MECHANICAL GAS FILTER - For those who missed it.

[Lifegrow]

Quick post for those who regularly ask about mechanical filters, but struggle to find archived info with the search feature.

I've tagged in as many relevant terms as I can, so hopefully people can always find this ever helpful design

Original thread here (bookmark it!)

-Life

 

[Biffa2001]

That is actually really cool, thanks for sharing. With the changes in gases/geysers etc and wanting to store gases now, this will be used a lot

[Lifegrow]

37 minutes ago, Biffa2001 said:

That is actually really cool, thanks for sharing. With the changes in gases/geysers etc and wanting to store gases now, this will be used a lot

Hey no worries bud, you helped me out with a fair few minecraft builds back in the day, so we'll call it quits  

[Saturnus]

One thing you neglected to mention is that with the double bridge version the rotation of the filter gas packets has to be opposite of the direction of flow in the pipe being filtered otherwise packets above 1000g minus filter gas valve setting will overflow the filter.

Eg., If you have the gas filter valve set to 1g and it sees as 1000g packet on the first bridge then with co-current flow 999g will get filtered and 1g sent to the next filter bridge then the gas filter packet, now being a full 1000g packet, will also have moved to the 2nd bridge location meaning the 1g packet will still have nowhere to go. With counter current flow however each time a packet moves the filter packet the 2nd bridge sees will be a 1g packet.

Note that this also means that the double bridge filter can only handle filter gas valve setting divided by 1000 full packets of the same kind in a row because there will be that amount added to the packet the first bridge sees at every cycle. 

As I write in the original thread. If you use a flow control valve you only need the original single bridge version as it then cannot overflow. And rotation of the filter gas packets becomes irrelevant.

Eg., if you set a flow control valve to 999g/s and the gas filter valve to 1g/s there's no way two merged packets combined can exceed 1000g.

Although I use both types I think most people that are new to building them should use the flow control valve version as you can effectively prime the filters with one packet of gas of each type by setting the flow control valve 0g/s first. Then as the first packet hits it you up the setting to 1-10g/s thereby priming the filter (or filters) before you turn it all the way up to 999g/s.

[Lifegrow]

I thought I covered this with the explanation about using a valve to restrict flow to 499g/999g depending on which version you use ?

[Kasuha]

2 hours ago, Lifegrow said:

Quick post for those who regularly ask about mechanical filters, but struggle to find archived info with the search feature.

I see you finally cracked the design, congrats.

By the way, if you're concerned about overflowing/spilling of the filter, it's also possible to build its "perfect" version. It's a bit more complicated but will not malfunction even if the filtered output blocks.

[Saturnus]

18 minutes ago, Lifegrow said:

I thought I covered this with the explanation about using a valve to restrict flow to 499g/999g depending on which version you use ?

But that's not accurate. You don't need to limit it to 499g/s ever (unless the filter gas valve is above 1g/s obviously). As I wrote, 1000 minus filter gas valve setting equals flow control valve setting.

As I also wrote above, you do not ever need the double bridge version if you use a flow control valve.

[TrustyFish]

Thanks for the video, and thanks to everyone who developed the tech. It makes perfect sense now that I see it, and is both useful and gives a great foundation for experimenting with similar systems as new problems present themselves.

[Lifegrow]

18 minutes ago, Kasuha said:

I see you finally cracked the design, congrats.

Ehhh, what? 

Well thanks for the clarity as ever @Saturnus @Kasuha  I was trying to keep it as short a video as possible, and rather instead refer people to the original thread than make it a 20 minute video. This is why I showed, discussed and shared the link to the original thread. Also, in my defence, it was silly'o'clock and an off the cuff video in response to a yt question  

I was under the impression that slapping a valve in before the first bridge and restricting the flow was a nice easy fail safe - especially with the current "stutter" bug that can often force packets/effect flow when any piece of pipe is altered on your piping network. Hence me mentioning that if you want this to be pretty much flawless, also reduce the input of the standard design to 499g/s.

Again, this was based off your own words way back when :

"If you make sure the incoming packet size is smaller than the valve maximum the filter will work indefinitely. As valve split to an average output that may be a bit higher or lower in each individual packet you may want leave about 10% remaining capacity for that so if the feedback filter valves are set to 100g/s then set the input valves to about 800g/s as an example."

5 minutes ago, TrustyFish said:

Thanks for the video, and thanks to everyone who developed the tech. It makes perfect sense now that I see it, and is both useful and gives a great foundation for experimenting with similar systems as new problems present themselves.

Glad it helped buddy - this was exactly why I wanted to share it. 

[Kasuha]

9 minutes ago, Lifegrow said:

Ehhh, what? 

Errrr... I guess I've been sleeping under the rock for a while but I still remember you like this 

[Lifegrow]

Just now, Kasuha said:

Errrr... I guess I've been sleeping under the rock for a while but I still remember you like this 

Why do you think I have it bookmarked?  I refer to that thread pretty much every time I build one - and I still often build my valves backwards  

[Saturnus]

8 minutes ago, Lifegrow said:

Why do you think I have it bookmarked?  I refer to that thread pretty much every time I build one - and I still often build my valves backwards  

That's just because back when it was posted the colours for inputs and outputs were reversed.

Anyway. I made a quick mash up of the simple filter type and its common extensions. Note these are spaced out for maximum clarity but can obviously be compressed a lot.

From left to right it's

- The standard one bridge filter.

- Flow control valve added

- Output block flow back added

- Filter line loop back added

[Lifegrow]

6 minutes ago, Saturnus said:

That's just because back when it was posted the colours for inputs and outputs were reversed.

Now that's something I forgot to mention in the vid - I did put it in the comment on my video, but whether people will read that or not is out of my control - I think I did add a "card" thing on the video explaining it too - but it's hard to add little edit messages now that annotations have been removed...

Sadly you can't just add a "message" it has to be a poll or a link or something  

[Arash70]

Thanks @Lifegrow

You should make noob friendly oxygen not included guides, your videos are awesome.

[Lifegrow]

2 minutes ago, Arash70 said:

Thanks @Lifegrow

You should make noob friendly oxygen not included guides, your videos are awesome.

Thank you buddy - I do have a "tips and tricks" playlist here : https://www.youtube.com/playlist?list=PLVLEZuI-7xWwi0zlPgzztWfI35Ue6xAVY

Honestly though it's more of a collection of things that people have asked me to do, or questions that come up a lot on stream (easier to direct people to a youtube video than explain it every time).

I would like to do a full "beginners guide" for ONI - but due to the fact that it's an ever changing game at the moment, I think it'd probably serve to misinform people with often outdated information more than help people. Nothing worse than releasing a video, then 1 week later it's "stealth hotfixed" into oblivion  

It's why I released no videos for the ranching update pre-release - look how many things have changed since day 1....

 

[Moggles]

Really useful thanks man. I read the original PVD thread last year but have still never built one.

I thought back then that it looked a bit fiddly to prime and have always stuck to the out of the box filter option. Your video really showed that it's pretty simple to prime. Well explained. Will you be streaming some more ONI when the update is released?

[kalafus2002]

This is an excellent idea. By using this mechanical filter you can avoid an issue with normal one, that is, once any output is full and the gas to that output enters it stops. This one will let gas pass through the main line and you can add another mechanical filter down the line for the same element to create priorities (e.g. first O2 filter to fill in exosuits , second to recycle back to base and third to send it to O2 storage tank). You can loop back the gas to prevent any gas loss and as long as you have a mechanical filter for every gas and send it to gas compressor (hopefully it still works after update) it will never stop. And if one pipeline can't handle the amount of gas you can easily make dozens more since it doesn't need any power (except the initial gas pump but you would have to build that anyway).

 

My expected setup:

Build things in vacuum (already learned how to make vacuum easily) and group them so that only single type of gas will be produced. When that is not possible use the mechanical filter. Valve to reduce gas packet size, array of mechanical filters for every element that can enter the pipeline (more than once to set flow priority and to send the excess gas to gas compressors), loop back to before the valve (basically your 'Filter line loop back' type) to avoid any gas loss. Repeat the setup if one pipeline can't handle the gas flow.

Also, what is the purpose of 'Output block flow back' ? I can't seem to figure that one out.

[Saturnus]

3 hours ago, kalafus2002 said:

Also, what is the purpose of 'Output block flow back' ? I can't seem to figure that one out.

It's if your system can have temporary blocks on the output line, you can loop it back to try to output it again later. Basically just an output buffer.