Sign in to follow this  
biopon

Melty Papa wants you to love your regolith

Recommended Posts

biopon    224

This, I think is what you should do with your regolith, if you don't mind snaking it down to your nearest volcano.

pupper.thumb.gif.bae0ae030414a3ef2edbd59cb1455426.gif

This is a sandbox build but should be easy to recreate in survival; anything to the right of the miner (the magma source) is left up to you. I just painted in a really overpressurized blob, ignore that, and picture a volcano magma storage instead.

Input is 20kg/sec regolith, or a bit less to be honest. Output is 1200C igneous of the same amount. The magma used is about 700g/sec when running at full speed, but it's only taken down to 1500C, you can still make use of it after that.

The output can run 20 turbines at full speed and leave you with 300C rock, which can be fed to hatches (if you build a hatch ranch in vacuum). Food for 85 groomed hatches comfortably (or many more if glum), producing 6000 tons of coal per cycle, enough for a whole lot of ceramics or 10 always-on coal gens.

The left side is just a big heat exchanger. The bridges into the diamond boxes are totally weird, but needed because conveyors are bugged and the bridges kinda fix the bug. 

Here's the right side:

image.thumb.png.c3639e8ee853f46207ddf5eddcdd2694.png

The fresh regolith comes in on the bottom right. It goes over some airflow tiles, this is where the molten regolith solidifies, the 300C regolith cools it just a tiny bit into 1407C igneous. Airflow tiles guarantee no heat loss from the rock. The regolith then exits to the heat exchanger, and comes back in at the top left. It bridges into the melting box where another bridge keeps it going around until it's melted. It drips down, solidifes, the sweeper picks it up and feeds it to the loader, after which it goes into the heat exchanger. The sweeper and loader are buffer-cooled here with dirt tempshifts and drops of petrol. If you build this for real you may want to pipe some coolant through the base tiles - you can lose the tempshifts then.

Heat: when the leftmost magma blob cools below 1500C, the door below it opens and it's vented to space. The door then closes, and the topmost door opens to let in magma until the hydro sensor detects any amount of liquid, at which point the topmost door closes, and refueling is complete. To the left, the inverted cross-shaped thing is an intermediary heat buffer that aims to keep itself at 1450C with the thermo sensor and the door to its right. Without this the temperature would spike too much in the melting box, wasting magma and causing all kinds of accidents. To the left, the sideways cross that's connected to the melting tiles aims to keep itself at 1421C. I figured this out with trial and error, this setting gets very close to 20kg/sec throughput.

 I had the miner at the top to help with priming as the initial magma drop solidified. It can be removed once things are up and running.

image.thumb.png.56a41994370026255593499618b30c9f.png

The buffer is 25sec, the top left filter is 20, bottom filter is 5. Ignore the miner, I just routed the wire through it for convenience. The rightmost door and the hydro sensor connected to it are only there for super high pressure shenanigans, ignore those too. The thermo sensors are above 1421 and above 1450, respectively. 

The automation might be rearranged to be more space-efficient and not stick out at the top. It's the first revision, so not a whole lot was put into that, sorry.

image.thumb.png.d42c7db105921c7af004316c04b5389b.png

Window tiles are diamond (duh), tempshifts also (except where used as cooling buffers, those are dirt). Doors are steel, so are wires and sensors. The loader & sweeper can be anything. The rails should be iron ore, wolframite or steel where they can get hot. (I circled the spots above, and obviously also in the heat exchanger where they're within diamond).

You want a few grams of water on the two thermo sensors on the left, before you encase them completely. And obviously, you don't have to build it in space (no need for the drywall then) but vacuum is a must.

The heat exchanger could be smaller. If you don't mind spending more magma, you can safely remove 1-2 boxes from it. As it is, the regolith is preheated to 1403.7.

EDIT:

This is the heat exchanger you want to use, courtesy of @nakomaru. 4 7x4 blocks, 1 apart horizontally, 2 apart vertically. It works onscreen and offscreen just as well. Obsidian tiles, iron and iron ore for metal. 

image.thumb.png.7a2d7505fd9ba487de8faeea43aa0bef.png

image.thumb.png.9e9561f923391ef5b03dc27dfe5eec8c.png

You also want automation wire bridges, obsidian gas and liquid bridges the same way. You can do whatever with obsidian pipes to make the "no output" bridge flashes go away.

A bugfix with the build itself:

image.thumb.png.961aa1011a7af86b839a5b9f5a9346bd.png

image.thumb.png.a4758a3512b79a5aaf59da0688cc1f6f.png

You want that airflow tile and that conveyor bridge. Otherwise, maybe, lava, all over.

 

Edited by biopon
reason
  • Like 4

Share this post


Link to post
Share on other sites
StarSquid    41

This is a very elegant little system, and barring the heat exchanger seems an awful lot smaller than most of the regolith melters I've seen. I feel like that magma consumption could be seriously reduced if the heat exchanger could be improved somehow.

  • Like 1
  • Thanks 1

Share this post


Link to post
Share on other sites
biopon    224
4 minutes ago, StarSquid said:

magma consumption could be seriously reduced if the heat exchanger could be improved somehow

Well regolith is pretty stubborn, I'm pretty sure the best way to get it to budge is to encase it in something very conductive, and with diamond this is how much it takes. Aluminium would be much better but it'd melt in anything but the 1st box (if it'd even last there). Thermium is the only step up, but if I had that much thermium I wouldn't need anything like this...

You can make it smaller and worse, or bigger and better, but short of going to thermium, I don't see how else to make it better.

I guess molten aluminium is a thing, I'll have to see how that'd work.

Share this post


Link to post
Share on other sites
nakomaru    1,381

(most) Entities inside of a cell q = k_low・ΔT・Δt・1000

Igneous rock only has a conductivity of 2, and will be k_low whether you use diamond or thermium. Going higher than diamond will improve the tile to tile transfer but I'm not sure that will matter much for this design. Plus, because rock has such higher capacity than regolith, you don't even need high efficiency, just enough time.

You could also cool the lava with the hottest regolith rather than the coldest regolith.

I like the build. Inspiring.

Edited by nakomaru

Share this post


Link to post
Share on other sites
biopon    224
2 minutes ago, nakomaru said:

will be k_low whether you use diamond or thermium

I suppose this means the exchanger boxes could be made of anything down to igneous... Good news and bad news I guess. Frankly, I should have known this, I'm a bit embarrassed. Also, I don't get to play with molten aluminium. :(

10 minutes ago, nakomaru said:

if you redesign your exchangers to use counterflow

It is counterflow right now, except compartmentalized... Ideally it'd be one long snake of 2 conveyor lines running parallel, encased in something, but the conveyor bug doesn't let that happen, so this is the next best thing I could come up with.

12 minutes ago, nakomaru said:

You could also cool the lava with the hottest regolith rather than the coldest regolith.

The hot regolith comes in at 1403+, and the magma drips down at 1420-ish, you don't want those two to meet. :) I'd get molten regolith on the cooling platform and the whole thing would grind to a halt. Besides, the cool regolith takes that heat energy with it into the first exchanger box, so it's not like it's lost or causing spikes.

17 minutes ago, nakomaru said:

I like the build. Inspiring.

Thanks, it means a lot.

Share this post


Link to post
Share on other sites
nakomaru    1,381
8 hours ago, biopon said:

It is counterflow right now, except compartmentalized

Yes, I mean you could even make the compartments counterflow, but acknowledged bug would prevent it from always working. With low conductivity doors you can get them to instantly transfer heat to the counter but only slowly to each other. But then I remembered the heat in the rock is so much higher than regolith that efficiency doesn't matter in this system.

Edited by nakomaru

Share this post


Link to post
Share on other sites
biopon    224
1 minute ago, nakomaru said:

efficiency doesn't matter in this system

It does, because the better the exchanger, the smaller it can be, and the less magma it needs. I'm happy with getting regolith up to 1403, it needs very little energy from there to melt - but I'm not happy with the heat exchanger's size. In the contraption I don't have too many wasted tiles, on the left side though, it's another story. So what's this about low conductivity doors? Something like a series of doors, where they don't really transfer to each other, but whatever's encased is equalized quickly?

Share this post


Link to post
Share on other sites
nakomaru    1,381

Sorry, of course you're right about efficiency. I'm trying to see if doors can be of any use here. It's tough. The real barrier is that regolith needs to spend enough time in tiles, and that sort of sets your minimum size. I think it might be worth it to utilize the fact that debris above tiles also transfers with the tile below (at 1/4th rate), to let your regolith spend more time transferring.

I did some comparisons to see if doors are worth it. I started with 200% the size of your heat exchanger and in pure counterflow. I used Iron Ore. I kept the system in full view to hopefully avoid the bug for now, and waited for equilibrium.

Spoiler

bigdoors.thumb.gif.88d0fa508feef7a26c5ca200ec654970.gif

For "100%" I cut off the top 2 rows. For "50%" I cut off the top 3, leaving only 2x8x4 blocks.

  • Inputs: Regolith@580K, Rock@1682K
  •  
  • 200% Iron Ore Door: Regolith@1681.1K (0.9ΔK)
  • 200% Diamond: Regolith@1681K (1.0ΔK)
  •  
  • 100% Iron Ore Door: 1.2ΔK
  • 100% Diamond: 1.1ΔK
  • 106% Diamond: 0.6ΔK
  •  
  • 50% Iron Ore Door: 21.2ΔK
  • 50% Iron Metal: 11.4K
  • 56% Diamond: 16.5ΔK
  • 50% Diamond: 15.9ΔK
  • 50% Diamond (non counterflow, like OP): 37.5ΔK
  • 50% Steel: 15.9ΔK
  • 50% Thermium Metal: 16.1ΔK
  • 50% Thermium Door: 16.1ΔK
  • 50% Thermium Mech Door: 16.4ΔK

There's some variation with play speed, and I'm posting with these at 20x.

As I suspected, even within the blocks it is a big deal to utilize counterflow: 37.5ΔK vs 15.9ΔK at half size.

So, I was clearly wrong about doors as they were soundly beat at 50% by more conductive materials. I imagine the physics of doors are actually two ordinary tiles, and the reported temperature is merely the average of those two tiles. So I no longer believe they instantly transfer heat across them. This is also supported by identical performance of thermium doors and thermium metal.

But when you get too conductive, it works against the counterflow. A surprising result to me is that Iron metal tiles were the best I could find. It seems there is a sweet spot between conductivity which allows you to transfer heat left and right quickly enough and at a small temperature gap, but not so quickly that the column approaches equilibrium.

Because steel and iron should give about the same results, I'm double checking those tests now.

The +6% tests were where I added a modification to the end like this:

Spoiler

mod.thumb.png.9fdfd8767434ee85d2e9e9b0e269fd76.png

Regolith (conductivity 1) goes on the outside and Rock (conductivity 2) goes on the inside. The idea is that we should probably want some balance between heat transfer of regolith and rock, so we want the regolith to be touching twice as many tiles as rock. It works! But only sometimes?! Not sure how bad it hits your FPS. Rails are some of the worst in the game atm.

With this, I'm going to see if I can find a more optimized exchanger that works to around 2~5ΔT.

Edited by nakomaru
  • Like 1

Share this post


Link to post
Share on other sites
biopon    224
7 minutes ago, nakomaru said:

o we want the regolith to be touching twice as many tiles as rock

By adding bridges to the middle line (and also removing some from the regolith line to keep the 2:1 balance) you could also enable (some of) the counterflow mechanic to work while offscreen. 

A question re. bridges: I always thought whatever was supposed to be on the bridge input tile does not matter thermally because it's not really there: it's instantly moved by the bridge to the output tile if the line is in motion. If the line backs up then both tiles will have packets, but otherwise it's just the output. Am I wrong? 

  • Like 1

Share this post


Link to post
Share on other sites
nakomaru    1,381
14 minutes ago, biopon said:

I always thought whatever was supposed to be on the bridge input tile does not matter thermally because it's not really there: it's instantly moved by the bridge to the output tile if the line is in motion. If the line backs up then both tiles will have packets, but otherwise it's just the output. Am I wrong? 

Ah, you're right again. It must have only worked because it was assisting the left-right heat transfer, or minimizing the bug. Or just having more tiles to transfer heat from the rock away helped. Or I made an error. Or going crazy.

7.thumb.gif.dbd3b8d333d944e13746a2c8d47bdba9.gif

Edited by nakomaru

Share this post


Link to post
Share on other sites
Yalp    49
2 hours ago, nakomaru said:

I imagine the physics of doors are actually two ordinary tiles

Yeah, the door has a block of its element for each tile. You can see/select it when clicking the tile twice. Each tile has its own temperature. The door's displayed temperature is the same as the bottom tile (root tile).

Spoiler

5dab0126365aa_DoortilesTemperature.thumb.gif.91782c2a852fbd04176bd77879e26aef.gif

In addition, when opening the door (or just set it to "open" and then "close" while paused) , it instantly sets both blocks' temperature to that of the bottom (root) one. (This was causing issues with the metal cannon a while back as well).

Spoiler

5dab02833c27d_DoorHeatAdjustment.thumb.gif.5d5331b33f7a78878066a990e33d9cdc.gif

 

Edited by Yalp
  • Like 1
  • Thanks 1

Share this post


Link to post
Share on other sites
nakomaru    1,381

I had always heard it places those blocks under the doors, but I assumed it was using a virtual building temp for them. Didn't know you click on the ore itself, although I'm sure I should have.  Plus, the reported temperature and the reset temp is just the root rather than any sort of average. Wow. Glad I can be taught so much today.

Edited by nakomaru

Share this post


Link to post
Share on other sites
Xenologist    1,237
1 hour ago, Yalp said:

 

  Hide contents

5dab02833c27d_DoorHeatAdjustment.thumb.gif.5d5331b33f7a78878066a990e33d9cdc.gif

That, with automation, could have endless heat deletion potential.

Share this post


Link to post
Share on other sites
biopon    224

It can certainly multiply a lot of heat or cold upwards or to the right.

Sour gas boiler & condenser with 300 doors, 600 tempshifts, and one thermo regulator?

Share this post


Link to post
Share on other sites
nakomaru    1,381

Best I could do in a slightly smaller footprint. I benched the original at 3.8ΔK, counterflow variant at 1.1ΔK, and this one at 1.5ΔK. Iron/Obsidian. Conveyer bridges steel. Should have filled in the left with rails, but I'm too tired now.7.thumb.gif.c905450e7ed10798a6f2639c0b6247ef.gif

Edited by nakomaru
  • Like 2

Share this post


Link to post
Share on other sites
biopon    224

Amazing. Looks like it should be OK for offscreen use as well. I'll test it and report back.

Share this post


Link to post
Share on other sites
biopon    224
17 hours ago, nakomaru said:

Best I could do in a slightly smaller footprint. 

So it works really well with the build while onscreen, and it's significantly slower than what I put together. I like the wire bridges, did not even think of that.

The issue is that if I move it offscreen for a couple of cycles, the rego exit temp plummets to the 1370s. 

It's tricky, because for offscreen you want something like this:

image.png.906d18ea8892a6c7a3b7783373a228d5.png

You have to break up the line with bridges into multiple segments, and put the bridge exit points right next to each other.

Or, better yet, something like this where you add length to the segments so all those chunks can exchange heat at the bridge exit preceding them:

image.png.4e1a7963e2da74e8de79c3af4f4c3d3e.png

Unfortunately this is very different from what works while onscreen, and it seems very hard to do both well while keeping the size optimal. 

 

Share this post


Link to post
Share on other sites
nakomaru    1,381

It can be done with the above design pretty easily. It's partially done for the bottom right for regolith. You can have two bridges pointing back to the center every third tile. I'll mod it tomorrow.

It takes up one more row on the top and bottom, but you might be using insulation there anyway. Hmm, center needs one more row too.

Edited by nakomaru

Share this post


Link to post
Share on other sites
biopon    224

I am not sure if the bug workaround works if you don't actually use the bridge, just snake through the exit point. it might though - but even if it doesn't, and you have to use the bridge, the input tile is dead so you lose 1/3rd surface area for onscreen. Unless you can move all bridge inputs to empty space between the boxes, the efficiency will go down. :(

Anyway I'll wait to see what you come up with. I'm tinkering with separate boxes, a bit like the original, but I don't think it's very compelling as they will have their internal temps equalized while not visible, so to get reasonably good exchange you need a lot of them.

I had this idea for swapping heat slowly but for large amounts at a time, like 25t chunks. You can move them through 2 series of doors, counterflow. Unfortunately it needs an insane amount of hot igneous to get going with, and the automation size turned out abhorrent so I abandoned it.

 

Share this post


Link to post
Share on other sites
biopon    224

That is pretty frickin amazing. 

image.thumb.png.c18f406b5b1dca2a349476404da8dc39.png

The input was 1408.8 so 2 degrees delta, that's fantastic. This was 90% offscreen by the way.

While you were busy with stuff of actual worth, I made this:

bewg.thumb.gif.14971d718b70917571ac34321a617744.gif

Inputs are 1407 and 207 because typo. That door temperature reset is a powerful thing.

  • Like 2

Share this post


Link to post
Share on other sites
nakomaru    1,381

That was just a sketch before I could workshop it. I worked on it a bit more and I'm very happy with this design now.

7.thumb.gif.495abd45d84b3a6441b27d747507836d.gif

Materials are all Obsidian/Iron/Iron Ore.

Results in 1.3ΔK on screen and 1.0ΔK off screen. If you extend it by 3 tiles this becomes 0.3ΔK always. (20x speed numbers)

Rail path:

Spoiler

5.thumb.gif.da121eee5940bd11656e994c975bb037.gif

Much easier to follow and aesthetic now too. I don't usually try to min max in sandbox, so thanks for the inspiration. :)

Edited by nakomaru
  • Like 2

Share this post


Link to post
Share on other sites
biopon    224

It's alive in survival :D

image.thumb.png.31bf54c1b793c7432e235246b1c48ce3.png

I guess I wasn't quite ready for the turbines to take up so much space, with the added 4 to cool the 2 ATs and take the igneous down to 125C. 

At least 40kg/sec desalination/pwater boiling is now an option I have...

  • Like 1

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this