The Water Duplicator - A Homemade Water Geyser Providing Almost 20kg/s at 3C

[mathmanican]

A special thanks goes out to @M.C. for starting the "Cooking dirt without digging" thread and @biopon for the bug post "Fertilizer doubling mass when it turns into dirt at 125C ". Also, thanks to @R9MX4, @nakomaru, @Oozinator, and @Neotuck for their comments and fun discussion that spurred on this project. 

I now present the Water Duplicator 23.7 (way to many different versions - finally got this one). Liquid, Power, Automation, and Shipping in spoiler. 

Spoiler

• 
• 
• 
• 

 

What are you looking at.

• This thing will produce just shy of 20kg/s cool water at about 3C (plus or minus a degree). The only reason the cap is 20kg/s is that this is the speed of the conveyor. The rate could easily be made to be 40kg/s, with some modifications that include 8 pumps and 4 aquatuners with crude (or 2 with super coolant).
• The exterior pumps on both sides take water near 1C to the top region. Aquatuner cooled oil quickly freezes the water to ice, which then gets stuffed in the conveyor loader.
• The rails, coupled with tempshift plates and solid doors on the bottom of the build, rapidly raise the temp of the ice and it melts while on rails. This is the fastest way I know of to melt ice (all other attempted versions were drastically slower, probably by a factor of 25 times slower - thanks @Yothiel for your temperature post). 
• Almost every time the ice melts, the mass doubles (the key bug). I am currently seeing >90% doubling rate, some cycles at 100%.  Sometimes doubling does not occur (not sure why).  I have noticed that if the ice melts one tile further left (so right next to the wall), then the entire mass is lost. The conveyor loader should not extend further right or left for this reason.  I think that the doubling must occur on the bottom row of liquid, as when I had conveyor receivers on the top row of water, I would not get 100% doubling. 
• Sometimes the oil cools the ice too much, which leaves a chunk of ice on the rails too long, reducing the 20kg/s rate a tad. Shouldn't be a problem though, as this is more liquid than any map seed will give, period. @babba, if you are still around, I think you'll like the ability to increase your water flow.
• The liquid temp sensors on the aquatuners are currently at -14C for the crude oil.  If the oil gets too cold, then the ice freezes too cold and takes too long to melt on the rails.  If the oil gets too warm, then it takes too long to freeze the water.  Both cause a slow down. Somewhere between -15 and -10 is an appropriate temp. More testing needed. Clearly super coolant could reduce the power requirements, but that requires going to space first. This can be done without. With supercoolant in your pipes, you can do this exact same build with crude oil, instead of water, to create crude oil at 20kg/s. 
• Because the water duplicates at negative temps, the tepidizer is needed. Even with full time running aquatuners, the temp of the entire bottom area will continue to drop. The tepidizer is connected to a 3C sensor, which keeps the water temp hot enough to melt the ice. At 2C, the ice takes too long to melt one run through the bath. Increasing the bath length (making this wider) would allow you to drop the temp to 2C, or even 1C. Increasing the temp would guarantee faster melting, but also increase freeze time (which could be offset by a lower temp on the aquatuner). 
• The hydrosensors control the liquid pumps that keep this thing running. If the water level is too high in the top region (above 50kg), or there is too much water in the bottom region (above 2000kg per tile), then the pumps shut off. This prevents the game from crashing because you have too much liquid pressure (this thing will get out of hand really fast if you don't pay attention to it.  After 100 cycles you could have 100000+kg in a tile, and at some point I know the game crashes). The hydrosensor in the top stops the pumps from adding too much water up top, should your cooling not be enough.
• There are two pumps in the bottom to extract the water at 20kg/s, if you can use that much, lol. 

Water is no longer a limited resource. If you want 600 dupes in your game, just build a few of these, and use @Lilalaunekuh's berry farm (see this post) for automated harvesting. Let your dupes play on the arcade and dance all day long. 

Here's a few other builds I played with along the way (versions 1-23.6, only a few shown).

Spoiler

First attempt was to avoid the pump entirely, and have the water push itself up with an escher waterfall. The aquatuner needs to be in the pool where the water comes from, otherwise the tepidizer just adds too much heat. The doors that froze the ice are long gone (went through several iterations)

The beast below original started from the one above, with just the aquatuner inside the bath.  Eventually I learned that separating the water into chunks to freeze at once saved power and sped up freezing. This thing has been running for a long time, and water collection room at the bottom has cells with over 90000kg of water. In this build I had to keep the water around 7C. Eventually I got things to melt just as the next batch of ice would drop. I had about a 3-4kg/s rate of water increase. Occasionally a batch of ice would build up over several cycles and become a 15ton chunk of ice.  Once it melted, I was glad the escher waterfall was there to slurp it all up almost instantly. 

I started playing with the automatic icemaker feature of the pitcher pump, thanks to the reminder from this thread. Fun times. It was a disaster at first.  

This beast can generate 5kg/s water, for very little power. It was too big to stop, but I thought I had it down. The top row of doors below the 860kg crude oil are on a 40/6 pulse (closed 40, open 6), which is just enough time to constantly reset the pitcher pump bug and enable cooling at full speed. At this point I decided there was no need for the escher waterfall, and started new builds to reduce the size. 

Here's one version of my melter. Worked pretty nice, but melting was still slow. 

While playing with just melting zones, I found that conveyors melt ice like crazy.  If the ice is in a solid tile, then the heat transfer is SUPER FAST. I could finally melt stuff at a rate of 20kg/s.  Water Duplicator 23.0 is shown below. 

A baby <10kg/s version is located later in this thread. Here's the basic idea.

Here's @Saturnus's "mini" version (averages about 8kg/s), and avoids over pressure issues. Click here to jump straight to the details for this one. 

[MRE_]

I figured this out months ago while messing with a regolith melting design, planned on a huge post about it, but never got around to it. Made a 20kg/s Igneous rock or refined metal duplicator that produced more than 5 unblocked steam turbines worth of heat when cooling the material down to 230 degrees celcius, and took nothing but one steel craft a day to keep running, with the 82 stone hatches needed to eat that much more than supplying enough coal and eggshells for.

[mathmanican]

25 minutes ago, MRE_ said:

planned on a huge post about it, but never got around to it

Would have been a fun read.   

[Lilalaunekuh]

First nice thread

I love it when someone takes a "bug" and makes something totally overengineering out of it^^

 

4 hours ago, mathmanican said:

If you want 600 dupes in your game, just build a few of these, and use @Lilalaunekuh's berry farm (see this post) for automated harvesting.

Thank you, that you like my farm. At least I am a bit a shamed of this "old" version of my farm.

48.000kcal a day with 8kg/s of water seems good till you realize you need no water while the plant is withering ...

Version 2.0 uses a water reservoir to feed each row of plants (in my case 40, but you could go for up to 41 or 42 if you use some additional pipes to buffer the water.)

48.000kcal a day with 5kg/s of water is the current production (still using 200kg/cycle of water more than I "really" need).

=> 5t of water / 1 reservoir is able to supply enough water to fully  grow 40 (up to 41) bristle berry plants

 

Spoiler

If I reach version 3.0 and solve my last problem I will make a thread.^^

Just building an automation based system to restart my farm based on it´s output. Withering got 2 stages of 2 cycles and each stage can reset on a reload.

=> It can take more than 4 cycles for an autoharvest to happen, so normal clocks are no viable solution.

 

Right now I am tinkering with a small room that will take in the bristle berries and seeds and will reset my farm after receiving 40 bristle berries before shipping them out to my kitchen.

(That would be an additional auto sweeper, 2 additional conveyor loaders and one refrigerator attached to each of the 6 rows ...

So now I need some advanced automation to keep these additonal buildings powered using my conductive wire^^

For now I have ~8kw attached to my wire and use 5 time slices to prevent any overload, but 11kw could slow things down.

If the autosweeper start to work to late the reset of my farm will be to late, resulting in less production.

And I am not sure if I want so spend my engergy on 6 always running refrigerator, just as a sensor^^)

 

Or is there a better solution I can´t think of ?

 

[Oozinator]

[badgamer123]

7 hours ago, MRE_ said:

I figured this out months ago while messing with a regolith melting design, planned on a huge post about it, but never got around to it. Made a 20kg/s Igneous rock or refined metal duplicator that produced more than 5 unblocked steam turbines worth of heat when cooling the material down to 230 degrees celcius, and took nothing but one steel craft a day to keep running, with the 82 stone hatches needed to eat that much more than supplying enough coal and eggshells for.

darn it it seem I finally know why my regolith melting turn out to be 100000kg of igneous rock....I tho I afk too long lol


back to topic....amazing .....could u double anything with this?i think someone will make a Nat duplicator soon.

[Saturnus]

6 hours ago, Lilalaunekuh said:

For now I have ~8kw attached to my wire and use 5 time slices to prevent any overload, but 11kw could slow things down.

And I am not sure if I want so spend my engergy on 6 always running refrigerator, just as a sensor^^)

For both of these there is another option.

You really only have to turn on the sweepers and fridge 4 cycles after the reservoir has run out of water. You can have a priority system where each row have decreasing priority. While this might delay some resets, it will also ensure a continuous flow, and the delay is likely much shorter than the waits involved in time slicing.

Specifically for the fridge it certainly doesn't even need to be on all the time while the sweepers are active. Powering it on for 2 seconds every 20 seconds should be plenty. So it's on 10% of the time to check if the system is ready to be reset but only after 10 cycles of being off, like the sweepers.

[Lilalaunekuh]

2 hours ago, Saturnus said:

4 cycles after the reservoir has run out of water.

That´s one of the problems. If I reload while my plants are withering, a normal 4 cycle timer will fail.

 

2 hours ago, Saturnus said:

You can have a priority system where each row have decreasing priority. While this might delay some resets, it will also ensure a continuous flow, and the delay is likely much shorter than the waits involved in time slicing.

Right now my time slicing is just used to run my autosweepers/loaders inside the farm.

(My time slices guarantee that each row has at least one active sweeper, so worst case it´s the one with the longest conveyor rail.^^ But since I loose no resources while waiting it´s ok. Lights are off and no water is beeing consumed.)

With some priority system on my columns (not rows^^) I would be able to achieve a samller delay.

=> If I keep the sweeper with the shortest conveyor rail always on and just use timeslicing for the rest, my delay will always be the smallest (I can achieve using this reset method) without much change to my existing design.

 

2 hours ago, Saturnus said:

Specifically for the fridge it certainly doesn't even need to be on all the time while the sweepers are active. Powering it on for 2 seconds every 20 seconds should be plenty. So it's on 10% of the time to check if the system is ready to be reset but only after 10 cycles of being off, like the sweepers.

My farm was optimized without water free auto-harvest in mind^^

=> My current "reset" system is in the spoiler below:

Spoiler

The system above had worked (using a full time working fridge ...).

I need no automation on my sweeper/loaders in this room since I know my auto-sweepers/loaders inside my farm will be finished.

Just the fridge is my "problem".

=> Like @Saturnus said there are good ways to cut the power usage by a huge margin

 

But I am still not 100% sure if that´s the best solution to reset my farm.

(A single smart battery attached to each row of auto-sweepers and conveyor loaders could yield much better results. If there is no other consumer on the line my normal automation to trigger a transformer/power generator could be used for the reset.)

Or do you know an other maybe even better solution ?

[Saturnus]

Just now, Lilalaunekuh said:

That´s one of the problems. If I reload while my plants are withering, a normal 4 cycle timer will fail.

I don't see the problem at all. The 4 cycle timer just gives the point which the harvest can start. Not when it has to start. Basically when the 4 cycle timer elapses the sweepers are ready if no other sections has already reported ready. When the fridge is filled it is reset. As I said. I don't see the problem.

6 minutes ago, Lilalaunekuh said:

My time slices guarantee that each row has at least one active sweeper

Completely unnecessary. Sweepers do not have to be on at all for 10 cycles after the system is reset and growth cycle start. It doesn't even need to check. You potentially waste a ton of time doing time slicing.

At the end of the day though. It doesn't really matter if you optimize to slice off a few watts of total power consumption or make things slightly more compact. Up until now the important thing was to minimize water consumption. With the thread, even that is irrelevant. Need more water? Just build more 10kg/s water geyser blocks.

[mathmanican]

4 hours ago, badgamer123 said:

back to topic....amazing .....could u double anything with this?i think someone will make a Nat duplicator soon.

Yep.  As long as the liquid can be solidified easily, by some thing else, this should work just fine.  Use super coolant (or maybe LOX - possible danger zone) for crude/petro/naptha, in the aquatuner loop. 

[Saturnus]

2 minutes ago, mathmanican said:

Yep.  As long as the liquid can be solidified easily, by some thing else, this should work just fine.  Use super coolant (or maybe LOX - possible danger zone) for crude/petro/naptha, in the aquatuner loop. 

I think you can safely use polluted water in the aquatuner loop as you don't as low temps due to higher SHC.

[Lilalaunekuh]

14 minutes ago, Saturnus said:

32 minutes ago, Lilalaunekuh said:

That´s one of the problems. If I reload while my plants are withering, a normal 4 cycle timer will fail.

I don't see the problem at all. The 4 cycle timer just gives the point which the harvest can start. Not when it has to start. Basically when the 4 cycle timer elapses the sweepers are ready if no other sections has already reported ready. When the fridge is filled it is reset. As I said. I don't see the problem.

Point for you (My thought but not my text was about using a 4 cycle timer JUST to do the reset, for everything else it´s ok.)

But since my systems power consumption has grown, why not just use some heavy watt wire ?^^

(There is no downside in keeping my autosweepers/conveyor loaders active all the time.)

 

14 minutes ago, Saturnus said:

32 minutes ago, Lilalaunekuh said:

My time slices guarantee that each row has at least one active sweeper

Completely unnecessary. Sweepers do not have to be on at all for 10 cycles after the system is reset and growth cycle start. It doesn't even need to check. You potentially waste a ton of time doing time slicing.

Completley unnecessary if I would just use heavy watt wire

But where is the problem if an autosweeper without something to do is active or not ?

(I only need one sweeper delivering one bristle berry to trigger a reset, so the only time I lose time would be when the furthest away sweeper is active while the plants auto-harvest.)

 

14 minutes ago, Saturnus said:

It doesn't really matter if you optimize to slice off a few watts of total power consumption or make things slightly more compact.

Like I said the fridge is the only thing to optimize. (Auto-sweepers have always the same worktime to do the fixed amount of work.)

My current base was using 100% uptime fridges just because I have a huge power surplus and no need to save power.

But it´s a game and my goal is optimization

 

PS: sry for hijacking this thread

[Saturnus]

2 minutes ago, Lilalaunekuh said:

Like I said the fridge is the only thing to optimize. (Auto-sweepers have always the same worktime to do the fixed amount of work.)

Then just cycle it on and off. Being on just 5-10% of the time should be plenty. You do it with a power switch.

[Lilalaunekuh]

Just now, Saturnus said:

8 minutes ago, Lilalaunekuh said:

Like I said the fridge is the only thing to optimize. (Auto-sweepers have always the same worktime to do the fixed amount of work.)

Then just cycle it on and off. Being on just 5-10% of the time should be plenty. You do it with a power switch.

My current thought is just if a power sensitive system would be superior.

If an autosweeper inside my farm starts working I am able to reset the water supply.

(No "reset"/sorting room, no additional power consumption besides the low daily power drain and I could still run everything on conductive wires.)

 

The version with the fridges works, but having 2 auto-sweeper that need to move each berry seems like a bad design.

=> Thank you for your input  and hope to hear from you when I start my own thread on this topic.

(Will start my thread when I am satisfied with my farm.)

[mathmanican]

14 minutes ago, Saturnus said:

I think you can safely use polluted water in the aquatuner loop as you don't as low temps due to higher SHC.

Original designs used polluted water and had to melt 7c water. I could not melt stuff fast enough. Now with a 3c bath i will try pw again. That was my original plan. I got too excited when I found this was possible.

I also have another optimization planned, namely in the  freezing area have the doors above with airflow below. This will prevent ice from getting colder. Version 24 and 25 coming soon.

34 minutes ago, Lilalaunekuh said:

PS: sry for hijacking this thread

No worries. Keep up the discussion. I need a way to use all this water.

[Saturnus]

6 minutes ago, mathmanican said:

Original designs used polluted water and had to melt 7c water. I could not melt stuff fast enough. Now with a 3c bath i will try pw again. That was my original plan. I got too excited when I found this was possible.

I also have another optimization planned, namely in the  freezing area have the doors above with airflow below. This will prevent ice from getting colder. Version 24 and 25 coming soon.

You also don't need it to be doors at all in the freezing area. Regular metal tiles should work fine. Especially when above the water.

If I was to give a suggestion it would be to half the output to just a compact 10kg/s cube. Very few people would realistically need more than that. And then they could just build more than one.

[mathmanican]

3 hours ago, Saturnus said:

If I was to give a suggestion it would be to half the output to just a compact 10kg/s cube

How about this?

Spoiler

This thing does not have an auto shutoff to prevent it from making enough liquid to crash the game, but since it is only half as big, you would have to ignore it for twice as long, or much longer, before that would happen.  Just turn it off when you have 50000kg per tile, and then turn it back on much later (or remove one airflow tile, lose the square shape, and put in protection). 

Some comments. 

• Hydro sensor currently set at 200kg.  You don't want to exceed 500kg, else the ice freezes into blocks.  If you exceed 200, then the system cannot produce enough cold to freeze the water. So far I've watched my water level hover between 40kg and 110kg. Never once has it gone up to 200kg (so the pump is running full time). 
• I currently have the liquid sensor on the aquatuner at -8. The PW will drop to -22C, but the system produces -20.6 to -20.8C water pretty consistently (haven't seen it once drop below -21).  
• With oil, I would need to spend more energy on power, but I could easily keep the water level down and maintain a 10kg/s inflow of water.  The sweeper arm runs on/off, but no worries as it's fine if ice is melted at 5kg/s or 15kg/s, as the average melt rate will be 10kg/s. 

To make this contraption, the biggest wait is a metronics engineer and some research. Pretty simple. 

[Saturnus]

11 minutes ago, mathmanican said:

How about this?

Very similar to my own experiment. I just left in room for hydro sensors. I'm not getting 10kg/s consistently though. I'll tweak it a bit more.

[mathmanican]

9 minutes ago, Saturnus said:

'm not getting 10kg/s consistently though

It's almost always a little less. I have not tracked down what makes the matter replicate or not. I do see that your conveyors go all the way to the right most water square. I also see them in the middle of your tank.

•  When mine extended to the left most water square, the ice would melt and disappear (complete mass loss).  I wonder if that happens on the right side as well. 
• When I melt stuff in the middle of the liquid bath, not on the bottom row of water, I also see a much lower rate of duplication.  The door row, and bottom layer of water, seem to be the spots where duplication occurs with highest frequency (anecdotal evidence so far - I haven't conducted thorough tests, but enough to make a conjecture). 

Glad that I made something you wanted to tweak.  I look forward to seeing your improvements. This is reminiscent of your Borg Cube.

[Saturnus]

2 minutes ago, mathmanican said:

I do see that your conveyors go all the way to the right most water square.

They don't though. I avoid the left- and right-most water tiles as suggested.

[mathmanican]

1 minute ago, Saturnus said:

They don't though

Face palm.  Oops.  Just looked again. You are right. \

I do wonder if the middle of that bath, rather than bottom row, plays a part.  At this point, the fact that we can make a water duplicator is what is absurd. Optimizing it is now just sheer fun and joy. 

With the next patch notes, I'll put in a comment to @Ipsquiggle, unless he wants to put a "thank you" in this post somewhere, just to make sure he sees the problem.  He's been pretty great recently about putting a thankyou on bugs/exploits like this.  They know the bugs are there (even if they haven't updated the bug tracker. 

[Saturnus]

Played with it a bit more. Compacted it and made it much more consistent by introducing counter-current cooling flow. Now it only freezes in the two right most tiles in the upper chamber where it picked up immediately as the sweeper doesn't have to swing around at all anymore. Water will fill the gap left behind but is cooled down on the way freezing very fast.

One thing however.  In order to compact it I did need to get rid of the overpressurized water reservoir so now the bilge pump will start at 1100kg. It's only averaging 8kg/s water at about 4C though but for the compactness I'm ok with that. And it does seem like it's picking up slightly over time.

[mathmanican]

5 minutes ago, Saturnus said:

Compacted it and made it much more consistent by introducing counter-current cooling flow.

It's a thing of beauty. I'm excited to see even more improvements. 

[MRE_]

On 3/10/2019 at 4:55 AM, badgamer123 said:

darn it it seem I finally know why my regolith melting turn out to be 100000kg of igneous rock....I tho I afk too long lol


back to topic....amazing .....could u double anything with this?i think someone will make a Nat duplicator soon.

You can duplicate anything that freezes into the same thing it melts into, eg Igneous rock, water and refined metal work, but trying to duplicate granite would give you Igneous rock instead, as granite melts into magma, which in turn freezes into Igneous rock.

 

I'll make the post tomorrow I suppose

[iAbaddon]

Thank you! I was quite concerned about the liquid oxygen mass in my frost room because it appears again and again and i can't find the reason where it from.