Fully automated fish farm giving one 1 egg per cycle for a cost of 1,33 kg algea

[AuroraFreya]

Hi! I have been playing around a bit with how to make a fish farm as efficient as possible. This is my first attempt so feel free to give me feedback on what to improve.

First a bit on my thinking. An egg takes 5 cycles to become a Fry, then the Fry takes 5 cycles to become a Pacu. Then this pacu needs 1,5 cycles to produce its first egg.

This means that the cycle from new egg to next egg produced is 11,5 cycles. When a pacu emerges it has 400 calories of food in its belly and it uses 100 per cycle. Thus it can produce 2 eggs at full speed (300 calories used up) and the last 100 calories is enough for it to produce one more egg after another 10 cycles (provided that we stop feeding it after those two first eggs are produced).

So if we want to use up those 300 calories of free food it would be good to have a new Pacu thrown into the feeding tank every 3 rounds. With an 11.5 cycle time from egg to egg 4 incubators will do the trick, the last one will only get to stay in the food tank for 2,5 cycles but who said life is fair...  

My setup is build in four stages.

1. The breeding part, here the egg transforms into a Fry and drops out when it becomes a Pacu.

2. Here the Pacu waits a bit until the Pacu in the breeding tank has dropped its next egg (this is important to make sure we get three eggs out of each Pacu).

3. This is where the Pacu spends is 3 breeding cycles popping out babies at full speed.

4. The graveyard... this is where it doesn't get more food but will pop out its last egg at age 17 and then starve to death at a ripe age of 20 cycles. (Providing us with some yummy Pacu Fillet (with this setup your duplicants will probably get quite fed up with their 7 days a week 3 meals a day fish diet after a while... oh, but that Omega 3 is so healthy for you!).

The automation might look a bit complicated at first, everything is set to standard settings expect the Filter gate for the doors to the graveyard, those are set to 1 second. (The XOR gate would open the door just a tiny bit in the end without this one).

To explain the automation from the top:

1. I pump out excess gas down to 500 g per tile... this is so that my water can flow freely.

2. The critter sensors in the breeding ground opens the floor when there is a creature in their pen... so that it can fall down nicely once it is ready.

3. The water is pumped up from the graveyard to cover 8 tiles (this is what is needed for a Pacu to not be crowded). I set it to be 600 kg so that I fill those tiles to 7,6 tones, the pump usually goes a little bit overboard before it stops and you don't need 8 FULL tiles, they just have to be wet for the Pacus to be happy).

4. Ok, here comes the first a bit move involved part. I start with an AND gate, when there is a fish in the tank the next time an egg passes the conveyor band it is time to drop out the old fishy (it has now laid its first two eggs). This is done by first opening the bottom doors with the XOR plate until 5 seconds have passed, at this time the top door opens which also closes the bottom one (XOR is fun!). This keeps open until the fish is gone (that resets the memory) but the buffer on the doors keeps them open a few more seconds to make sure enough water has followed).

5. At the same time the fish resets the memory for the conveyor sorter. The thinking here is that as soon as a fish has dropped down from the tanks we want the next egg to head up there and become a new breeding fish. The filter gate is just to make sure the first egg gets sent up to the tanks before the conveyor sorter shuts down until the next fish drops down from the breeding tanks.

The last part is just to fill up the water tank when you initialize the process. I would go for 48 tiles filled with water to have some space for the starving fishes to not get a cramped debuff. (That way you could have 6 dying fishes in there... my calculations says that it will rarely be more than 4 and never more than 5... so you have one in margin for when the system crashes due to a bug or something).

So this will produce one egg per cycle (or actually slightly more). (The breeding tank gives one every 1,5 turns and the starvation tank gives the rest). The total cost is 1 kg of algea per cycle in the main tank and 1 kg of fishfood in the pretank every 3 cycles. So total cost is 1,33 kg of Algea per egg you get out. Or well, only 2 out of three eggs actually leave the system so you could see it as 2 kg of Algea per produced external egg. (But the meat you get from the producing ones too of course).

If you don't save your hard earned eggs for a free fishpool later but instead put them in an egg-cracker you get 3733 calories of omelette and 1067 calories of cooked fish per turn, or in total 4800 calories per turn. If you instead save your egs for the future they will provide 64 calories per cycle indefinitely through rebirth every 25 cycles, this is the option I would go for, a 1000 calories diet duplet would need 15,6 fishes to stay fully feed, with this system you reach that in 23,4 cycles... so each 24 cycles you can add a duplet to your group and not worry about feeding them, or if you setup two of these babies you can add a duplet every 12 cycles... provided you can also give them oxygen etc...

 

Ok, nuff talking, here are some screenshots of the setup:

 

 

 

Give me some feedback on how I could improve this. (Of course it requires quite a lot of tech researched and quite some refined metal for all the Auto-sweepers etc, but still a fairly easy way to get high quality food where you basically can throw away the key once you started it...)

 

Oh, regarding starting it... those two manual switches are to turn the solid filter on the conveyor belt on or off... you want to space it out so that every second egg that gets produced is sent to the breeding tanks on your first fish... so that they will pop down with 3 cycles spacing later on. I guess you COULD automate that too but I was a bit too tired right now to figure it out.

[Nxf7]

Hey nice one. It's a bit hard to follow what's going on though. Could I get a save file?

I don't have any advice just yet. Below are my designs though. The 2nd one is for a single breeding unit which possibly the same goal that your build is trying to achieve. And the 3rd design is the scaleable version.

 

[degr]

Each time I open such thread because I want to see nice setup for pacu farm, but each time I see how complex is it and want to kill myself.

[AuroraFreya]

Here is the save file, thanks for the link to your builds!

Fish farm.sav

[Occam Blazer]

6 hours ago, degr said:

Each time I open such thread because I want to see nice setup for pacu farm, but each time I see how complex is it and want to kill myself.

The complexity comes from extreme min/max'ing. Here's a simple breeder tank that pops out pacu rapidly enough. If you set this up early you won't feel the sting of waiting around for the population to expand because it will grow with your dupes.

Spoiler

The tank. I'm growing gulp fish so I put it next to a cool salt water geyser to keep the pool cold. You can skip that part (and the overflow mesh tiles).

This is from the DLC and I'm feeding them seeds. The feeders are set to 1kg of a seed I'm making a lot of.

Automation. Keeps the conveyor chute closed if there are more than three pacu/eggs in the room. All other eggs are shipped out for sorting.

Shipping. Loader in the tank scoops everything. Loader above the tank carries everything except food and eggs.

 

[ghkbrew]

So if I understand correctly you're relying on the 11.5 cycle egg to first egg time and setting up 4 staggered cycles of that to keep 1 breeder in the tank for ~3 cycles at a time.  How stable is that in the long term? Is there something that keeps them from drifting out of sync? 

20 hours ago, AuroraFreya said:

Give me some feedback on how I could improve this.

I think you could compactify the build quite a bit by taking out the incubators and just dropping the eggs on the ground.  Give each egg a 1 wide room on top of an mechanized airlock with a dedicated delivery chute.  Then control delivery by selectively disabling chutes with automation.  It would also remove the need for sweepers in the upper chamber.

 

[AuroraFreya]

Yes you are completely right with what I am trying to do.

 

The reason I am using the incubators is because they keep the fry in there... Not sure how to make a stable timer to wait with opening the door until the fry has become a pacu otherwise.

As soon as they are Pacu I want to start feeding them though so that they really do produce 2 eggs in exactly 3 cycles. A delay would probably work out though as long as it is not more than 1 cycle long since that would mess up the “2,5 cycle” one.

 

What can mess things up a bit is that filter gates reset to zero when you load a file so could happen that a trap door in the bottom part doesn’t open right because of that. If you run it in debug speed (10x) it also freezes up sometimes so that the chute to the graveyard doesn’t open as it should. Haven’t seen that in normal mode though.

[Nxf7]

You reckon every 4th Pacu only gets 2.5 cycles to feed because you think there's a 11.5 lifecycle, but I don't really see how that 11.5 is forced. If anything there's a lifecycle of 3 that occurs (12 cycles per 4 Pacu). Also if a Pacu was only getting 2.5 cycles to feed, it wouldn't be able to lay a 3rd egg before dying of old age.

On 2/25/2021 at 6:49 AM, AuroraFreya said:

The last part is just to fill up the water tank when you initialize the process. I would go for 48 tiles filled with water to have some space for the starving fishes to not get a cramped debuff. (That way you could have 6 dying fishes in there... my calculations says that it will rarely be more than 4 and never more than 5... so you have one in margin for when the system crashes due to a bug or something).

This isn't a concern. They only need 1 tile of water and 8 tiles of room size.

8 hours ago, AuroraFreya said:

What can mess things up a bit is that filter gates reset to zero when you load a file

I haven't experienced this, but be wary of critter sensors. They seems to toggle to the opposite state for a varied amount of time, usually less than 0.7 seconds. After each critter sensor, a filter or buffer or both may be necessary to prevent anything disastrous.

[Occam Blazer]

9 hours ago, Nxf7 said:

I haven't experienced this, but be wary of critter sensors. They seems to toggle to the opposite state for a varied amount of time, usually less than 0.7 seconds. After each critter sensor, a filter or buffer or both may be necessary to prevent anything disastrous.

I built a shove vole condo that would auto-murder residents after they laid their replacement egg. Every time I loaded the save I had to reset each of the twenty rooms because the critter sensor switched the room to kill. A filter gate would break the room because *engineering* so I just lived with it.

[AuroraFreya]

12 hours ago, Nxf7 said:

You reckon every 4th Pacu only gets 2.5 cycles to feed because you think there's a 11.5 lifecycle, but I don't really see how that 11.5 is forced. If anything there's a lifecycle of 3 that occurs (12 cycles per 4 Pacu). Also if a Pacu was only getting 2.5 cycles to feed, it wouldn't be able to lay a 3rd egg before dying of old age.

I actually just realized my model above is flawed... because if I put the second egg up my cycle time is 13 cycles and it will fail me on the fifth iteration.  Because then the new fish will wait for the old fish’s third egg and that will take too long so the whole system will choke. Guess I need to set the gate so it is the second egg passing after a new fish has dropped that should go up (i.e. the first egg of the new fish).

You are right in that they all have 3 cycles, what will happen is just that one of them will stay in the top tank for a bit longer than the rest... so the output (when I have fixed so that it actually IS 11,5 cycles and not 13) will be 12 eggs in 11,5 cycles.

What the system still optimizes on is that I only feed the fishes only the times that the food gives me meaningful value in increased production, which I guess is a bit fun if you really want to reduce fish-food cost

12 hours ago, Nxf7 said:

This isn't a concern. They only need 1 tile of water and 8 tiles of room size.

You are totally right, I should have thought of that... still 8 tiles per fish so almost the same room size but at least I conserve a bit of water...

12 hours ago, Nxf7 said:

I haven't experienced this, but be wary of critter sensors. They seems to toggle to the opposite state for a varied amount of time, usually less than 0.7 seconds. After each critter sensor, a filter or buffer or both may be necessary to prevent anything disastrous.

Hmm, Okay, I’ll add that to the lower critter sensor, the top ones doesn’t really matter if they flip wrong for a while... Big thanks for all your suggestions btw! Try nesting a few filter gates on 200 seconds and then save and reload somewhere in the middle of the cycle... when you load it will be back to zero again unfortunately.... in the beginning I had all these plans to use timers but gave up that idea due to this fact. (Plus that this is simpler=Hopefully more fail safe...)

[Nxf7]

2 hours ago, AuroraFreya said:

You are totally right, I should have thought of that... still 8 tiles per fish so almost the same room size but at least I conserve a bit of water...

Actually I thought of one benefit to giving them 8 water tiles per Pacu. If the feeding Pacu is dropped in there early it won't lose its reproduction buff.

So one possible design option is to have each Pacu only in the feeding chambers for 2(.01?) cycles only, feeding it algae at the start, middle and end of that period. And of course the retirement home on the bottom needs sufficient water.

I could try that to double or even triple the output of my Pacu farm, but probably at the cost of double or triple the vertical height at least. Probably not worth it but it could be fun.

2 hours ago, AuroraFreya said:

Try nesting a few filter gates on 200 seconds and then save and reload somewhere in the middle of the cycle... when you load it will be back to zero again unfortunately

This might be a new bug with DLC. I don't have DLC and loaded a save file which has nested buffer-filters, and there was no problem.
Also the bug might not be with filter gates, but whatever is giving them the ON signal, it must be turning off briefly upon load.

[AuroraFreya]

Oh, so the reproduction needs 8 tiles of water? In that case I do need it because the third egg is being produced in the graveyard (or retirement home if you prefer). If I would have them less than 3 cycles with a feeder they would slow down their calorie consumption to early though, I want them to burn through that first 300 calories at high speed to make sure they have time to give me three eggs before producing dying.

And yes, could be that it is a DCL thingy... feels like they are trying to reduce the amount of information they store in the save files. They remember if it was active or passive but always restarts at 0... not a big problem in most cases I guess but if you try to nestle them to run for 3 cycles as I did then you definitely noticed.

 

On 2/26/2021 at 7:56 AM, Nxf7 said:

You reckon every 4th Pacu only gets 2.5 cycles to feed because you think there's a 11.5 lifecycle, but I don't really see how that 11.5 is forced. If anything there's a lifecycle of 3 that occurs (12 cycles per 4 Pacu). Also if a Pacu was only getting 2.5 cycles to feed, it wouldn't be able to lay a 3rd egg before dying of old age.

 

I actually just realized my model above is flawed... because if I put the second egg up my cycle time is 13 cycles and it will fail me on the fourth one