The Exspilement: self-powered Experiment 52B tamer

[Fradow]

That new Experiment 52B tree is super interesting: it allows transforming Food into Water (plus some Isoresin, that I personally never use). Boiling Resin also has a very interesting property: it creates heat (from 1.11DTU/g to 3.46DTU/g, about 3 times as much). And something that creates heat above 125°C can be used to run a Steam Turbine. Using advanced mechanics, that’s a great candidate for a self-powered build!

Enter the Exspilement

The name is a wordplay on “Spile”, a tool used to get water from trees and “Experiment” (thanks to Ishamoridin for the name). The Exspilement does that exactly that: it takes Food and transform it into Water and Isoresin, while being entirely self-powered.

How it works
There are two main ideas used in this build: pinning the tree close to 100°C, and superheating the Resin to 200°C to create heat.

First, the tree is pinned to around 97°C to 100°C, so that Resin is created at that temperature and requires less heat to heat up. This is done with a heat bridge to take heat from the Steam room, and use the fact that the tile just below the middle of the tree exchanges heat with it to precisely control its temperature.
Note that the heat exchanger uses regular tiles rather than metal/window tiles, to have a gentle heat introduction.
The tree room is vacuumed with a pump for finer temperature control. It isn’t absolutely required but it makes it easier to control the temperature.

Then, to maximize the heat created by the transition, Resin is superheated to 200°C (just below Isoresin melting point) before being allowed to change state by piping it to 1kg/s or lower.
There is a very small heat exchanger (which is enough to cut Aquatuner usage by 2 compared to directly sticking the Steam Turbine on top), which is kept at a low pressure to reduce heat exchange, and the Steam is gulped by the Turbine, creating Water and Power in the process.

That water is then used as a cooling dump by the Aquatuner, to cool the Turbine, and the excess is ejected, at around 60°C to 80°C.
The Isoresin is counterflowed in the Steam chamber as well to get back some heat, then through the Turbine room to lose a bit of heat, and ejected without temperature control. It should still be slightly above 100°C because there aren't a lot of cells to exchange heat, and it was not a goal of this design to get low temperature Isoresin.

Priming the build

• The Tree room doesn’t require anything special, as there is a Gas pump to empty it.
• The Steam Turbine room atmosphere is irrelevant, as it uses a puddle for cooling: any gas or a vacuum above the puddle will do.
• The Steam room needs to be vacuumed out. Since it has 4 tiles of height, you can use Brine, Salt Water, Polluted Water and Water stacked on top of each other, or vacuum any other way and put a bit of water.
• When starting the build, since it needs to be heated up by the Aquatuner, you’ll have a lot of excess cooling to get rid of. The simplest method is to keep the loop opened and cool down a pool somewhere. If it gets too cold, add a Tepidizer to negate that cooling.

Is it really self-powered?
When starting the build, of course you will need a consequent power input to heat up everything. After it’s up to operating temperature, as long as there is a steady supply of food, yes it should be self-powered.
I’ve only tested it for about 10 cycles (note that previous versions were not self-powered and could only last 3 cycles on a single Jumbo Battery), but I’m pretty confident that it will remain self-powered indefinitely.
On the other hand, that’s a really precarious equilibrium: when building in survival, I’d advise having a Manual Generator connected to be able to quick-start the build again if it somehow manages to lose too much power.

What about the food?
My initial goal was to use this water creation to oxygenate a few dupes. This is severely limited: my tests were done with Cooked Fish: to have a constant food supply for the tree, you’d need a population of around 750 Pacus, which is a lot. Then a Dupe, even with 10 Cuisine, is not enough to sustain 100% Tree uptime. If you somehow get 100% tree uptime, you will only have enough water to oxygenate 2 dupes using an Electrolyzer.
So while it makes it possible in theory to have a mini-base on a Marshy Asteroid with just 2 dupes, it’s in practice very tight and a huge investment to get there.
You could have more Pacus and feed the tree Pacu Fillet, skipping the cooking step, but then you can only oxygentate one Dupe.
You could have better food (more calories per kg), like Surf’n’Turf, but now you need more Dupes for ranching and Natural Gas for the Gas Range, and I really doubt there will be enough Dupe Labor to compensate. The only plausible option would be Barbeque using a starvation Shove Vole farm. Let me know if you try this out!

Materials
The regular tiles are made out of Sedimentary Rock, but any mineral should work.
The Window tile under the tree is Diamond, but any Metal tile will work as well.
Everything in the Steam Room must be made out of Steel.
As usual, the Steam Turbine foundation is made out of Ceramic, and all the other insulated tiles/pipes are made out of Igneous Rock as it’s plentiful.
Radiant Pipes are made out of Copper, but any other refined metal should work.

Automation details

• The Liquid Pump has a Hydro Sensor to avoid pumping partial packets.
• The Valve is set at 1kg/s, but you could set it lower for better heat exchange if you know you’ll have a lower throughput.
• The Thermo Sensor for the tree temperature should be set at a temperature to avoid overheating the tree. In Sandbox I could use 98°C but I recommend 97°C in survival.
• The Aquatuner is set to heat the room up to about 200°C, and will overshoot the target a bit, that’s why it's at 198°C. You do not want to make the Isoresin melt into Naphtha.
• The Auto-Sweeper is automated with a Timer Sensor to avoid running it all the time with small packets.
• The self-resetting Conveyor Meter is set at 1 unit to have better heat exchange.
• The Steam Turbine works only when there is enough Steam. There is a Buffer Gate to avoid pulsing the Steam Turbine, as pulsing reduces the Steam Turbine power output (enough to matter there).

Conclusion
I don’t expect that build to become mainstream as it’s more complicated than just boiling the Resin in a 125°C room and calling it a day, but the self-powering part is really nice because you don’t have to connect it to a main grid.
The water output is very disappointing compared to my initial hopes though, a Dupe must be a great cook and have loads of Pacu Filet at its disposal to even hope to sustain itself using that water. It’s a far cry from my initial objective of having a small colony thriving off of that tree.

 

[ghkbrew]

On 2/28/2022 at 3:01 PM, Fradow said:

What about the food?

I recommend pacu omelets for feeding the experiment 52B. As I described in this thread.

You can even avoid the huge pools by keeping the pacu flopping 90% of the time, only briefly submerging them to eat and lay eggs.

[Fradow]

8 hours ago, ghkbrew said:

I recommend pacu omelets for feeding the experiment 52B. As I described in this thread.

[SNIP]

You can even avoid the huge pools by keeping the pacu flopping 90% of the time, only briefly submerging them to eat and lay eggs.

That's a very interesting idea, that I automatically discarded because I don't tend to feed Pacus.

Which prompts the question: how do you make a renewable food source for those Pacus? And what's the size / labor cost of that food source? I don't see those questions answered in the thread you linked (though I might have skimmed too fast).

 

Note that I runned the number for grooming other critters (hatches/slicksters, since their food source is not an issue) to produce omelettes, and those don't look good: with napkin maths, a single high husbandry rancher will take care of 4 ranches of 8 critters. Wiki says that's 420kcal of omelette per cycle per hatch, so a total of 13440kcal/cycle. That's 67.2kg/cycle of Resin, or about 50kg/cycle of water. Which means 75g/s of Oxygen, while a single dupe need 100g/s.

[ghkbrew]

5 hours ago, Fradow said:

how do you make a renewable food source for those Pacus?

With seeds, I was using arbor trees.

Quote

Wild arbor trees with stacked liquids covering their branches.  The trees have 4 of 5 branches covered in liquid.  This causes the branches to drown and break immediately/repeated.  Each time a new branch grows there's a chance for the tree to spawn a new seed for the pips to rummage through them and find.  Each tree is producing a little over 1 seed/cycle this way.  I'm leaving 1 branch uncovered on each to let the pips eat.  That way you can just leave a bunch of tame, ungroomed pips in the room to extract the seeds.

But wild nosh beans are an option too.

5 hours ago, Fradow said:

And what's the size ... of that food source?

Well, here's the design I was playing with:

With this cooker:

But that was a rough first draft. That's way more trees than I needed and the giant pool could be eliminated by using flopping pacu. The cooker could probably be compacted by using storage bins to crack the eggs since they hold 10x as many.

I was targeting 29 fed pacu.  Each pacu lays 13 eggs in it's 25 cycle life, minus one to replace itself and each egg makes 5600kcal of omelet.  So 29 pacu produce 29 * 12/25 * 5600 = 77952kcal/cycle.

6 hours ago, Fradow said:

what's the ... labor cost of that food source?

That's the best part! 0 dupe labor.  Seed collection, pacu breeding and omelet cooking are all completely automated.

 

 

[Fradow]

Oh I didn't understand that part, that's why! A super nice idea. It handily wins against my Cooked Fish method (which, as I said before, I wasn't convinced with in the first place).
 

I'll revisit the food source using the fed Pacus idea when I got the time. The arbor trees are certainly a very attractive seed source.

[tuxii]

I am still playing around with Pacu and have come up with something quite interesting.

How many tame Pacu fed seeds does it take to max out the 52B on omelettes?

I have much more advanced version of what @ghkbrew demonstrated, although that seed farm is genius.

[Fradow]

Let's do some maths!

Experiment 52b consume up to 30kg per cycle. With Omelette 2800kcal/kg density, that's 84 000kcal per cycle (or enough to feed 84 normal difficulty dupes, that tree is a glutton). According to ghkbrew formula 2 messages above, you'd need around 32 fed Pacus to achieve that, or around 11 seeds/cycle.

According to the test you posted on Discord, where you say that 4 Arbor trees generated 56 acorns over 10 cycles, you'd need 8 Arbor trees to generate that much, so 2 times your small setup:

@Saturnus posted a minimal flopping pacu design on Discord as well:

Combining all 3 builds together should now be trivial, all that's left is for someone to do it and endurance-test it. With some Sweeper automation to reduce their power cost, it might even retain the "self-powered" property for all 3 setups together.

In theory, it should produce 420kg/cycle Resin, meaning a bit above 315kg/cycle (525g/s) water, enough to oxygenate 5 Dupes. That's way better than Cooked Fish, especially as there is absolutely no Dupe labor once the setup is built.

[Fradow]

After testing it for a while with a real food source (a full write-up will be available in a few days hopefully), I made an important modification: once the room is vacuumed out, the gas pump is no longer useful.

Instead, it's more useful to ship Rot Pile out before it turns into Polluted Dirt and starts off-gassing. To make some space so that the Auto-Sweeper and Conveyor Loader could be in contact with Resin (to evacuate their heat), I had to replace the Thermo Sensor with a Pipe Thermo Sensor:

The rest of the build is unchanged. It works perfectly.

Note that I've lowered the temperature a bit to 95°C. In longer session, it's safer, and the impact is minimal. I was tired of fiddling with it to get the perfect temperature.

[Nogard78]

Why don't use a mesh tile insted of windowed tile? (mesh doesn't exchange heat with the tree and so does food) With mesh you just preheat the tree to 99C put mesh tile and drop food. Even better, you can deep freeze your food so that it can't go bad and it just gets simpler.

(The pumps are here just for "if something goes wrong".)

[Fradow]

In my opinion "just preheat the tree" is more complicated and brittle. That's certainly a solution, but you need to build a one-use heater specifically for that. And then if anything goes awry and your tree temperature change, well you have to do it again. I like builds that "just work" with minimal fiddling, and don't consider a build complete if it doesn't include priming.

Deep-freezing the food is another condition that takes more infrastructure: again that's certainly possible, but a needless complication in my opinion. Especially when sometime you'll have a food packet that for some reason doesn't get deep-freezed, then rot, creates some PO2, and now your carefully initialized 99°C Tree isn't at 99°C anymore, and your food is thawing.

There are obviously more ways to build around Experiment 52b. This build is optimized to be self-powered, reliable and have the least requirements possible. If you optimize for different things, you can get to vastly different builds.