My abyssalite flaker with pumped tungsten output: Pre-space and either pre-steel or pre-uranium

[Zarquan]

I've seen some rather interesting abyssalite flaker and I decided I wanted to try one of my own. 

For those who don't know, an abyssalite flaker is a machine designed cause a tile of abyssalite to "flake", or partially melt, by exposing it to a gas at a higher temperature than its melting point.  This is useful because abyssalite is common and kind of worthless (except for geotuning volcanoes) and when it melts, it melts in to liquid tungsten.

One thing I've wanted for a while is to make aerogel (tiles of 0.1 g mass) out of tungsten, so I decided to incorporate the ability to pump the tungsten in to pipes.

Note that I built this version in debug and sandbox, but I placed the gases by destroying pipes and tiles (with a few minor and trivial exceptions), so the gas layout is entirely possible and I see no reason why this arrangement of aerogel could not be made.  Also, this appears to delete materials on load due to a bug involving droplets, but if you turn it off before saving you should be fine.

All pipes are insulated ceramic (though the ones on the input sides of bridges could be anything).  The glass tiles are 0.1 g aerogel.  The refined carbon is 100 kg for no particular reason other than the fact that I wanted it to be thermally conductive.

Here are a couple videos of the setup and the 2 steps required to get the system working.  This is a uranium version, but it could be made using steel in every place where uranium is used.  I will provide typed commentary to go with each video below.  

 

Spoiler

One of the most important things to have set up beforehand is the 10 grams of water in the middle with the less than 1 gram of hydrogen and a pipe full of hydrogen of uranium ore.  This will be important later, as the tungsten will be too cold to be pumped when it first appears.  The water will be deleted by the molten tungsten and the pipe will be broken later to heat up the tungsten

Now, I got the gases in a very particular arrangement that ends up being a form of infinite storage with perfect insulation with an open top that can receive the flaking tungsten.  There are 5 different gases present to comprise the storage, each at 0.1 grams.

Above the 5 gases is a 1 gram cell of oxygen.  This protects the micro-quantities of gas around the pump from being destroyed by the 16 kg/tile hydrogen above.

A similar setup exists below to protect the lower pump which pumps uranium, but all the gases are less than 1 gram per cell (though there is more hydrogen than other gases.

Both pumps are made of copper ore, so this system is not introducing too much heat.

The uranium is picked up by the lower pump and sent up to the liquid shutoff.  If the uranium is cool enough, it will continue up to the metal refinery, where it gets heated and sent back in to the system.  If the metal refinery is full, the uranium continues to another vent.  If the uranium is too hot, it goes straight back in.  I used different vents for each one, as I didn't want any uranium to ever back up. 

The sensor is set to below 3980 C, as I am refining steel, and that gives me plenty of leeway if something goes funky with the automation.  It will be set differently if the uranium is replaced with steel or you are refining something else.

I didn't go over it in the video, but both pumps are covered by a waterfall which is currently looping with a 10 kg/s flow from a pump (the dev pump isn't doing anything.)  This flow can be throttled to 100 g/s to save 99% of the power from the pump, but that isn't really necessary for this build.

To protect the wires, I put in a temp shift plate behind each pump made of diamond, which prevents the micro-quantities of gas from exchanging any extreme heat they pick up with the wires and melting them.

I use refined carbon to speed up the heat transfer and, in the future, to inject heat in to the tungsten, but I don't want that active at this phase (hence the micro-quantity of hydrogen over the pipe).

After a little bit, the abyssalite flakes.  You see that the first drop annihilates the water, resulting in 4990 g of tungsten, or 10 g less than the 5 kg that was produced when it flaked. 

The second video is a couple cycles later.  To finish priming the system, I have to wait for the abyssalite to be tapped out (have less than 5010 kg of mass).  Between these videos, I found that the abyssalite was getting hot, so I added a couple of temp shift plates on the water side to keep it cool.  This does heat the water, but a marginal amount compared to the amount of cooling you can provide a real long-term version of this build.

Spoiler

At this point, the abyssalite is tapped out.  There is now 1994.99 kg of tungsten on that tile.  It will not flake again until you load more abyssalite in to the automatic dispenser and reload (I think they stopped the debris from merging during the same save.

So, I "send in a dupe" to destroy the gas pipe.  (I actually forgot to open their path, so I do that later in the video.)  This creates a functional thermal bridge between the hot hydrogen room and the infinite storage instantly solidifies the tungsten and melts the uranium ore in to depleted uranium.  At this point, we now wait for the system to heat up the tungsten so that it melts. 

One thing that surprised me is that a small amount of tungsten debris ended up in the liquid uranium.  I think this has something to do with the melting of the pipe causing the destruction of the hydrogen in a manner similar to the replacing of a tile, but I'm not sure.  Regardless, the tungsten is now being heated by the hydrogen room and the precise gas layout surrounding the pump is undisturbed.

It takes some time, so I do some things while I wait, including showing how the dupes could have destroyed the pipe, setting up an automatic system to deliver abyssalite, and some other things.  I also noticed that my wire was overloading, but this could easily be on two circuits so I didn't care for this proof of concept.

When the tungsten melts, I wait for it to gain some additional heat.  I had multiple failed runs due to it freezing in the ceramic pipes before it could reach the reservoir.

I do not know of a good way to detect when the tungsten is hot enough or massive enough to pump (and there has to be liquid tungsten there forever or the system breaks), so the tungsten pump is controlled manually in this prototype.

I believe I started spinning up this system with Ren at about cycle 2 (as I had to make sure I could set up the gases naturally), and I was able to get pumped tungsten out by cycle 18.

This post is long, so I won't go over how I got the gases in position in detail with a video, but I will say that I started with having the waterfalls and other liquids in place with just enough space in the rooms for the gases and then I broke pipes or opened canisters where the airflow tile is before I placed the airflow tile, only releasing the next gas when the water forced out the previous gas.  I started with the heaviest gases and then did the lighter ones, with a special note to do the 1 g of oxygen before the 0.1 g of polluted oxygen and wait for the oxygen to get to position before moving on.  I also put 2 g of hydrogen in the room with a tile cap to ensure there was room for all 7 gases, which did not destroy the 0.1 g of gases, and let me reliably open up the system to the higher pressure hydrogen.

Another thing to note is that without any special cooling system, the metal refinery room reached about 50 C over the course of 16 cycles.  That means that this system can easily be cooled and does not require atmosuits to construct or run.

Here are three saves from various points in the priming process. Sandbox.sav Sandbox2.savSandbox3.sav

Edited December 17, 2025 by Zarquan

[Zarquan]

So apparently when the tungsten debris was formed in the third video at around 0:14, it superheated the glass aerogel under it, resulting in it melting and vanishing on load of Sandbox3.  I think the best way to resolve this is to order the construction of a pipe behind the tile once the tungsten is gone, let them deliver the material, then cancel the order before they dig up the tile.  The debris stored in the build order will cool the glass down to a reasonable temperature, although you will have a buried object in there for the rest of time.  And don't save and load before the process is done. 

Edited December 23, 2025 by Zarquan