Discoveries and considerations about LH/LO2 production

[suxkar]

Hello!

After dozens of threads in which I asked weird and complicated things on this forum, I think it's time to share some results I got so that I can be of help to someone.

I've spent basically a day (thank you quarantine) experimenting with different setups and made some pretty useful discoveries.
To spice things up a bit, this time, I decided to go for a different approach to LH/LO2 production then usual. Instead of piping the gases into 2 thermally regulated rooms, I wanted to try the "freeze on the stop" approach. I've seen some setups around so I decided to give it a go. Basically, you make a room full of electrolyzers and cool down to output as it comes out to around -260, so you get liquid hydrogen and solid oxygen. The solid oxygen is shipped into another room and heated up by the super efficient liquid tepidizer. The advantage of such a setup is A LOT a power spared (you don't have to pipe or sort the gasses around, only ship solid oxygen) and a marginal save in calculations (no endless gas pipes). This was the result built in sandbox (ignore the top, it's more testing stuff):

I've put A LOT of liquid pipes and thermo aquatuners just to soo how fast the setup can be. As you can see, the idea in itself is very simple, you can even use the LH to cool down some of the steam turbines. The use of liquid tanks for the supercoolant makes the output temperature highly stable and allows the system to be set to cool down to -261 without risking to solidify the hydrogen.

There is just one, small, gigantic problem: Most of the hydrogen is deleted.

I've run many tests, but in almost every single setup MORE THAN HALF of the hydrogen is not there. Instead, there is sometimes more oxygen then expected (most of the times still at a total mass loss, a couple of times, somehow, with mass gain, compared to the input water).
The setup you see is the most efficient version I could build. I've tried with doors, metal tiles and airflow tiles instead of mesh tiles, much worse.
I've tried to run the pipes inside the tiles/doors, even worse. In the specific case of mesh tiles and pipes running through them (unlike in the photos) I obtained the worst result of all. For 1ton of water input, I received 798kg of O2 and a ridiculous 36kg of H2 (instead of the theoretical 112kg). Aluminium tiles instead of mesh tiles had, somehow, similar results and is tied in worst case scenario. This was due (still talking about mesh tiles with pipes IN them), I discovered, to the fact that hydrogen appears to get stuck inside the mesh tiles. Even though the debug tools detect hydrogen mass inside the tile, deconstructing the tiles doesn't free it. It's basically ghost mass.

Among my many attempts, as I said, this setup is the "best". The best result was obtained limiting the flow of water at 4kg/sec: between 5.5% and 8.1% (inconsistent) of the input water was converted into hydrogen (instead of the ideal 11.2%) and between 83% and 85% into oxygen (instead of the ideal 88.8%). 
Another very curious result is that both limiting to 1kg/s and to 10kg/s (no limit) produced equally bad results: 5.3% hydrogen, 81.5% oxygen and  total mass loss of around 13%. Notice that the water flow is equally split into 4 pipes, so the 1kg/s result seems to prove that the problem is not too much gas production.
The 4 aquatuners seem to provide enough constant cooling even in the 10kg/s case, as the gas pressure inside the chamber never goes above 300g per tile (although it does increase at times).
Ah, also, for the record, electrolyzers have almost 100% average uptime.

My only guess is that the output oxygen pushes hydrogen in the tile where the output hydrogen is generated and somehow the 2 actions have a conflict and less hydrogen is spawned.

I'd be very glad to have someone cross reference my results or show me a similar setups that doesn't waste that much hydrogen.


In the end, I think I'll go for a good old Rodriguez setup with separate chambers. I've build this (I've seen many people do the same) so that there is no need to pipe the extra oxygen.

Even with the doors open, it seems to be working at the same speed, and there seems to be almost no hydrogen loss and no mass loss at all.

Comments and critiques are very welcome 

Here is the map, should anyone feel to play around with it.

PetBoilerEndTube.sav

[Angpaur]

In my prevois colony I created build, where oxygen was liquified in same chamber as electrolyzers, but hydrogen was flowing to separate chamber where it was liquified:

I also observed hydrogen deletion and no matter how I was rebuilding it the deletion persisted. In some cases I was geting almost zero hydrogen

So yeah, as it seems as a very good idea to get rid of gas pumps a some kind of hydrogeb deletion bug makes it useless

[suxkar]

3 hours ago, Angpaur said:

In my prevois colony I created build, where oxygen was liquified in same chamber as electrolyzers, but hydrogen was flowing to separate chamber where it was liquified:

So yeah, as it seems as a very good idea to get rid of gas pumps a some kind of hydrogeb deletion bug makes it useless

I am "happy" to hear that I'm not the only one experiencing this problem and that there is some confirmation.
That is a nice and smart setup, too bad it can't be made to work because of this issue.

While waiting to see if someone has an explanation/solution, I've built this more classic setup:

It works perfectly, uptime on eletrolyzers is not 100% of course but it is still very high (around 85%). It consumes around 6-7 kg/s of water. Most of the time there is a mass gain, there is a bit more hydrogen and noticeably more oxygen then the water that was input,  but that might just gas "staying" in the Rodriguez between different test sessions. Liquid tanks are not essential in the oxygen tank since there is no risk of solidification, but they are in the hydrogen one. I've made some segments of the hydrogen pipes out of thermium to make it so that the oxygen tank does a bit of prechilling, so its aquatuners will work even when the osygen tank is full. Funnily enough, a single segment of thermium already brings the average hydrogen packet from 30C to -190C, so it might be better to use a less conductive metal to spread the heat in the tank.

The automation simply makes it so that if no more oxygen is needed, the doors will stay open and the oxygen pumps will stop in order to save energy, provided that there is water in the small overflow loop up right, to ensure that the doors open only if the electrolyzers are working, otherwise the gas balance in the Rodriguez will get messed up and then become vacuum.

[psusi]

Yea, I really wish they would fix the gas deletion bug.  Apparently it used to convert one gas into another, so if you had a little oxygen in your steam room, over time steam would convert into an ever enlarging supply of oxygen.  Instead of actually fixing it they decided to just delete the steam instead.