[Setup] Fish'o'matic station : a way to auto-breeding fishes

[StalightStryker]

@SackMaggie Not exactly, they do not get crushed to death, they are very much alive within the doors, but you don't get that pesky critter starvation notice and they don't count as present in the pool and will die eventually of age/starvation. For all practical purposes, they may as well have been crushed.

I unfortunately couldn't find any reliable way of strait up killing the pacu except freezing or overheating them, but ever since we've lost our lovely abyssalite I'm hesitant to use anything with a high or low temperature near temperature sensitive things like this.

I could use a tile of vacuum to isolate it from the pools but the chance of polluted dirt gassing off and messing with the vacuum is present.

Thinking of this, and combining it with @Nxf7 's rightful claim that pacus don't like x3 speed at all and may bug sometimes (especially if left unobserved for some magical reason), I may have to do a slight remodel to make it bugproof and actually kill the pacus for real. 

As it is, the system would fix the bug automatically but you'dd miss out on a pools production for a few cycles due to overcrowding which is, of course, not where I was going for. I just want food to be a worry of the past and focus on building my meteorite defence system.

As for the purple circle, it's not disconnected since it's not meant to be powered at all, thus automation crossing it is, well, irrelevant. That wire is purely for the memory cells. Those loaders are just there for the sweepers to have an input for dropping the eggs.

Good thinking on the memory cells, I will definitely be streamlining that.

[SackMaggie]

15 minutes ago, StalightStryker said:

Not exactly, they do not get crushed to death

I forgot that the living things in this game is magical being.

[StalightStryker]

@Nxf7

I took the time to look at your pacu farm and man that is amazing, and so much more efficient than mine! I think I'dd better forget all about my design and copy yours. I even tried if I could make like any serious improvement to your design but it's just... better, I guess? Seems a shame of the time I invested in building my contraption.

The one thing I don't like is the pacu delivery system. It does bug every now and then. I think I'm gonna build something to improve it so I can keep a shred of dignity after this ordeal.

Fun fact though; there is one thing my farm can do that has possible use: separating the pacus from their water (with a little added automation, that is) so you can use the all-done-reproducing pacus as coolant. Yes. Seriously. This worked out surprisingly well since they're made of ooze (3.470 heat capacity), weigh 200 kilograms and can be heated by 50 degrees (from 30 to 80) before they kick the bucket. Unfortunately I'm not really in need of coolant now since I've found a slush geyser but hey, it's something, right?

Anyway, thanks for the design, definitely using that from now on

[Nxf7]

I appreciate the kind words @StalightStryker. It's taken a lot of iteration, so trust me when I say I've wasted a lot of time on poorer designs because I was tunnel visioning on particular limitations and solutions. All fun though.

Which design concept is better is still subjective, mine was to be "survival friendly" and not need too much refined metal relatively speaking. But unfortunately the issues it has on x3 speed is not to be ignored, and the Pacu sharing a feeding pool makes this worse, because the idea was they all eat at the same time or else half of them could eat everything. Perhaps Klei can throw me a bone here and not make Pacu drunk when on x3 speed.

Other concerns: they do wilee coyote and don't fall when on x3. Storage containers can bug and get "unreachable" so sweepers ignore them (but this is manageable).

Please do improve it  

Just going to link it again for others and convenience: https://m.imgur.com/a/s7bkqF7

[StalightStryker]

 @Nxf7In this build I tried combining the best of both worlds, going for total reliability and zero maintenance. I quickly ran into trouble with your spacial efficient multipacu-pools, since the pacu do not seem to share their food quite readily, like you said.

In an attempt to work around this I went back to my less efficient single pacu pools, but this time with your design in mind.

Also, the egg timer, though fabulous, was not completely reliable over longer amounts of time, I've noticed the egg dropping may lag sometimes or the fish would flop around for a full pacu cycle and when that happens the entire system shuts down (it is easily reset but I was going for zero maintenance). So I went back to the water clock.

The dropping system has come along nicely, I've tried to get it foolproof and so far it's working out I would say. The trick is (in my experience) not having to rely on flopping fishes dropping off a ledge because those damn fishes never do what you would expect them to. To fix that issue I have doors push the eggs off a ledge before they hatch so they won't need to flop. Also, I set the droptimers to 60 seconds each, which is more than enough time for even the Wiley-e-coyotiest of fishes to actually fall. I've had it run for over 100 cycles without any problems now, so I'm willing to call it a day and enjoy the basically free food.

I still have a question though, how many eggs do the fish lay in your 'waiting for death' pool? Mine lay one (their third, if they didn't during active breeding) and that had me wondering about what replenishing time would be more efficient, 5 cycles for the full use of calories still produces three eggs: (67% x (440 calories / 100 calories per cycle) = 293% after which they die due to starvation in 10 cycles (10 days * 7% a cycle = 70%) with the total being 363% so three eggs with 5 days of active breeding.

If we keep them breeding for three days we will use 300 calories and produce exactly 2 eggs. The fish is then dropped into the waiting pool where it lives the rest of its life 7 (140 calories at 20 calories/cycle due to glumness) + 10 (days until it actually starves) = 17 cycles. 17 x 7% = 119%. That would put that total at 319% per fish but with more fishes going through the active period, and should yield appreciably more eggs and meat per cycle for the same cost in algae. Must admit I have not tried the set-up thoroughly with these settings (reducing the hydro sensor to 18 should work) so it may not work properly. Use at your own risk.

It does, like you said, cost quite a bit of refined metal to build all those sweepers/loaders/automation required for single pacu pools and that is definitely a disadvantage of this build. Though with a metal volcano or some gold refining it should be manageable I suppose.

Anyway, here goes: 

PS: the power grid looks a bit drunk, that's because at first I intended to hook every line up to a transformer, but then I realised most devices can't ever work simultaneously and a single line suffices . Feel free to adjust it.

[Saturnus]

@StalightStryker You can use the new critter sensor to detect when eggs drop in the breeding pool, feed that signal to a simple toggle switch so the system is triggered by the pacus actually dropping the second egg. Or more exactly when the 2nd egg has been picked up and carried out of the breeding pool room by the conveyor. That way the timing is always perfect and will never need adjustments.

[RonEmpire]

3 hours ago, StalightStryker said:

 @Nxf7In this build I tried combining the best of both worlds, going for total reliability and zero maintenance. I quickly ran into trouble with your spacial efficient multipacu-pools, since the pacu do not seem to share their food quite readily, like you said.

In an attempt to work around this I went back to my less efficient single pacu pools, but this time with your design in mind.

Also, the egg timer, though fabulous, was not completely reliable over longer amounts of time, I've noticed the egg dropping may lag sometimes or the fish would flop around for a full pacu cycle and when that happens the entire system shuts down (it is easily reset but I was going for zero maintenance). So I went back to the water clock.

The dropping system has come along nicely, I've tried to get it foolproof and so far it's working out I would say. The trick is (in my experience) not having to rely on flopping fishes dropping off a ledge because those damn fishes never do what you would expect them to. To fix that issue I have doors push the eggs off a ledge before they hatch so they won't need to flop. Also, I set the droptimers to 60 seconds each, which is more than enough time for even the Wiley-e-coyotiest of fishes to actually fall. I've had it run for over 100 cycles without any problems now, so I'm willing to call it a day and enjoy the basically free food.

I still have a question though, how many eggs do the fish lay in your 'waiting for death' pool? Mine lay one (their third, if they didn't during active breeding) and that had me wondering about what replenishing time would be more efficient, 5 cycles for the full use of calories still produces three eggs: (67% x (440 calories / 100 calories per cycle) = 293% after which they die due to starvation in 10 cycles (10 days * 7% a cycle = 70%) with the total being 363% so three eggs with 5 days of active breeding.

If we keep them breeding for three days we will use 300 calories and produce exactly 2 eggs. The fish is then dropped into the waiting pool where it lives the rest of its life 7 (140 calories at 20 calories/cycle due to glumness) + 10 (days until it actually starves) = 17 cycles. 17 x 7% = 119%. That would put that total at 319% per fish but with more fishes going through the active period, and should yield appreciably more eggs and meat per cycle for the same cost in algae. Must admit I have not tried the set-up thoroughly with these settings (reducing the hydro sensor to 18 should work) so it may not work properly. Use at your own risk.

It does, like you said, cost quite a bit of refined metal to build all those sweepers/loaders/automation required for single pacu pools and that is definitely a disadvantage of this build. Though with a metal volcano or some gold refining it should be manageable I suppose.

Anyway, here goes: 

PS: the power grid looks a bit drunk, that's because at first I intended to hook every line up to a transformer, but then I realised most devices can't ever work simultaneously and a single line suffices . Feel free to adjust it.

your setup looks interesting.   do u have individual timers for each of the fish pool rooms?  or they're all randomly staggered and you're estimating.

[Nxf7]

8 hours ago, StalightStryker said:

 If we keep them breeding for three days we will use 300 calories and produce exactly 2 eggs. The fish is then dropped into the waiting pool where it lives the rest of its life 7 (140 calories at 20 calories/cycle due to glumness) + 10 (days until it actually starves) = 17 cycles. 17 x 7% = 119%. That would put that total at 319% per fish but with more fishes going through the active period, and should yield appreciably more eggs and meat per cycle for the same cost in algae. Must admit I have not tried the set-up thoroughly with these settings (reducing the hydro sensor to 18 should work) so it may not work properly. Use at your own risk.

Looks great! As for the above quoted, only issue with that is eggs are still going to take 5 cycles to pop and another 5 to become an adult, so the farm would have to be designed with that in mind. Otherwise if you're shaving off 4 cycles, you'd be dropping an age:1 fish instead of age:5 in for feeding and they aren't going to lay anything for ages. The math looks sound though.

I haven't extensively tested results of fish fighting over algae at x3. If it's not that bad my "Economies Of Scale" build could probably still function at x3 with a few tweaks taken from yours. Each door layer being held open for ~60 seconds is heavy-handed but made necessary in this case  

4 hours ago, RonEmpire said:

your setup looks interesting.   do u have individual timers for each of the fish pool rooms?  or they're all randomly staggered and you're estimating.

They are synced so that automation doesn't need to be duplicated. When first creating the system the pacu's ages might all be different, but after a generation or two they'd all be the same age, forced upon them by eggs being stored halting incubation progress no matter when the egg was laid.

[RonEmpire]

34 minutes ago, Nxf7 said:

Looks great! As for the above quoted, only issue with that is eggs are still going to take 5 cycles to pop and another 5 to become an adult, so the farm would have to be designed with that in mind. Otherwise if you're shaving off 4 cycles, you'd be dropping an age:1 fish instead of age:5 in for feeding and they aren't going to lay anything for ages. The math looks sound though.

I haven't extensively tested results of fish fighting over algae at x3. If it's not that bad my "Economies Of Scale" build could probably still function at x3 with a few tweaks taken from yours. Each door layer being held open for ~60 seconds is heavy-handed but made necessary in this case  

They are synced so that automation doesn't need to be duplicated. When first creating the system the pacu's ages might all be different, but after a generation or two they'd all be the same age, forced upon them by eggs being stored halting incubation progress no matter when the egg was laid.

Oh I see, you store them until storage is full and then you distribute them at the same time roughly.

Now I've always wondered about the viability bit.   Is that % chance it will survive or will any % = hatch.   0= raw egg always I figured.  But is there any chance  anything less than 100% = raw egg?

 

[Nxf7]

No chance @RonEmpire as far as I'm aware. It's basically just a 10 cycle timer at which point it cracks and drops raw egg. Someone please correct me if I'm wrong but I haven't observed otherwise and I was under impression it could have 1% viability left and be ejected from the storage, then once incubation hits 100% it'll still be a Pacu Fry.

[Saturnus]

54 minutes ago, Nxf7 said:

No chance @RonEmpire as far as I'm aware. It's basically just a 10 cycle timer at which point it cracks and drops raw egg. Someone please correct me if I'm wrong but I haven't observed otherwise and I was under impression it could have 1% viability left and be ejected from the storage, then once incubation hits 100% it'll still be a Pacu Fry.

Correct. Vitality and incubation rate are mutually exclusive. Eggs either lose vitality or they gain incubation rate. Not both.

[Nxf7]

4 hours ago, Nxf7 said:

12 hours ago, StalightStryker said:

 If we keep them breeding for three days we will use 300 calories and produce exactly 2 eggs. The fish is then dropped into the waiting pool where it lives the rest of its life 7 (140 calories at 20 calories/cycle due to glumness) + 10 (days until it actually starves) = 17 cycles. 17 x 7% = 119%. That would put that total at 319% per fish but with more fishes going through the active period, and should yield appreciably more eggs and meat per cycle for the same cost in algae. Must admit I have not tried the set-up thoroughly with these settings (reducing the hydro sensor to 18 should work) so it may not work properly. Use at your own risk.

Looks great! As for the above quoted, only issue with that is eggs are still going to take 5 cycles to pop and another 5 to become an adult, so the farm would have to be designed with that in mind. Otherwise if you're shaving off 4 cycles, you'd be dropping an age:1 fish instead of age:5 in for feeding and they aren't going to lay anything for ages. The math looks sound though.

Ok I did some napkin math and a "3.0 cycle" cycle is borderline not workable, but slightly longer would work and definitely be reliable like 3.5 or even 3.3. The issue is you want the recently dropped adult fish to be fed and benefiting asap, but the water still needs to be "pumped" back up there and any time that takes is time they aren't getting 66.6% repro and that's the time they lose once they on their last day (assuming a 3.0 cycle still) on top of an extra 60 seconds early for that set of doors to eject them; they will still have time left on their reproduction boost but they'll lose it instantly as soon as they get overcrowded in the retirement home (I like this name for it, I'm taking it).

So some more ****ty math: a 3.5 cycle would look like this:

 

Now that I think about it, probably don't have to worry about a starting offset with 3.5, likely won't matter at all if it's half a cycle older than it should be before it starts feeding.

[StalightStryker]

1 hour ago, Nxf7 said:

Now that I think about it, probably don't have to worry about a starting offset with 3.5, likely won't matter at all if it's half a cycle older than it should be before it starts feeding.

@Nxf7 nice find, I had not given much thought to the 3 cycle timer beyond the general concept and with the 3.5 timer we would lose only 3.5% of reproduction because it won't effect the active breeding cycle. Since we have the 19% buffer that would work out just fine. Come to think of it, following this logic even the 3 cycle timer would still work (cutting it close at 305%, perhaps a bit too close for comfort since the water refilling does cut on their active breeding time) but increase yield by 66.6% compared to a 42.8% gain in the 3.5 cycle timer. Both are very tempting (though I prefer the 3.5 for reliability reasons) and of course the increased spacial efficiency should also work wonders on your already compact Economies design. Though you may have to use that fancy new critter sensor (??? I'm not on the testing branch so I hadn't heard of it) or a water clock in place of your eggtimer.

The second major issue, like you noted, is the foodtimer, which would need it's own wiring designed with perhaps an extra hydro sensor/critter sensor (or looped off of the egg dropping timer, that could work?). But I'm sure that at this point neither of us would have a lot of trouble designing that. (I love how the designs keep improving every time something's posted though)

A third minor issue would be the extra incubation cubicle needed, but it is not that big a deal.

Though I'm confident we can all agree the best find of the day is the name 'retirement home' for the second pool. I can now pretend I'm not treating my fish horribly

 

[Saturnus]

You can help the process along greatly if you pump a single tile of water into each incubation chamber that will drop at the same time as the pacu. This will fill up the pools with water almost immediately so the breeding process can start much faster.

With 12 incubation chambers it will take a single pump 2 cycles to pump the water necessary.

[StalightStryker]

@Saturnus What a great idea, that would allow using the 3 cycle timer without any fear of the system falling apart. Thanks sir.

[Saturnus]

32 minutes ago, StalightStryker said:

@Saturnus What a great idea, that would allow using the 3 cycle timer without any fear of the system falling apart. Thanks sir.

Also note that while making the research for the original implementation of the omelette factory (see link) I noticed that Pacu life cycles are based on the cycle timer. A pacu hatched at any point during the cycle will have a counter set to 0/25 until the first dawn where it shifts to 1/25.

For example a Pacu hatched at 25% of a clock timer will be 0/25 until the clock timer is back at 0% where it switches to 1/25.

A pacu will actually live for exactly 25 cycles so the above pacu will die at 25/25 + 25% cycles but it becomes an adult when the clock timer is 0% and the pacu cycle reads 5/25 no matter at which point in the cycle it hatched so maturing to a fully grown pacu can actually be cut down to just a fraction over 4 cycles if incubation is always started just before the clock timer reaches 0%

 

[StalightStryker]

How very interesting, with that in mind, we could reduce the time they need to mature to 9 cycles + a fraction of a cycle, which would enable using only 3 incubation cubicles instead of four and they will be (almost) perfectly mature when dropped in. Thanks for the contributions, good finds Saturn.

[Saturnus]

Well. I just tested it out. It seems that now maturing is based on actual age from when they were hatched. So back to square one.

[StalightStryker]

18 hours ago, Saturnus said:

Well. I just tested it out. It seems that now maturing is based on actual age from when they were hatched. So back to square one.

Yeah, I had noticed that during testing, it was a bummer. But fear not! I've build a prototype that works (I must admit that it barely works, the fish drop their final egg about a minute before they die of old age in the 'retirement home', and I find it hard to change that. However, since they all have the exact same lifespan it should work for all of them).

So if any of you see some manageable change to fix that to a more reasonable amount of time, I would be most grateful.

I've abandoned the water clock in favour of a pure automation one, since we need to time 3 days exactly anyway. Do note: I'm not an automation expert. Everything I know I thought of myself and there probably are more efficient versions to be made.

Also note I've only build a single cell. You should in theory be able to add many more, but I've already spent a lot more time on this project than I intended to and this shows all it needs to.

Edit: @Saturnus Better?

[Saturnus]

@StalightStryker One of the images (the first one) doesn't show up.

[Nxf7]

The effort you go to label these amuses me. Even the NOT gate is labeled but I'm not sure why 

[Saturnus]

7 hours ago, StalightStryker said:

Edit: @Saturnus Better?

I'd still use a critter sensor as the basis of a simple toggle counter instead of a 3 cycle clock because that way you are certain that 2 eggs have dropped before you trigger the whole system regardless whether there have been an unforeseen delay somewhere in the system.

Something like this. The thing that look like a clock sensor is the critter sensor. The bottom left atmo sensor is a manual toggle. Every time it switches on it switches the output state. The atmo sensor at the arrow is a manual override that will keep the output low if turned on.

[StalightStryker]

@Nxf7 Yeah, about that Not-gate, I had drawn the lines before I wrote the text and then realised it was a not gate. The effort to go back and fill that spot up was a bit too much but a line with no text seemed silly. I wasn't too satisfied myself in the end.

1 hour ago, Nxf7 said:

The effort you go to label these amuses me

I just wanted people to be able to reproduce it if they liked it, even if they don't know much automation. Also, not sure whether this is compliment or not. 

@Saturnus Yes, that looks like a decent option if a few issues are resolved. But I'm not into the experimental releases and thus won't be testing that any time soon.

About that: wouldn't you need to hook up every single pool to it's own automation this way? The fish may take their merry time to fall or fall instantly and thus start their egg cycles on different times, creating a spread among the critters. A timer based off of that would surely tend to diverge over time if they don't all have their own automation. It's not undoable, but it certainly wouldn't add to the charm of the hatchery in my opinion.

But in all fairness my system is too close a call. I just thought of the fact that if you add more pools the retirement pool will get more pacu eggs laid in it over time, giving all pacu the expecting debuff for a few seconds, but quite often. That will probably be enough to offset it. It's gotta be possible to get this system working with the resources we have at the moment. Sigh.

[Saturnus]

5 minutes ago, StalightStryker said:

 

@Saturnus Yes, that looks like a decent option if a few issues are resolved. But I'm not into the experimental releases and thus won't be testing that any time soon.

About that: wouldn't you need to hook up every single pool to it's own automation this way? The fish may take their merry time to fall or fall instantly and thus start their egg cycles on different times, creating a spread among the critters. A timer based off of that would surely tend to diverge over time if they don't all have their own automation. It's not undoable, but it certainly wouldn't add to the charm of the hatchery in my opinion.

Here's a system that detects each individual cell. And only when all cells have dropped 2 eggs does the output become positive.

[ruhrohraggy]

Throwing up an updated version of this for use with the critter sensor. Thanks to SamLogan for the initial build, helped a bunch...Used it to help me build a more simplified / flexible version using the critter sensor.

 

Spoiler

 

Spoiler

 

 

Spoiler

 

I use 2x screen doors to make a small open room in the upper left corner with the critter sensor in it. The critter sensor enables the sweeper in the upper left to drop eggs beneath it when there are "xyz #" critters in the room, turns off once desired number is reached.

What's nice about this, is you can set how many fishies you want to use for breeding. You can easily choose now by adjusting the critter sensor, and you no longer have to worry about egg-shells messing with the weight-plate, so you can collect the egg-shells at your leisure. Set sensor for 0 critters to re-cycle 1 fishy, 1 critter to re-cycle 2 fishies etc...

Once desired fishy eggs are chilling in the upper left, it will then turn the upper left sweeper off and activate the one in the middle. This will deliver the eggies to the right...Very much the exact same as SamLogan's right-side setup.

I've also modified this fishy farm to extract the meat and the eggs automatically, and deliver it to the dupers at a spot of my choosing. (You can see 2x conveyor loaders on the right, 1 is for the eggs, other is for meat/shells)

Though...Looking at it, I'll probably need to add an airlock door under the left side, so the eggshells can drop down into the pool where the other sweeper can get 'em.

Updated with the door. Thinking about it further, It's probably possible to automate the door with a filter gate to activate a short time after the room empties out to dump the eggshells automatically...I'll have to work that in too...

Done...ezpz.  Should now open the trap door a short time after the fishies drop outta the room. Might not be great for more than one fish though...especially if the eggs aren't hatching all at once...But this might stabilize if you have more than 1 fish going at once...Meh, oh well, will just recycle them back up top again anyway...Will watch it for a bit and see how it goes.

The automated trap-door for egg-shells did not work out as well as I had initially hoped...But i'm tired...I'll give it some thought after some sleep. Everything else seems to be doing it's job though.

 

Once everything is setup, you can seal off the right side of the room. I setup the ladder on the right above the Storage Container so the dupers can still deliver coal until I sort out central storage to do it automatically. The desired goods are delivered outside of the room.

I just kinda winged it in a live playthrough, we'll see if it's reliable or not. (I'm thinking it will be based on initial impressions).