Some feedback on airlock mechanisms

[dfountain62]

TL;DR Basically I am writing this to give an overview of the hard limits to the current system regarding airlocks. These limits are mainly centered around the 1 gas type per pipe tile. If players are to be able to make good airlocks this needs to change in some way or form. It drastically limits the power of a gas pump in terms of useful work being done while forcing the pump to consume quite a bit of electrical power. This makes mechanical (i.e. not fluid obstruction) airlocks almost always worse than u-bend fluid-displacement locks. Other issues are the gas element sensors being unable to differentiate between the presence of an unwanted gas and a vacuum, vacuum being slow to achieve and duplicant checkpoints being late to research and requiring dead-space on either side of the lock. 

This mainly comes from extensive testing of different airlock designs that don't involve fluid displacement of gases (i.e. water-locks). There's 3 main ways to go about making these: vacuum based, non-vacuum based and a hybrid of these two. Why might one want an airlock that doesn't rely on a u-bend? Stress reactions from being wet, ease of moving the airlock (no need to drain the water), being somewhat more realistic, speed, setup time (esp for Co2-locks) , temperature insulation, working in space exposure, etc. 

• Vacuum based locks (3rd lock in image):
  • Summary: Works by having a pump inside a small (2x3 or 2x2) room with automation checks to prevent dupes from entering until the room is full. 
    • Pros: Prevent thermal transfer, prevent any gas leakage (both ways), no sopping wet or wet feet.
    • Cons: Slow, require duplicant checkpoints or locked doors the latter of which breaks navigation if gas leaks are to be prevented, Dupe checkpoints need to be set back from doors to allow them to close on auto, requires a lot of pump time, suffocation risk, requires a lot of space. (Problem 1 see below)
• Non-Vacuum based locks (2nd lock in image):
  • Summary: Works by running a pump which filters unwanted gases (to be vented to the outside) from wanted gasses (which are kept inside the lock or on the desired side of the lock)
    • Pros: very good at keeping gas out, no sopping wet or wet feet, very fast for dupes. Works in a 2 high hallway (or 4 high if insulation is desired). 
    • Cons: Not very good at keeping gas in (i.e. is a one-way lock), especially if outside pressure is low. Allows for temperature changes
      • Requires one of the following:
        • Pressure plate under outside door which has very high thermal conductivity. Either requires raising the lock (which slows dupes) or dealing with the thermal loss/absorption from whatever is below the lock. Is open-loop meaning you have to trust that running it while the door is open + after it closes for X seconds is enough to keep the gas out.  
        • 2 pressure sensors. Since pumps can only pump one gas at a time this decreases the efficiency of the lock by increasing the size of the room which increases the different types of gases that can leak in. More types of gases leads to some types of gases bypassing the lock once the gas pump's output pipe is blocked. Requires additional automation and is prone to a vacuum infinite loop. (Problem 2 and 3 see below). 
• Hybrid lock (1st lock in image): 
  • Summary: Basically the best of both worlds but has a downside of needing 2 pumps and a larger footprint. Works by having a vacuum on the outside of the lock to dramatically reduce leakage from the one-way lock. This is as near as can be got a perfect mechanical airlock and I would say it works pretty well.
    • Pros: Works both ways well, no sopping wet or wet feet, very fast, near-perfect thermal insulation, very low loss of internal gases. Internal pump has low duty cycle. 
    • Cons: High peak power usage due to second pump see problem 2. Much larger footprint (2x10), automation is a pain just like in the non-vacuum lock, some gas loss from internal space to external space. 

Problem 1: Dupe checkpoints come in late at a tier 5 research and the setback distance creates dead-weight. 

Solution 1: Dupe checkpoint should come earlier (tier 3). Ideally should prevent doors from opening if a dupe is standing next to a door in the checkpoint, though I imagine this type of exception is difficult to get right. 

Problem 2: Because of the way gases and pipes work, it's very difficult to take advantage of the full power of a pump when there are multiple gases in a room. Similarly when filtering gases there is a problem with pipes getting backed up with multiple pumps because each segment of pipe only holds 1 type of gas.  This leads to a gross waste of energy and a need for higher peak power consumption. It also leads to gas leakage once a pump gets backed up due to "pipe blocked"

Solution 2: It would be nice to have pipes that work more like reservoirs and hold up to 1000g of any combination of gasses. Gas pumps already seem to absorb 500g of anything in the lower-left cross area so pipes need to combine up to 1000g of anything per segment. Even having multiple outputs for a pump would help in this department though that just pushes the problem down the line to whenever they need to be filtered. Since we already have a gas filter for filtering out a certain type of gas there is already a way to separate the gases. Unfortunately vents would need a rework (maybe be size 4 and be in the same layer as gas pipes) since outputing them the same way reservoirs do (1 type at a time) would lead to a bottleneck and pipes backing up just like before. Also why are vents not in the gas pipe layer? They obstruct so much when dealing with airlocks.  

Problem 3: Gas element sensors are difficult to use next to pumps because the pump has it's input as a cross centered on the lower left corner of it. This leads to a vacuum in the adjacent bottom-left tile and left-bottom tile. When you are trying to detect the absence of a certain gas (like o2) this leads to an infinite loop. A vacuum is usually not a problem the presence of anything not O2 is. Correcting for this leads to asymmetry (a layout works differently than it's mirror image).

Solution 3: Gas element sensors should have a setting for to set their behavior in a vacuum. 

Another suggestion Things that would make this stuff easier and more approachable: Gas and liquid drains (basically the inverse of a vent). They require connection to a pump which sucks in fluids from the input and delivers them to the pump. Due to the already difficult to understand nature of pipes, bridges, shutoffs, etc this should probably be it's own type of pipe/input (gas/liquid drain line and drain input/output) and it would be nice to make explicit the tiles which are "drains" on current pumps with the same indicator. This separates fluid lines into "incoming" and "outgoing". The benefit to this is that the suction area can be made A) explicit on current pumps and B) can be customized (see 2nd picture; imagine the vents were drains). Another benefit to this system: currently capturing certain gases requires moving the pump to where it is located. However it would be nice to simply re-adjust the input drain location instead.

 

[Yunru]

You could just have a CO2 u-bend.

The only reason to use a vacuum is to isolate heat, in which case you'll want envirosuits anyway.

[Greybear]

I almost feel like c02 and water u-bends are cheats, exploiting the weirdness that comes with the one-gas-per-title system. I would also love for the mechanical air locks to be a better option, but gas pumps are so cumbersome and ineffective.

[Lawnmower Man]

Drains would be OP.  Currently, pumps are a "tax" and a *tech gate* to get fluids into pipes.  Pipes themselves have no overheat temp, so the only thing stopping you from shipping magma around your base is having a pump that can survive 1000+C.  I guess you could specify that when a fluid hits the pump, it would cause overheat damage.  But if you could "pull" fluids from the other side of the base, you could possibly get them to their destination without ever hitting the pump itself!

But pumps are also an intentional space constraint, limiting the layouts of designs.  So, I think they would be great, but they would probably damage game balance too much.