Scrubbers are very slow. They take 5+ minutes to filter or siphon a 5 x 10 room

[juliangiebel]

Description

Scrubbers work very slow, even when set to wide net. I timed them with my phone on salterns upper arrival section with closed firelocks. They take around 6:30 minutes to filter out 11.0753 mols of Plasma and around 6:65 to reduce the same room to 3kpa from 100kpa.

In both cases the scrubbers where set to wide net.

A room of that size shouldn't take more than a minute or two to fully empty out.

Also removing gases completely from a room without using the air alarms replace mode is neigh impossible because the scrubber gets clogged with atmosphere that doesn't contain any gas that could be filtered. This should be taken into account by setting the default thresholds on unwanted gases to at leas 1% from 0.

Screenshots Before filtering:

After filtering (took 6:30):

After siphoning (took 6:59):

[ike709]

IMO it should take a single scrubber like ~2.5 minutes tops to siphon a 5x5 room to near-vacuum from STP and the other modes should be derived from that baseline.

[juliangiebel]

IMO it should take a single scrubber like ~2.5 minutes tops to siphon a 5x5 room to near-vacuum from STP and the other modes should base their balance off of that.

With or without wide net switched on?

[ike709]

With or without wide net switched on?

Can't remember if panic siphon enables wide net or not.

[vulppine]

With or without wide net switched on?

Can't remember if panic siphon enables wide net or not.

currently does not

[ElectroJr]

A room of that size shouldn't take more than a minute or two to fully empty out.

IMO it should take a single scrubber like ~2.5 minutes tops to siphon a 5x5 room to near-vacuum from STP and the other modes should be derived from that baseline.

What do you mean by "near vacuum" / empty out? The main problem is that the siphoning just leads to an exponentially decaying pressure, where the rate of pressure decrease is proportional to the amount of gas present on the scrubber tiles, where every update they can scrub 200L of the 2500L on a tile (so ~8% of the gas on accessible tiles). Going from 100 to 10kpa will take about as long as going from 1 to 0.1 kpa. So the definition of vacuum here matters quite a bit.

E.g for the saltern area that was originally mentioned the pressure goes as:

Currently it takes ~60 seconds for the pressure to drop low enough for people to start suffocating. If you want an atmos set up that could take the pressure in that room down to below 1kpa in ~2.5 min using only scrubbers, then it would only take ~15 seconds for people to suffocate. That's probably long enough for people to notice and put on an O2 mask?

In terms of vacuuming a room faster its better to use a pump that moves a constant number of moles every update, or one that causes a constant pressure change, rather than filtering a constant volume of air. Gas vents currently target pressure changes, and for that reason are actually much better at vacuuming all the air out of a room than scrubbers. Though having a pump that magically results in a constant pressure change is probably not very realistic. There should probably just be different pumps that work well at different pressures. I.e., atmos should probably get a portable vacuum pump that works well at low pressures or something like that.

[juliangiebel]

I was talking about reaching a pressure of ~3kpa as I'm fully aware of that problem. That's why I specifically ran the stopwatch until it reached 3kpa.

[juliangiebel]

People have enough time to put on O2 masks after an explosive depressurization (not counting the fact that you are incapable of protecting yourself from pressure damage in the slightest but that is going to be fixed afaik). So I think that is not much of a problem as poisonous gasses (or the ensuing plasma fire) is also really dangerous.

[Partmedia]

"Low enough" should be the pressure at which hazardous gases (e.g. plasma and tritium) are no longer hazardous (i.e. lightable with a torch).

The minimum concentration of plasma and tritium that is lightable is defined in Resources/Prototypes/Atmospherics/reactions.yml to be 0.01 moles. So, the scrubbers in siphon mode should bring the reference 5x5 room down to 0.01 moles in ~2.5 minutes. Alternatively, we can consider raising the reaction threshold.

[Partmedia]

In scrubbing (not siphoning) mode, things are even worse.

The recurrence relationship for scrubbers working in scrubbing mode is:

$n[t+1] = n[t] - \frac{TR}{CV} \frac{S}{N} n[t]$

Where:

n is the number of moles of bad gas to be filtered TR is the transfer rate, 200 L CV is the cell volume, 2500 L S is 1 for one scrubber, 9 for one scrubber in widenet mode, etc. N is the number of tiles, 25 in our hypothetical 5x5 example t is the time in seconds

This says, every second, sample TR/CV moles of gas (see GasVentScrubberSystem.cs:106) from S tiles of a total of N tiles.

Factoring the expression then rewriting it yields:

$n[t] = n[0] (1 - \frac{TR}{CV} \frac{S}{N})^t$

Where $n[0]$ is the initial number of moles of bad gas. The following figure shows this plotted for $n[0]$ = 1 mol for different S on a semi-log-Y scale:

The plasma concentration starts at 1 ( $10^0$ ) mol. Note that the lightable threshold is 0.01 ( $10^{-2}$ ) mol. This leads to the observation that:

• It takes nearly 25 minutes for a single scrubber to bring plasma down to a safe level.
• With widenet on, it only takes 3 minutes. However, air alarms are not mapped so few people will be using widenet.
• Remember, this is for a reference 5x5 room. Most stations have far less than 1 scrubber per 5x5 area.
• This is for 1 mol of plasma. A gas leak has more than 1 mol of plasma [citation needed]. A full plasma canister has 1800 moles of plasma.
• How many minutes does it take a single person in a station with 50+ people to light the plasma?

This is why CE with an RCD is much more effective.

I would suggest:

1. Make widenet on by default until air alarms can be mapped.
2. Increase the lightable threshold of plasma from 0.01 to 0.1.

[juliangiebel]

@Partmedia Don't forget that scrubbers on scrubbing mode get clogged by gasses they aren't filtering because the atmos tiles next to the tiles that are getting scrubbed dont get woken up when scrubbing, only when the amount of moles changed.

This essentially means that say oxygen and plasma flows into the tiles that are getting scrubbed when that tiles moles lower. The tiles will contain more and more oxygen than plasma and slow down the rate plasma is getting filtered out of the room.

That is the reason for the 1% threshold but even before reacing that the clogging slows down the rate bad gasses are filtered out at

[UbaserB]

it got buffed a bit ago so i assume its fixed?

[MACMAN2003]

they're still a bit slow

[juliangiebel]

Scrubbers should be benchmarked again before claiming the issue is fixed