Gas/Liquid Filters Replacement

[blash365]

On 27.4.2018 at 5:27 PM, KittenIsAGeek said:

It doesn't even need to be that complicated.  You only need one gas shutoff and a gas type sensor.  And it works like a charm.  The advantage of an air filter is you can get them going rather early in the game.  The advantage of this method is power efficiency.

I've never had a packet slip.  The sensor turns on the valve for carbon dioxide.  Otherwise, the packet continues along the pipe.

This does not work if either of your outputs has a "jam".

In that case the valve will open (depending on sensor), but the flow will follow the exit that is currently "unjammed". It works in both ways and can create unwanted packets in your otherwise well-filtered stream.

The op's version might allow you to avoid this jamming effect.

[MidnightSteam]

30 minutes ago, blash365 said:

The op's version might allow you to avoid this jamming effect.

They both failed the clogged pipe test.

One advantage koningkrush's build has is that it won't let any gas pass if there's a power outage.

[wachunga]

Loop back excess resulting from a clog with another sensor/shutoff. Prioritize the loop by having your input bridge into it. Both sensors are set to the gas you want filtered, second sensor runs through a not gate to the second shutoff. Left is input, middle is selected gas, and right is everything else. You can do the same with just bridges, but it's uglier and bulkier. 

Also as Dude42 and Midnight said the OP's design is flawed as the shutoffs are after a T.

[BDelacroix]

Quote

 

I've had a few leaks using this but it is not terribly egregious.

I don't know why an empty quote seems to be attached here.

[clickrush]

This build is super useful. However as others mentioned the two shutoffs need to be inline/serial for it to work w/o fail.

[XEVEN]

On 4/26/2018 at 6:15 PM, Koningkrush said:

I don't know if this is already known or not, but gas and liquid filters can now be created in a 20W version by using two Gas Shutoffs, a NOT Gate, and a Gas/Liquid Pipe Element Sensor.

I like this a lot better than the single valve design. Sure, technically hydrogen should never back up because of hydrogen generators but I think in most use cases you want your O2 to back up after its pressurized everything. Shutting down your electrolyzers and gas pups is ideal because it uses a lot of power and water. If you're using up all of your geyser water you have no pool to temper it with and it generates a lot of heat.

[nivodeus]

On 5/24/2018 at 12:40 AM, clickrush said:

This build is super useful. However as others mentioned the two shutoffs need to be inline/serial for it to work w/o fail.

could you please explain more? what do you mean by inline? i tried the design, it works at first but it leaks after a while. At first i thought maybe because the pump wasnt pumping constantly creating a back log, but even with constant pumping, it still doesnt work 100%. 

[Le0n1des]

1 hour ago, nivodeus said:

could you please explain more? what do you mean by inline? i tried the design, it works at first but it leaks after a while. At first i thought maybe because the pump wasnt pumping constantly creating a back log, but even with constant pumping, it still doesnt work 100%. 

Instead of having the shutoffs on a split, place them in parallel, one besides the other on the same pipe.

Here's a nice guide (in Russian)

 

[Santhor]

Problem is with all designs the same: once it backs up and clogged one pipe, it will slip wrong elements through the other output.

Reason is simple: If the first valve-pipe after the sensor is clogged, it will pass on in the same pipe to the next valve even if the sensor is OPEN the clogged valve and it should go through but cant-> it moves on-> there is the next valve which is now still closed BUT it will stay on the "in" of that shutoff until it will open due the element sensor is 2 tiles earlier and will as soon as it register one for that shutoff open it no matter whats currently waiting on the inline of it.

 

Only solution to make it very safe is add an additional layer behind on the pipe where you want your desired gas/liquid with a check sending everything back into a loop with a priority in the pipe so the flow into the loop is stopped- if something slips through. Though there better is no power blackout or this could fail too however its very unlikely with that setup. Your power usage will be with that 30w (still a good less then 120w from the filter).

[0xFADE]

Sounds like multiple mechanical filters are still the best if these have similar backup issues

[wachunga]

Sounds more like people don't read threads. Look up 8 posts.

[0xFADE]

I can't claim I don't heavily skim content. Loop backs are going to clog as well.  Still mechanical filters require no refined metal or power. 

I see that clogging the system that way is the point. It is a nice solution. 

[wachunga]

I can't conceive a way that that loop will clog in normal game play. If you deliberately go out of your way to clog it, then sure I guess.

Mechanical filters are gamey and not trivial in understanding or in priming/changing. People choose not to use them. If they work for you, great. But don't spread misinformation about their alternatives just because you can't be bothered to read a thread. There's more than enough misinformation on these forums for that very reason.

[0xFADE]

I can dismiss something while emphasizing something else just as well as you apparently.  I would argue your loopback design is much more non-trivial than a mechanical filter that by it's nature only needs to be primed once.

[nivodeus]

On 5/19/2018 at 10:32 PM, wachunga said:

Loop back excess resulting from a clog with another sensor/shutoff. Prioritize the loop by having your input bridge into it. Both sensors are set to the gas you want filtered, second sensor runs through a not gate to the second shutoff. Left is input, middle is selected gas, and right is everything else. You can do the same with just bridges, but it's uglier and bulkier. 

Also as Dude42 and Midnight said the OP's design is flawed as the shutoffs are after a T.

 Santhor, do you mean like this one poster by wachunga? 

I tried the inline method posted by Leonides, it works fine so far. But I am interested as in what could possibly clogged the system? Power outtage or could inconsistent pumping also caused clog? I mean when you put atmor sensor on the pump so it doesnt pump the materials all the time. 

 

[PhailRaptor]

@nivodeus I believe "clogged system" is another way of referencing a pipe that has backed up from, for example, an overpressure vent.  The packets can't exit the pipe, so they build up like cars at a red light, until they reach the filter.

[Santhor]

Exactly, if something cannot exit the vent due overpressure/or not being used by the machine you're delivering or whatever else it starts building up in the pipe until past to the shutoff starting the the problems. You can minimize the risks up to a point where its really negligible that it goes wrong with a loop and if you look for no power outage.

The difference my design to wachunga's would be i use 3 shutoffs and double check the filtered pipe where its important for me that nothing wrong is deliverd with the second sensor and build the loop there sending back anything wrong. It is as said very very unlikely that something wrong is deliverd with that filter and costs 30w power.

[nivodeus]

@Santhor do you mind post a picture or screenshot of it? Do we add priority to the main line ( i.e. our main line in which we our desired packet to be sent to) or to the loop? It should be the main line right, otherwise the packets will just keep looping. 

So it does look similar to wachunga design only with additional shutoff preventing access of unwanted packets to our main line? right? 

 

[abud]

I'm sorry but I don't understand why not just serialize it normally? It doesn't suffer from clogged or power outage problems, and using same amount of power, 10W for each output.

 

We can make multiple input and output, in my screenshot there is 2 input (from left and pump bottom) and 2 output.

 

I know it take so much space, but we can just centralize it because multiple input/output. And we can make multiple condition, for example by temp other output by liquid type. We can make it branching for even more complex rules. When clogged it will just circling there.

[abud]

Since I changed coolant geyser to oil, I add oil filter and made some modification. Basically we can just stack these filter as much as we need with another 10W power.

 

I cannot edit my previous reply to add this screenshot, probably because not approved by moderator yet. So, new reply it is...

[camelot]

I tried this layout several times and the only problem was it led first wrong element went through. Need some batteries to make this never fail.

[martosss]

First let me advertise my thread, where I also discovered hot water:

There are a ton of explanations, as well as the "worse" version that uses 2 shutoffs(+ another ton of explanations for it).

So, to be clear, I am using Kittensisageek's version of 1sensor+1shutoff - cheap and easy.

Now, where to start ... there's so much stuff wrong here ... sooo much   OK, let's start with some theory:

• If there is a normal split of input pipe into 2 output pipes(no bridges involved) liquids will travel 1 packet in direction A and 1 packed in direction B, alternating.
• The element sensor sends a logic signal for 1sec to the shutoff, but it starts at 0.1(when the packet enters) and ends at 1.1(when the packet has exited).
• Currently shutoffs require to be connected to a working grid, but don't use power, so they cost only the building materials and the "idea" of power.

1. Here's what happens in OP's case, which is VERY bad:
   • 0sec: packet 1 of type A comes to sensor => goes in a RANDOM DIRECTION(actually UP Down Left Right repeat, just tested), sensor doesn't work yet.
   • 0.1sec: packet 1 is traveling in it's "random" direction, but the sensor activates and correctly opens/closes both shutoffs
   • 1sec; packet 1 arrives at the end of the line, where the shutoff is(remember, packet 1 is of type A)
     • if 1 => B, it will stop and wait until the the shutoff opens, so this packet goes in the wrong direction for sure.
     • if 1 => A it goes in the right direction - no worries.
     • In both cases valves will open at A and close at B.
     • In the same time packet 2 arrives but the sensor hasn't updated the signal yet
     • Packet 2 sees A open and B closed, so packet 2 goes in direction A, regardless of packet 2's type.
   • 1.1 sec: Sensor has updated shutoffs so now they correspond to packet 2's type.

As you can see, sensor updates with a 1 packet delay, so every time packet types switch, but the first packet went in the right direction, the next packet that is wrong will go in the wrong direction. This makes this filter totally worthless, as it's actually not filtering.

Here is a simple test - both gases going in both directions - colorful, but not what we want!(I tested with cutting pipes directly in sandbox mode, as well as stopping  pumps with atmo switches - stopping 1 pump and letting the other flow causes mislabeling)

BTW in this setup gases will flow first up, then right, then alternate until no more gases in the pipe from the green bridge(beginning of this pipe) up to the shutoffs. After that when flow starts again, they will again flow first up, then right and so on. That way you can figure out which case is bad for you, but you'll definitely have a bad case, which you don't want with filters - you want them to actually filter!

• 
• My advice - DON'T USE THIS IN ANY CASE, unless you want 2 things to flow in both directions randomly.
• I won't even discuss power failure/clogging cases here

On 4/28/2018 at 1:24 AM, Kabrute said:

so its basically either one way or the other, the not gate ensures that its always one and not the other, it would be me space convenient to lay it behind the ladder to the left of the setup and just wire it direct but this shows it more obviously

I'm sorry, but ... LOL at this, did you test any of it, or you just "thought" what might happen? When I was building this for the first time, I might also have thought the same, but .. the reality with the sensor and its 0.1sec delay is very grim.

On 5/19/2018 at 8:15 PM, blash365 said:

This does not work if either of your outputs has a "jam".

In that case the valve will open (depending on sensor), but the flow will follow the exit that is currently "unjammed". It works in both ways and can create unwanted packets in your otherwise well-filtered stream.

The op's version might allow you to avoid this jamming effect.

The OP's version doesn't work, so ... nop, it doesn't allow you that.

On 5/19/2018 at 8:45 PM, MidnightSteam said:

They both failed the clogged pipe test.

One advantage koningkrush's build has is that it won't let any gas pass if there's a power outage.

That's its only advantage as it doesn't generally work.

On 5/19/2018 at 10:32 PM, wachunga said:

Loop back excess resulting from a clog with another sensor/shutoff. Prioritize the loop by having your input bridge into it. Both sensors are set to the gas you want filtered, second sensor runs through a not gate to the second shutoff. Left is input, middle is selected gas, and right is everything else. You can do the same with just bridges, but it's uglier and bulkier. 

Also as Dude42 and Midnight said the OP's design is flawed as the shutoffs are after a T.

THIS is really cool! That works properly 100% of the time. It stops only with a power outage and never lets the wrong element. Kudos!

On 5/25/2018 at 9:35 PM, 0xFADE said:

I can dismiss something while emphasizing something else just as well as you apparently.  I would argue your loopback design is much more non-trivial than a mechanical filter that by it's nature only needs to be primed once.

The mechanical filter uses an "assumption" that the bridge has precedence over a normal pipe, which is not trivial(I'd argue exploit, as bridges and normal pipes should have the same precedence). Wachunga's version doesn't use this hidden preference, it just uses logical elements and is still 100% failproof.

@abudNice work, similar to @wachunga, but for liquids. If you put the 2nd sensor between the 2 shutoffs and remove the bridge before it, you'll get exactly wachunga's version.

Alright, enough filtering!

[Gurgel]

On 4/27/2018 at 9:49 PM, Kabrute said:

the ops design is specifically for anyone who experiences backups in their system, it will never fail afaict

It probably does exactly the same as a "real" filter does internally. 

[Kabrute]

testing shows the second valve should be placed after the filter valve in the pipeline assuming you choose to use 2 valves in the first place.  A single valve and sensor will work fine to keep a particular line clear of all other gas so in that regards going simpler may in fact be better.

tldr: I was wrong before

[martosss]

1 minute ago, Kabrute said:

testing shows the second valve should be placed after the filter valve in the pipeline assuming you choose to use 2 valves in the first place.  A single valve and sensor will work fine to keep a particular line clear of all other gas so in that regards going simpler may in fact be better.

tldr: I was wrong before

Actually using 2 shutoffs is best, but not like the OP. Look at wachunga - he's got the proper solution. It also uses an extra sensor and some logic, but never mislabels, which should be the most important factor in choosing a filter.
And if you are OK with sending everything to 1 pipe as long as the flow doesn't stop, you can still take the "cheap" 1 shutoff version and put the filter on the pipe that you want to stop flowing if electricity stops.