New "Free" Gas Sorting

[KittenIsAGeek]

With the recent addition of automation ports on liquid and gas vents, and conveyor chutes, it is now possible to do some quick "free" gas filtering.  For example, here I have an electrolyzer that is directly under one gas pump.  Automation on the pump tells it to turn on if there's Hydrogen gas nearby.  It also is paired with an AND gate that will stop pumping should a gas reservoir fill up.  Yes, this is also showcasing a use for the new automation output on reservoirs.

OK, so here's the filtering part. 

Spoiler

 

The gas element sensor is set to "Hydrogen".  Since a green signal opens the vent, we want the vent to only be closed if there is hydrogen, but open if there is anything else.  This will let oxygen (or CO2, etc) escape, but hydrogen will continue on to the reservoir.

Here's some SS's of it in action:

Spoiler

And a SS of the contents, just so you can see there's no oxygen in here:

Visually, the sensor lights up the instant that the sensed gas starts to move into the sensor's gas pipe.  As you can see from the SS's above, it APPEARS that the hydrogen will get vented but the oxygen kept, but that's an illusion. 

 

Here are some tips if you plan on using this sort of filtering:

1) There must be a straight shot across the sensor and vent.  This shape does not work and you will sometimes get the wrong gas:

2) Your vent can't be overpressure.  Kinda obvious, but if the vent can't release the gas you don't want, then it will continue on and into your supply.  However, if you're capturing gas using a water lock, then you could very easily chain several capture rooms together on a single pipe.

For example:

Spoiler

This room could easily be modified.

Something like this would allow whatever gas to get captured in the room, while letting all other gasses continue past.  I just haven't had time to fix it yet.

 

Anyway, I hope this is a useful idea for others to play around with.

[SamLogan]

I've tried this filter and it let pass some hydrogen sometimes in the gas vent, probably due to the lag or a bug.

[Oozinator]

5 minutes ago, SamLogan said:

.. probably due to the lag or a bug.

[Saturnus]

Just now, SamLogan said:

I've tried this filter and it let pass some hydrogen sometimes in the gas vent, probably due to the lag or a bug.

Yeah, vent sorting isn't perfect like shut off valve filters. I thought it was just a bug in the preview but it seems it's either not fixed yet, or it's not going to be fixed which is why I've not made vent sorter build to present on the forum yet.

However, in places where you don't need perfect sorting or where you just need pressure equilisation that doesn't need to be absolutely perfect as in letting water into steam chambers when pressure gets too low for example the vents do a good enough job on their own.

[KittenIsAGeek]

Strange. I've only had it fail if there was either over pressure or there was a bend in the line.  I'll try and do some more rigorous testing after work and see if I can break it consistently.

[Neovik]

also temprature filter on rail doesn't work, i think he lets through the package before ,seems all buggy to me

[Gus Smedstad]

That's an interesting approach, but not usually what I want. Usually I want my filtered gasses to remain in the vents, I'm just choosing which vents.

For example, if I have a central electrolyzer facility, I need the filtered oxygen to go into my vent distribution system, not just into the local open air. Air pressure hasn't been enough for distribution, in my experience. Maybe if you used high pressure vents limited to 4 kg to avoid popped eardrums, but I've always started ventilation well before I've had high pressure vents.

There are going to be some cases using vents to filter would work. For example, a gas pump at the bottom of the base, mostly meant to capture CO2 and store it, and maybe trace gasses. Dumping any oxygen that gets captured by accident is great, though again you want high-pressure vents or it may fail.

[TheRickestMorty]

You could have the vent be in an infinite storage room, cant do the same bridge magic to avoid a pump for outlines, but if the room is pressured enough might it even kill any erroneous gasses?

[RonEmpire]

I think the biggest issue  / problem with  sensors is that they're not actually 100% reliable.   If you're trying to 'sort' something,  the sensor has to be literally on top of the pipe before the vent or shut off value  white block to actually make it work.     Otherwise  the distance between the sensor pipe to the white input block could leak the element incorrectly. 

It's only slightly useful if the entire elements are flowing in a uniformed increasing or decreasing temperature. 

Ideally, the shutoff value (or vent)  should have its own senors on the actual white input block to really make it work correctly.

The only other way is to figure out the time it takes for the element to reach the white block,  you'd have to use a buffer or filter gate to time it to open and close. But this is also unreliable. 

[KittenIsAGeek]

Update: I have played around with this for a while and it appears to be quite reliable for me. One caveat is that my game is running at around 60fps and I haven't had the pipe section off-screen for any extended periods of time.  As far as I can tell, as long as the vent can release the gas when it opens, then it filters as it should.

[TheDeamon]

On 3/6/2020 at 8:36 AM, Neovik said:

also temprature filter on rail doesn't work, i think he lets through the package before ,seems all buggy to me

I've noticed it seems to share an issue with the aquatuner(when the aquatuner is used with a bypass). The first item on the line at the time of activation continues along on the bypass, while anything else that follows, so long as the aqua tuner/conveyor shutoff remains active, will traverse across the desired path without further incident.

Given the process that should be involved in that particular quirk, the issue is likely present for the other conveyor sensors as well(or even simply using the enable/disable switch for that matter), as the problem is likely to the be a timing issue on the shutoff itself. Which will result in all kinds of frustration for people trying to build a sorter on a conveyor line using a mixed input.

So long as you get multiple "packets" of the same type, every "packet" except the first in a given group should make it through on the intended path. But if you're trying to sort between eggs and some other resource, where you only have 1 egg moving through at a time. Well. It's probably going to fail to properly sort the egg, every time.

---------------------

To clarify, as the above discussed two different forms of the same issue:

Scenario 1: Aqua Tuner or Conveyor Shutoff are enabled and powered on, no additional filtering, with a "bypass" connected.

Sequence is Item, Item Item, no gap in between them.

After running the test, the bypass will have 1 item, and the output for the AT or Shutoff will have 2.

The items can be dissimilar in this case, as the two devices were enabled for the entire duration of the test.

Scenario 2: AT and CS are enabled and powered on and enabled, no additional filtering, with a bypass connected,

Sequence is Item, 1 second gap, Item, 2 second gap, Item.

Expected outcome based on my own experience is that the Bypass line will have 3 items present, the output for the AT and CS will have none. 

Scenario 3: Dropping the AT as it isn't used for filtering things, but the CS is.

Now we have a conveyor rail with two sensors, two shutoffs(connected to their respectively positioned sensor) and a bypass area large enough to handle the unsorted content.

Sequence is now 2 loads of coal, 1 egg, and 2 more loads of coal. Sensors are looking for coal and the egg type being used.

Expected outcome based on my own experience is that the conveyor shutoff for the line collecting coal will have 2 loads of coal. The line that is supposed to be receiving eggs has nothing, and the bypass has 1 coal, 1 egg, 1 coal.

Scenario 4: Basically same setup as Scenario 3, just the quantities have changed.

Sequence is now 4 loads of coal, 2 eggs, and 6 more loads of coal. Sensors are looking for coal and the egg type(s) being used.

Expected outcome is the line for handling coal will have 8 loads of coal, the line handling eggs should have a single egg, and the bypass will have 1 coal, 1 egg, 1 coal.

Scenario 5: We're repeating scenario 3 again, only this time we're staggering the release of the contents so that there is a 1 second delay between each item.

Expected outcome is the coal line has no coal, the egg line has no eggs, and the bypass line has 2 coal, 1 egg, and 2 coal.

[TheDeamon]

20 hours ago, RonEmpire said:

I think the biggest issue  / problem with  sensors is that they're not actually 100% reliable.   If you're trying to 'sort' something,  the sensor has to be literally on top of the pipe before the vent or shut off value  white block to actually make it work.     Otherwise  the distance between the sensor pipe to the white input block could leak the element incorrectly. 

It's only slightly useful if the entire elements are flowing in a uniformed increasing or decreasing temperature. 

Ideally, the shutoff value (or vent)  should have its own senors on the actual white input block to really make it work correctly.

The only other way is to figure out the time it takes for the element to reach the white block,  you'd have to use a buffer or filter gate to time it to open and close. But this is also unreliable. 

Someone else seems to have found a solution for some of the issues, we need to allow for an extra tile of space between the sensor and the shutoff..
 

On 3/6/2020 at 6:11 PM, Crapig said:

signal delayed
move a sensors to the left(1 tile)
use this until it is fixed

 

[Saturnus]

4 minutes ago, TheDeamon said:

Someone else seems to have found a solution for some of the issues, we need to allow for an extra tile of space between the sensor and the shutoff..

I wonder if you can put two sensors next to each other, and then 2 shut off valves next to each other in order to save space.

[TheDeamon]

Probably, just don't cross the wires.

Just be prepared to do a bit of demo/rebuilding if/when they address the timing issue.

[Saturnus]

6 hours ago, TheDeamon said:

Just be prepared to do a bit of demo/rebuilding if/when they address the timing issue.

You'd have to do that in either case.

[TheDeamon]

Moving a sensor is a lot less involved than moving a shutoff valve, sensor, and rail line.

[Saturnus]

29 minutes ago, TheDeamon said:

Moving a sensor is a lot less involved than moving a shutoff valve, sensor, and rail line.

Not if you plan ahead. 

[RonEmpire]

Well that graphic shows things moving in some what uniformed processing. It needs to be tested where you have 1 unit of each item      and see if it actually sorts them correctly back to back.    And if the items are backed up say in a receptical    it won't work.  In the above image they're being removed via a chute.   So if ores are backed up  will it still work toggling between items.  If things are backed up and the ores sit there waiting on the rails things end up getting clogged or completely blocked as temperature changes over time 

[0xFADE]

Nice design but mechanical filters still work for free.  Don’t have to worry about the vent going over pressure or use refined metal but that sure is a simpler design.  Very nice. 

[TheDeamon]

On 3/9/2020 at 6:17 AM, RonEmpire said:

Well that graphic shows things moving in some what uniformed processing. It needs to be tested where you have 1 unit of each item      and see if it actually sorts them correctly back to back.    And if the items are backed up say in a receptical    it won't work.  In the above image they're being removed via a chute.   So if ores are backed up  will it still work toggling between items.  If things are backed up and the ores sit there waiting on the rails things end up getting clogged or completely blocked as temperature changes over time 

Based on the graphic, 1 unit packets should sort properly. Just so long as there isn't a backup that happens on the line(which wouldn't with conveyor chutes), but the system would break down the moment things start to stack up on the line itself.