cycles (or how long puft prince lifespan takes), and allow duplicants to enter puft prince chamber until puft ranches countain less then 6 critters. After it, counter reset, and duplicants are allowed to enter only chlorine puft chanmber and normal puft chamber.

[Ayreon98]

4 hours ago, degr said:

I think a little bit, it could be improved with timer sensor, counter and critter sensor. It should just count 75 cycles (or how long puft prince lifespan takes), and allow duplicants to enter puft prince chamber until puft ranches countain less then 6 critters. After it, counter reset, and duplicants are allowed to enter only chlorine puft chanmber and normal puft chamber.

I remember seeing a design by EchoRanch with a cycle counter. I think the concept is similar to what you described, but also involved timing the doors to the incubators and whatnot.

I did a little more messing around with liquid + critter drop-offs, too. One thing I found interesting is that the critter drop-off building doesn't register as flooded unless two tiles of the same liquid stack above it, which would make it a little easier. I thought it was related to pressure - that is, how much liquid is on the first tile - but it seems to allow any amount, so long as the tile above it is a different substance.

My thought with that is each ranch could have a small loading section beneath it or to the side, and three tiles high for three different liquids. Drop-off limit of 1 for the desired morph so that they only get dropped off one at a time, and then automate locking the door when the ranch reaches its limit. Liquid has to cover the door in whatever orientation is has to prevent other critters from pathing back into it, and to force the delivered critter into the room. It wouldn't be hands-off like my first iteration, but would allow the use of incubators to speed up critter replacement.