Show me your LOX

[Occam Blazer]

Spoiler

I'm testing a liquid hydrogen / oxygen machine, but before I present it to y'all I want to see how it compares to what's already out there. I've done some googlin' already and I think I have a build that spins up way faster.

My build uses about a ton of super coolant for chillin and can be made without thermium (but it's a little more efficient using about 1,600 kg of the stuff). It produces liquid at 1 kg/sec about three cycles after you turn on the feed.

Whatcha got?

[degr]

It process 1kg/s oxygen and hydrogen.

[Fradow]

The most comprehensive I know of is Hotep Thunderbolt's one (he doesn't come on the forums, but is okay with his designs being shared here).

The Integrated rocket refueler

 

Having large LOX/LH tanks, it obviously doesn't emphasis spinning up fast. It has some power-saving by merging the Hydra and the LOX/LH maker, and according to his descriptions, has runned 8-9 hundreds cycles on triple speed without any noticeable issue, including save/load cycles (there is an OnLoad detector)

[DimaB77]

I don't understand the point of such schemes. 1000 sensors, kilometers of pipes. Combine one with the other, so it would be difficult to understand and unrealistic to maintain?

I already gave you a simple scheme:

1 kg/s of liquid hydrogen and oxygen. The first batch after 2...3 cycles. Works with gases from -20C to +80C (and maybe more in both directions). 2 sensors (hydro-sensors can be thrown out). Minimum of supercoolant and other materials. No transformers or powerful wires.

[Occam Blazer]

I am finally getting around to running some tests and have initial results. I need to rejigger the tests so I'm making a fair comparison and I should be able to finish up tests this weekend. Anybody else want to see how their rocket fuel setups compare?

[degr]

yes sure

Delete please drywall from my setup. I think drywall will take additional time to become cold enough, I add it because it was build in space.

Also, DimaB77 have no drywall at all, only tempshift plates

[mathmanican]

On 6/14/2021 at 12:09 PM, Occam Blazer said:

spins up way faster

If the goal is to get them to spin up faster, then every build above will benefit from using airflow tiles between the liquid and insulation walls. Why? By preventing the liquid from touching the solid insulation walls, you bypass partial evaporation (flaking) completely.  It just cannot happen. Hence, every build above will "spin up faster" with a layer to keep the liquid away from solids.

It does make each build bulkier (so that may be a downside), but will improve every build you have above if your metric is "spin up time."  The fewer tiles you have cool via flaking, the faster the build will start producing for you. 

[Gurgel]

On 6/16/2021 at 2:09 PM, DimaB77 said:

I don't understand the point of such schemes. 1000 sensors, kilometers of pipes. Combine one with the other, so it would be difficult to understand and unrealistic to maintain?

One thing well-known in engineering is that the more experienced and skillful the engineer, the simpler the designs get. The same is true for coding (which really is engineering): Solving a problem with tons of elements is easy. Solving it simple, small and compact is hard. On the other hand, the simpler, the more reliable, cheaper and easier to build.

[Autoskip]

10 hours ago, Gurgel said:

One thing well-known in engineering is that the more experienced and skillful the engineer, the simpler the designs get. The same is true for coding (which really is engineering): Solving a problem with tons of elements is easy. Solving it simple, small and compact is hard. On the other hand, the simpler, the more reliable, cheaper and easier to build.

My dad has a Russian Dnepr motorcycle (the "D" is silent), and it's often the very definition of this - at one point the speedo was playing up, so dad looked up how they worked and all the tiny little pieces all working together in a complicated mess - and then opened up the Dnepr's speedo and found out that it had about two parts.

[Hokaeru]

15 hours ago, Gurgel said:

One thing well-known in engineering is that the more experienced and skillful the engineer, the simpler the designs get. The same is true for coding (which really is engineering): Solving a problem with tons of elements is easy. Solving it simple, small and compact is hard. On the other hand, the simpler, the more reliable, cheaper and easier to build.

Only true if you do not have OCD and want a flashy symetrical thermium overkill design  

[Occam Blazer]

Something changed in the most recent update. I'm getting liquid teleportation in a build that used to work perfectly. Ideas how to prevent this?

There's a little pool of hydrogen that appears two tiles up and one tile right of the air vent.

 

Spoiler

 

[mathmanican]

1 hour ago, Occam Blazer said:

Ideas how to prevent this?

Remove the airflow tile under the metal tile (by the liquid reservoir).

[Occam Blazer]

14 minutes ago, mathmanican said:

Remove the airflow tile under the metal tile (by the liquid reservoir).

Thanks. Bet that also explains why the puddle of supercoolant on the oxygen side now gets replaced by liquid oxygen sometimes.

[Autoskip]

8 hours ago, Occam Blazer said:

Thanks. Bet that also explains why the puddle of supercoolant on the oxygen side now gets replaced by liquid oxygen sometimes.

Yup - gasses liquifying in airflow tiles tend to tunnel directly upwards until they find somewhere they like.

Speaking of, you might be leaking some LH and LOX over the airflow walls too...

Source: https://steamcommunity.com/sharedfiles/filedetails/?id=2154398396

[Tranoze]

Anyone else use door to pump and compress liquid hydrogen/oxygen? Why everyone just use big tank....
Kinda not have any screenshot rn because deleted that world months ago, but it look like this from afar

[Occam Blazer]

On 7/4/2021 at 10:04 AM, Tranoze said:

it look like this from afar

That's very far away. (-:

                                                                                                     

Now, on to the results. This is my test apparatus. Left to right, we have @degr, @DimaB77, the Integrated Rocket Refueler (IRR), and my design.

I let everything reach operating temperature before starting the trial runs. Each machine is simultaneously started and independently timed. They received one pipe each of oxygen and hydrogen at 80°C. Once the machine generates 6,000 kg each of liquid hydrogen and liquid oxygen its timer stops. The theoretical minimum time is 6,000 seconds on one pipe of gas.

The IRR was the consistent winner at around 6,300 seconds per run. I ran a bunch of races, tweaking my design here and there, but couldn't get better than 6,562 seconds. The IRR just throws a lot of cooling power at a small amount of gas. I did have to tweak it from the manufacturer's recommended settings because it kept freezing hydrogen. Once I dialed that in and added airflow tiles to the walls the IRR was consistently faster.

Degr comes in third position with around 7,800 seconds. The design is not that much different, so with a few tweaks it can easily achieve the same throughput as the IRR.

DimaB's machine only uses one aquatuner, so it was significantly slower than the other three. Slap another AT on there and add flaking abatement and it'll get just as quick.

Takeaways

The most significant thing you can do to speed up the process is to prevent flaking as suggested by @mathmanican. It saves you a lot of cooling time. All you have to do is watch out for gas condensing in airflow tiles. The resulting liquid will travel straight up until it finds a suitable tile.

Just run a bunch of supercoolant at the precise right temperature through a bunch of gold, steel, aluminum, or thermium pipe.

Side note

I had to bypass the door pump on the hydrogen side of the IRR. In my initial runs it kept pace with oxygen, but lagged far behind on hydrogen. The feed pumps showed about the same uptime as the other three feed pumps, but material was missing. I believe the automation timing is deleting gas.

                                                                                                     

I've learned some lessons from the IRR and am gonna tweak my design to see if I can beat it.

[gabberworld]

On 7/3/2021 at 8:43 PM, Occam Blazer said:

Something changed in the most recent update. I'm getting liquid teleportation in a build that used to work perfectly. Ideas how to prevent this?

There's a little pool of hydrogen that appears two tiles up and one tile right of the air vent.

 

  Reveal hidden contents

 

it can be bug related to the this 

 

[Occam Blazer]

1 minute ago, gabberworld said:

it can be bug related to the this

I'd like to see what's at the bottom end of that column of tiles where the polluted water comes from.

My problem is probably this:

21 hours ago, Autoskip said:

Yup - gasses liquifying in airflow tiles tend to tunnel directly upwards until they find somewhere they like.

 

 

[gabberworld]

6 minutes ago, Occam Blazer said:

I'd like to see what's at the bottom end of that column of tiles where the polluted water comes from.

thats the point. nothing. as you see in that video he not have pipes at there

or you mean even more bottom, like some transportation, like it happen some other items

[Occam Blazer]

43 minutes ago, gabberworld said:

even more bottom

Yeah. Go straight down to see if liquids could be doing weird things.

[Fradow]

Beads teleport upward when trapped.

Two known ways to do that are:

1. Condensing a gas into an Airflow tile. That principle is used in the Saltuner, for example.
2. Trapping liquid in Mechanized Door. That principle can be used for quick pumping of any liquid into an infinite storage without any pump.

There might be other ways to perform that, I'll gladly add them to the wiki if anyone find one.

Lengthier explanation on the wiki.

 

To stay on subject: great experimentation on those designs I passed down the info to the IRR creator.

[ghkbrew]

Using gold metal tiles instead of airflow tiles works well too. You still end up flaking to cool them down, but they have 20x less specific heat than an igneous insulated tile. And you don't have to worry about gas condensing them and teleporting away.

[Autoskip]

13 hours ago, Occam Blazer said:

Yeah. Go straight down to see if liquids could be doing weird things.

I downloaded their save file, and look what I found even more bottom!

Yes, that door is directly below the leak, yes the leak happened as the door closed, and yes the leak was fixed when I dug out the tile immediately above the door.

It's a fairly nice base, but there's no LOX yet so not really interesting in this conversation.

[Occam Blazer]

Rawr

[DimaB77]

I am trying to find the advantages of this scheme to add it to the site. The scheme works, interesting, but there are disadvantages, imho:

• 2 ATs and the need to use improved wires
• big amount of supercoolant, insulation, termium
• big amount of automatics, bulky scheme
• inconvenient removal of pipes from below (the rocket is usually at the top)

Here's my version (it's a concept, there is still room for improvement):