What's wrong with this germ killing contraption?

[vasil78]

Hi

I'm trying to create a contraption that kills germs in water. What I came up is this:

The room is filled with chlorine.

Germ-infected water comes in via the shutoff on the lower left and goes into the containers. 

When water leaves the containers it is checked by a germ sensor that sends a green signal when there are less than 0 germs in the packed. If this is true then the shutoff on the right-hand side is opened and the packet goes to the clean water storage. Otherwise the shutoff on the upper left is opened and the packet goes back into the containers. 

The automation logic looks like this:

 

The shutoff on the lower-left is there to prevent blocking the flow if both containers are full and a packet with germs are in the sensor.

 

Still setup leaks germ-infected water from time to time. I don't understand why however :(

Any ideas?

 

Best regards 

Vasil 

[TripLykely]

What’s the bottom middle input for?

[psusi]

The reason it leaks germy water is that when a packet of germy water hits the sensor, it flows to the input cell of the shutoff before the shutoff actually closes, and then it is stuck there until clean water hits the sensor and opens the shutoff.  For filtering, your shutoffs need to have an overflow path so liquid does not get trapped in the pipe.  You also only need one shutoff and a valve rather than 3 and all of that logic.  You just wrap the pipe past the input of the shutoff back around to the input of the tank so that if the valve is closed, the flow automatically recirculates.  The input pipe just needs to enter via a pipe bridge so that the recirculating water has priority.  Have a valve limit the amount that is allowed to flow back ( half or 2/3rds max should be fine ) so that some new water always has room to enter, otherwise the tank will never fill up.

[vasil78]

34 minutes ago, TripLykely said:

What’s the bottom middle input for?

It's used to manually drain the system in case it locks up.  

[Urist McPilot]

The T-junction under the germ sensor is not controlled in any way by the sensor, it sends packets left and right in a round-robin fashion (provided the next pipe segment is empty) regardless of germ count. Sometimes it sends a germy packet to the right, which has no way to leave that segment; next time a clean water packet arrives at the sensor, it opens the shutoff, and the waiting germy packet leaves.

Solution is to let the germy packet leave if the shutoff is closed: build a pipe segment leaving the input of the shutoff towards another input, and all germy water will go that way, making space for the incoming clean water packets that will be let through the shutoff. The left hand side of the T-junction can be completely removed.

The germy water stream still needs to be handled, but beware: if you block the pipes that drain the germy water, you get into the same situation, germy water that cannot leave the shutoff input will be let through if a clean packet comes afterwards. If you just route the germy water back towards the input, like the left side shutoff does now, the merging junction on the left edge will block the pipe. If you use a bridge section to give priority to the germy water, you will have problems with filling the reservoirs. My usual trick here is to mix the input into the output from a reservoir, and connect the germy filtered water to the input of the reservoir; this way the reservoir output gets blocked from time to time, filling the reservoir slowly without blocking the germ filter.

[vasil78]

Thanks!

I was never aware that the shutoff blocks the output, not the input  . 

[Urist McPilot]

3 hours ago, vasil78 said:

I was never aware that the shutoff blocks the output, not the input  . 

When you  have a loop from the output to the input, the two are the same

[psusi]

6 hours ago, vasil78 said:

I was never aware that the shutoff blocks the output, not the input  . 

It doesn't block the output or the input.  It either moves packets from the input to the output, or it doesn't.  If it doesn't, then the fluid remains in the pipe and can continue flowing in search of another input port to take it.

[Lilscratchy]

Honestly the most reliable way i found is one that requires a bit of extra setup water. 

You chuck down 3 reservoirs and fill them with water. Of course you put them in a chlorinated room. To be on the safe side i make sure that once theyre all filled i remove any water stuck in pipes as they back up while filling. 

Now when theyre all filled wait for all the germs to be killed off. Any packet of water that enters the line of reservoirs is going to have its germs basically diluted to such an extend that by the end it comes out the other end the bacteria are dead. While at the other end, already desinfected water flows out. 

The result? You can desinfect up to 10kg/s of germy water if you can feed it into the series.

Keep in mind though you have to keep these three reservoirs full. I usually chuck in a shutoff connected to a 4th reservoir that deactivates the shutoff if its 1 percent filled (50kg) and allows water to flow through if it gets filled past 50 kg. 

I feel this method is the most straightforward, only downside being that you require 3 to gather three reservoirs full of water once. 

Hope it helps

[Fyrel]

Easy set up is to build the tanks on horizontal doors with the germ sensor just before the tank connected to the door through a not gate and a filter gate.

Just need the filter gate to hold the door open long enough for the water to disinfect.

Tanks built on open doors will accept water but won't let it flow out.

[darknotezero]

15 hours ago, Fyrel said:

Easy set up is to build the tanks on horizontal doors with the germ sensor just before the tank connected to the door through a not gate and a filter gate.

Just need the filter gate to hold the door open long enough for the water to disinfect.

Tanks built on open doors will accept water but won't let it flow out.

This. I currently have three liquid reservoirs in my base built on horizontal doors in a chlorine filled area. (Forgive that gas pump in there, i was using that to eliminate some polluted oxygen that accidentally got in that area and was too lazy to take it out.)

Spoiler

and here's the automation layout. Two of the filters are set for 115s, the third one is set for 0s because that one is flowing from carbon-skimmer PW which I was not aware of until after the fact produced no germ side effect.

So water sieves convert PW to W until the reservoirs get full, sieve turns off, filter waits 115s before closing door, once door closes clean water exits until empty. once empty, door opens and sieve turn back on, cycle resets.

Signal switch to OR gate allows me to keep that Sieve which is the no-germ water on indefinitely if i want, which i turn on if my water reserve is getting low but turn off to minimize heat production

Spoiler

 

[nakomaru]

I'm sorry to bring this up every time, but you guys are already using three nearly full reservoirs in chlorine... As long as you are only using one element, there's no automation or doors or sensors needed. Just piping. And that gives you 10kg/s flow.

Only the bottom right 6x3 room is chlorine.

[darknotezero]

oh brilliant. I wouldn't have understood why this worked until my recent discovery of PW without germs, i was under the false assumption that the PW would need to be converted to regular W before decontamination was possible. This is much more elegant, will definitely tinker with this in future builds.

[TripleM999]

18 minutes ago, nakomaru said:

I'm sorry to bring this up every time, but you guys are already using three nearly full reservoirs in chlorine... As long as you are only using one element, there's no automation or doors or sensors needed. Just piping. And that gives you 10kg/s flow.

Only the bottom right 6x3 room is chlorine.

I tried this build, as i loved its simplicity... but it has one minor flaw... as long, as the input pipe is full, and this can happen, the output will not pipe back into the system, and after 20kg (2 packets) the "flowback" will overflow. Even, and especially, when the system is not full.

[nakomaru]

31 minutes ago, TripleM999 said:

I tried this build, as i loved its simplicity... but it has one minor flaw... as long, as the input pipe is full, and this can happen, the output will not pipe back into the system, and after 20kg (2 packets) the "flowback" will overflow. Even, and especially, when the system is not full.

Don't know what you tried to do, but the only way this will seize is if the bypass (bottom left) is blocked. It cannot flow back if you add another bridge, like so.

When you want to add extra input (need to convert extra PW), add a bridge on the right, like this.

The septic system takes priority, the feedback loop will never seize, and only when there is extra time, the extra load is processed. The system can only process 5kg/s because of only using one sieve, but adding a second sieve would reach a steady 10kg/s.

[TripleM999]

3 minutes ago, nakomaru said:

The septic system takes priority, and only when there is extra time, the extra load is processed.

The way, the bridges are built, the input, the one from washing room, has priority over the flowback. And three containers in chlorine only kill all germs, IF they are nearly full. If nearly empty, the germy water will mostly flow through. So a condition to make this work is to fill the containers before one connects/opens the outflow to the sieve.

[nakomaru]

2 minutes ago, TripleM999 said:

So a condition to make this work is to fill the containers before one connects/opens the outflow to the sieve.

Yeah, if you want the system to work right you have to build the system correctly and not a different, faulty system. True.

Once there is 15T of water in the reservoirs, it will never leave.

[TripleM999]

1 minute ago, nakomaru said:

Yeah, if you want the system to work right you have to build the system correctly and not a different, faulty system. True.

Even if built right, but not used as meant, it will not work. I do not want to say, it wont work, i only wanted to add, that it needs full containers to work, as intended, empty ones will not do, so it has to be prefilled/used, before fully opened.

[nakomaru]

Just now, TripleM999 said:

Even if built right, but not used as meant, it will not work. I do not want to say, it wont work, i only wanted to add, that it needs full containers to work, as intended, empty ones will not do, so it has to be prefilled/used, before fully opened.

Fully agree. Do not build a faulty system that is superficially similar to the one I am recommending.

[Yunru]

An easy way to ensure the three are full (and thus leaving the system as fire and forget) is to put a door under the 3rd tank, connected to the automation port of the 1st.

[psusi]

1 hour ago, TripleM999 said:

I tried this build, as i loved its simplicity... but it has one minor flaw... as long, as the input pipe is full, and this can happen, the output will not pipe back into the system, and after 20kg (2 packets) the "flowback" will overflow. Even, and especially, when the system is not full.

The whole way you get clean pw out of the loop is because the input pipe is full and so the pwater can't circle back around through the bridges, so instead it exits to the left.  Also once it is full, it can't ever become not full again.

[TripleM999]

26 minutes ago, psusi said:

The whole way you get clean pw out of the loop is because the input pipe is full and so the pwater can't circle back around through the bridges, so instead it exits to the left.  Also once it is full, it can't ever become not full again.

Imagine, what happens, when system is freshly built... and empty. And fully connected on input side and output side. And now put a steady stream of germy water into the input.

Maybe i was stating the obvious... for some. The system works if full, but is mostly not able to fill itself.

[Yunru]

1 hour ago, TripleM999 said:

when system is freshly built... and empty.

So... Built wrong. 

 

If it's not full, it's not fully built. 

[TripleM999]

4 minutes ago, Yunru said:

So... Built wrong. 

 

If it's not full, it's not fully built. 

Built, but not prepared, ok... but built wrong, sorry, but no. I think, this discussion leads to no further knowledge.

[KittenIsAGeek]

11 hours ago, nakomaru said:

I'm sorry to bring this up every time, but you guys are already using three nearly full reservoirs in chlorine... As long as you are only using one element, there's no automation or doors or sensors needed. Just piping. And that gives you 10kg/s flow.

Only the bottom right 6x3 room is chlorine.

This is a very elegant design.  I've set things up similarly, but I never thought of using a bridge  as an overflow valve.  Nice!