Pumping gases without a pump

[tzionut]

Today when i tap the bottom hot area for my last patch of crude oil i accidentally find this:

Using the same principle as @SamLogan liquid pump using 2 gases, you can pump gases up using the dripping effect of liquids. I attached the save file. It can be used for future designs of steam turbines using condensation and this dripping effect for pumping steam under the steam turbine whit a power free gas pump.

Best regards.

Post edit:

Is possible that the steam goes up because of the higher pressure. I must get ready for work. I will test it more when i am at home. 

Post . Post edit:

As you can see i put chlorine down, and hydrogen up. If you let only one liquid (water) drip on the wall nothing happen. If you put the second liquid(petroleum, oil, naptha, visco gel) on the dripping tile, the void created push the chlorine gas up. This setup pump bottom gases up. Useless for steam I think. 

I will make another setup and try again for the steam engine

Post post post edit (too much editing):

Muhahahaha (evil laugh)  The SSST is created. Have fun.

 

Utopia.sav

[mathmanican]

On 3/11/2019 at 8:22 PM, tzionut said:

Using the same principle as @SamLogan liquid pump using 2 gases, you can pump gases up using the dripping effect of liquids.

This my friend is quite awesome. I've been playing with tests for the last two hours. Any two liquids can be used. A few rules are in place.  I can make a pump that pumps much faster than 10kg/s with this trick. It's quite insane.

Has anyone read or seen other posts on this topic? If have, please leave a comment.

I've been hunting through the archives, and using google, but I don't know what key words people are using.  If anyone knows of post related to this topic, using liquid on a wall to pump gas upwards, please add a comment.  Here's some screenshots with explanations if you are interest in learning more. Lots of spoilers, and time for massive exploits, if you open this spoiler. 

Spoiler

In the above 4 pictures, notice the bead of water on every other tile of the walls.

• The first two pictures use polluted water and water.
• The second picture has the cross hairs from debug to verify that no matter is lost.
• The third picture shows that without a second liquid, the beads do not form. You must have a second liquid, and then a wall to cling to at the same height as the second liquid, for the bead to form.
• The last picture use crude oil and petroleum.  @tzionut These two can be used in the SSST, so you don't need supercoolant.  You can use a steel liquid pump and avoid space age materials, provided you keep your heat sources low enough. 

Q & A

• Do you need both sides to be walls?
  • Nope.  All you have to do is form the bead. Once it is formed, it doesn't disappear, even if you can't see it. In the picture below, we need one tile under the crude to hold the liquid. We need the tile to the right of the petro to form the bead of liquid. The tile to the right of the crude is needed to allow pumping, otherwise the gas will just move right when it hits that tile. Once you stop either the right wall, or the left wall, you cannot start it again without destroying the bead. So you can stop the pump action on purpose by leaving a gap on either side of one tile. 
    
• How many tiles do you need to make the pump?
  • Not much.  You need air above the liquid. The liquid pump can dump it's contents directly on the tile. Then you need one tile below the tile holding the crude for air to enter the pump.  This is probably a minimal setup. As a side note, with 800g/s petro, I pumped the room to the right to vacuum. With water, you'll only need 80g/s. 
    
• How much do you set the liquid valves to?  
  • I think I'll be making a post about this soon.  Each liquid has a type of viscosity. You can measure the viscosity by droping the liquid on single tile and letting the excess run off.  For water, this leaves you with around 35-38g. Double that and round up a tad, and you get the needed liquid to make this pump work without backflow. Stay close to, but under this doubling, and the matter converter works. More to come when I'm done testing.  For crude and petro, they both naturally want to have about 360g (plus or minus 20) when on a tile alone. So doubling this and adding up gives 800g.  At 700g, backflow occurs quite often. Over 800g, and I've only seen once in 10 cycle a few mcg slip past.  This means I need to go higher than 800g a bit.  You can run the liquid pump at full 10000g (waste of power), but you won't see any slippage.  
• How fast does the pump work?
  • Depends on how many beads you have falling. If you really want to utilize the pump for high pressure, then make sure you liquid outlet is in the liquid (otherwise your gas pressure WILL overpressurize the liquid vent, eventually). Each bead causes a percentage of gas to move upwards. Leave both sides open, and more gas gets gobbled up.  It's a percentage game, so increasing the beads causes an exponential increase in the pump speed.  It's quite amazing.  A minimum size pump like the one above took several cycles to pump to vacuum. The open on both sides build that has 6 air tiles below took 500kg pressure (20 tiles, so 10000kg) to 200kg pressure in seconds, and pumped the room to vacuum in less than half a cycle.  INSANE!!!!!!!  Take that you steam turbine with your measley 10kg/s.   

There is way more to learn.  Here's a few more pics for fun, adding on to the SSST. (I'm using heated neutronium to simulate a heat source, so I can ignore the temperature game.)

This one is slow in pumping, but if you increase the gas pressure inside, eventually the percentage game will win and guarantee that the top always has a lower pressure than under the turbine. 

No reason to stop at one.  I think you can stack AS MANY AS YOU WANT horizontally. All you have to do is add enough liquid to make sure the percentage game wins (might be quite a lot of liquid - use a homemade geyser). Next time I build this, I'll make sure the pump sits in the middle of all the turbines, and is open on both sides for maximum pumping power (rather than running down a wall). 

I'll be exploring fun uses of this over the next week or so.  CRAZY!!!!

 

Here's a save file (cycle 9, with lots to play with). 

Viscosity.sav

 

[OxCD]

This does seem crazy. Knowing those many interesting cases study you have already done on steam turbine, I cannot wait to see what you're going to explore with it especially. That could bring a new dimension to its potential. However, careful, if your discovery and set-ups are too much exploit, we can consider a patch from the devs that could change flow mechanics ahah.

[Angpaur]

10 minutes ago, OxCD said:

However, careful, if your discovery and set-ups are too much exploit, we can consider a patch from the devs that could change flow mechanics ahah

I think that this is the exact point of doing this stuff. So such big exploits are patched sooner than later

[Yunru]

50 minutes ago, mathmanican said:

Has anyone read or seen other posts on this topic? If have, please leave a comment.

Yes, it was quite a known thing, was even referenced recently in another thread (by you even!). But damn does searching suck.

The first I'd seen of it was someone posting an Escher waterfall as an alternative to door pumps.

[tzionut]

Quote

These two can be used in the SSST, so you don't need supercoolant.  

First i use petroleum and crude oil but i mess whit the lava temp and the wall start melting, and oil converted to petroleum. So i changed a little. My only purpose is to see if is possible and power efficient to move sufficient steam in this way for powering the steam turbine. Your Battery transformer power loop is way more efficient in create power. But as i said. I see this dripping effect by mistake when i tapped a sealed oil area, and remember the cascade effect. So i tried it in sandbox mode to see if i can recreate it, and have fun whit it. If is fixed or not, klei will decide. 

[mathmanican]

3 hours ago, Yunru said:

Yes, it was quite a known thing, was even referenced recently in another thread (by you even!). But damn does searching suck.

The first I'd seen of it was someone posting an Escher waterfall as an alternative to door pumps.

When I first read this post, I thought it was like an escher waterfall. It's not. It has a very different mechanic. It's similar, but also different.

I'm guessing @FIXBUGFIXBUGFIX (apparently R9MX4's new name - didn't know you could change your screen name) has seen this and worked with it.  I'm also guessing that my recent exploit posts have taken his frustration to a new level.... 

3 hours ago, Angpaur said:

I think that this is the exact point of doing this stuff.

Bingo. Some of the bugs are super well known, but the extreme uses of these bugs may not have been explored.  So I'm working on marketing to hopefully get them more attention. The actual fix to some of them should be quite simple, while others are more engrained, and just a side effect of one-element-per tile mechanics. This particular exploit, with beads of water rolling down walls or hovering in mid air, may or may not have an easy fix. 

59 minutes ago, tzionut said:

First i use petroleum and crude oil but i mess whit the lava temp and the wall start melting, and oil converted to petroleum. So i changed a little. My only purpose is to see if is possible and power efficient to move sufficient steam in this way for powering the steam turbine.

That's what I guessed. Your post was great.  Thanks for putting it out there.  

[Yunru]

1 hour ago, mathmanican said:

When I first read this post, I thought it was like an escher waterfall. It's not. It has a very different mechanic. It's similar, but also different.

Oh no, it used this too. It just used an Escher Waterfall rather than a pump.

So something like this:

 

@mathmanican not sure if you caught my edit.

[Yunru]

Refined it a little (cutting out the Steam Turbine part):

Before:

Spoiler

After:

Spoiler

(As you can see, I didn't quite use enough petrol. Stupid escher waterfalls being finicky. @mathmanican how do you do it?)

 

EDIT: Okay, the bottom left waterfall has to be two tiles lower than your gas input.

[tzionut]

@Yunru  i think your problem is the quantity of liquid falling. I try to recreate your design.

[Yunru]

5 minutes ago, tzionut said:

@Yunru  i think your problem is the quantity of liquid falling. I try to recreate your design.

My original one which I just C+P'd and didn't try to refine works like a charm, if you want to try the whole thing without tinkering with the waterfall.

[tzionut]

As you can see is work and power free using water and the cascade . As @mathmanican said is all about the quantity of liquid that fall. I used water and crude oil for the cascade and oxygen and chlorine for the cascade. Is a little redesigned. I save the setup for you to try and see.

Spoiler

Post edit. I think is more power efficient using the escher waterfalls and the drip effect for pumping gases (no power used)  

Rocky Panopticon.sav

[Zarquan]

This seems to work like a real world Sprengel Pump, but backward.  A Sprengel pump would slowly vacuum the top chamber and pressurize the lower chamber.

[Yunru]

Might not want to be using water to move steam.

[mathmanican]

8 hours ago, Yunru said:

(As you can see, I didn't quite use enough petrol. Stupid escher waterfalls being finicky. @mathmanican how do you do it?)

I saw your example earlier today and thought, "Perfect.  The example you gave illustrates 3 principles perfectly, and shows the similarities and differences." I'll use this to help show how this new water pump mechanic is actually new.  @tzionut gave a great fix above. I'd like to compare and contrast the principles, as it illustrates the how/why each principle is used, which is important for making use of this cool feature/bug/exploit/awesome sauce. 

Here's your picture that I'll refer to. If you want the summary, then just read the "Key Points" in bold below. Otherwise, grab some popcorn and enjoy the read. 

1. Escher Waterfall - Trap two gasses at the ledge of where a liquid will fall. The liquid falls without checking any pressure requirements. The liquid basically "disappears" from the game momentarily, and then reappears at the bottom of the fall. It's teleportation of liquid (or the liquid vanishes to a different dimension while falling). Note the liquid does not interact with surroundings while falling. I currently use this to collect all water in swamp biomes.  I build a single water collection station at the bottom of the swamp, and then just dig out the whole biome.  All the water ends up in one tiny spot, ready to pump where I want.  Simple water storage. 
   
   Key point - An Escher waterfall uses gas and the "one element per tile rule" to force liquids to travel via teleportation. You can't click on the liquids as they fall, as they don't exist in game at that moment.
    
2. (Liquid pressure mechanics) Pressures equalize on the same row, and gain a fixed amount of max pressure based on depth.  When liquid bodies of the same type touch, then liquids at the same height should have the same pressure. The speed at which these pressures try to equalize is indirectly proportional to the viscosity (how much liquid wants to sit alone on a single tile if everything else is allowed to fall off).  Water and polluted water have a very low viscosity, so they try to equalize very rapidly.  Liquid metals have an extremely high viscosity, so they don't try to equalize pressure very rapidly at all, in fact they won't even cause any change without passing a rather high threshold. 
   
   In your example above, the petro at the bottom of the escher waterfall (leftmost waterfall) is gaining in pressure every time liquid teleports to the bottom.  Since this tile gets high pressure, the tiles equalize at the same level.  However, then the tiles above notice that the lower tiles have really high pressure, so the petro moves up. As the petro moves up and reaches the top, then the tiles at the same level try to equalize, causing the liquid to move right. It eventually hits the ledge and falls down another waterfall, NOT an escher waterfall. Instead, the new liquid bead pump.
   
   Key Point - Liquid pressure mechanics is what causes liquids to rise, as the escher waterfall creates a high pressure imbalance in the lower tiles.
    
3. The new thing - How about we call it the "Liquid Bead Pump"?  (We'll use the bead definition "a small ball-shaped body", as in a bead of sweat). The liquid bead pump forms a bead of liquid to the right of the petro (because that's how two liquids interact). The liquid did not vanish, rather stays as an entity and interacts with the environment the entire path downwards (until a wall ends, and then starts again). Because liquid is heavier than water, and the liquid has not vanished (entered the hidden ONI dimension) it swaps spots with the gas below (tile swapping rules). As long as the beads form fast enough to create a "Bead - gas - bead - gas - bead -gas" pattern, the pump will form a perfect vacuum eventually.  This is all directly related to the liquid's viscosity rating (not sure if this is the best term, but I'm happy with it for now). With liquid metals, the amount of liquid needed to make this pump will be >150kg/s, so not very feasible to pump gaseous gold this way. However, with water all you need is about 80g/s, and with crude or petro a little more than 800g/s is enough. 
   
   Key Point - The "liquid bead pump" uses water mechanics to make beads, and then "tile swapping rules applied to one-element-per tile" to create the upwards gas pump. 

In summary, (1) escher waterfalls use liquid teleportation to work, (2) liquid pressure mechanics cause high pressure liquids to rise, and (3) the bead pump causes liquids to form beads (instead of teleporting) so that the liquid can interact with it's environment on the downward descent.   

I believe this pump will make a pretty sweat gas storage facility for all gasses (prototype in progress), will make very slick turbine stacking builds (another prototype), and much more. I've heard people talk about how germy water that falls into a pond will infect every tile the water passes through (so teleporting liquids can leave behind a nasty trail of germs till they hit the bottom of the pond - but you can't click on the liquid...), however I bet this liquid bead pump will NOT leave behind a trail of germs but rather stop the second it hits the TOP of the pond, instead of the bottom (gotta test this).

I've got plans, and will post as I have time, probably starting a new thread titled, "the liquid bead pump" if that name sticks. We also need a thread on "liquid viscosity and its effects".

Too much to do. So little time. ONI rocks.

[Craigjw]

You can also use a Wheeze to pump gasses, as it's input is the lower tile and it's output is the upper tile.  There is a post describing this effect, but I can't remember it off hand.

[mathmanican]

4 hours ago, Zarquan said:

This seems to work like a real world Sprengel Pump, but backward.

Thanks for the reference.  That was fun to read about. 

Just now, Craigjw said:

You can also use a Wheeze to pump gasses, as it's input is the lower tile and it's output is the upper tile.  There is a post describing this effect, but I can't remember it off hand.

Thanks for the post.  Yep, I know about this one.  It's a different principle than this liquid bead pump, but definitely related to pumping gas upwards.  Again, thanks for the post. 

[Yunru]

A weird thing I found was that (at least the working smaller version I eventually built) it only works when falling right. When falling left the petrol simply refuses to bead.

[mathmanican]

On 3/13/2019 at 5:54 PM, Yunru said:

A weird thing I found was that (at least the working smaller version I eventually built) it only works when falling right. When falling left the petrol simply refuses to bead.

Check my steam turbine builds above (in the spoiler).  I had it working perfectly fine on both sides. If it's still not working, post a screenshot and I'll trouble shoot it.  I got it to work fine on both sides.

Edit:  The pump is not symmetric.  It behaves differently on each side. See here for more details. 

[mathmanican]

40 minutes ago, nakomaru said:

I wonder if anyone has a code source for those values.

This would be extremely useful. With it, I think we can predict exactly what g/s value is needed on a valve to cause the element converter to work, or the liquid bead pump to work.  If the value is too low, neither works.  Hit a specific threshold and the matter converter works.  Hit the next threshold and the matter converter stops precisely when the liquid bead pump starts working to get perfect vacuum (current hypothesis).  If you want to not have a perfect vacuum, but want a specific ratio of gasses, like 4:1 or 900:1, I bet we could predict this exactly too.

I could, I guess, decompile the code and then give an exact description, but that leaves out half the fun.  Though here I wouldn't mind having a list of values, as I know at this point how to estimate them. The overflow mass (viscosity) is not always the same, and depends on some other conditions, but it is pretty easy to estimate with a small tolerance.

[0xFADE]

You can use a wheezewort to pump gas as well if temps are not an issue.