More Atmos Suggestions

[Shadeykins]

 

1) Passive pumps should be... Passive pumps.

 

Currently passive pumps are useless because they cap at 4,500kPa effectively doing the exact same thing as a regular gas pump.

 

Passive pumps should act much like the passive vent, which is to say they should be used to direct passive flow without pressure caps.

 

Remove the 4,500kPa cap, and make them operate at the same speed as passive vents. This would give atmospherics three distinct options, 4,500kPa, 200l/s, and passive.

 

2) Fix digital manifolds, passive vents, and pumps.

 

This could probably be slung into a bug report. Digital manifolds spawn uvents when selected, and passive vents don't work. Pumps stop working at random in rooms that have had the environment power toggled. Tully can confirm that weird things happen in the turbine room because of this, all the time.

3) Make air injectors run off air alarms.

 

Air injectors in their present state draw from environment power, meaning to turn them off you need APC Access.

 

Atmos Techs do not have APC access, and therefore can't disable or remove their own air injectors.

 

They are also a giant pain in the butt to turn on in their first place, and should have a manual option of enabling and disabling rather than the current signalling nonsense they use.

 

4) Allow 'doubling up' of pumps, or remove double up'd pumps in atmos altogether.

 

Doubled pumps cause some weird bluespace shenanigans and don't work. Want to make a line not draw anything? Just add in another pump!

 

A great example is the turbine gas line, which doesn't actually work properly due to the presence of two pumps, one near the outtake for the gas tank, and one in the turbine room. This is probably a bug, but there's no real reason to have two pumps in the first place - if you want 'throttle' control in certain lines they should be hooked up with a pump placed somewhere sensical, and valves.

 

5) Remove the Turbine tank.

 

It's not really necessary, and it actually reduces efficiency by capping what can be put in the line at 50L/s, rather than 200L/s.

 

I have yet to see someone serious about the turbine make use of this tank, it's always cut out from the line.

 

6) Fix overpressurization, and add a space line to the scrubbers network.

 

Yesterday we had someone get griefy with N2O and air alarms, by setting a few alarms to 1,000kPa. The end result was that the scrubbers line had so much pressure in it, that it was ineffectual. It was working at 4,5000kPa per tick, and in 20 minutes the pressure didn't visibly reduce (The metres and all forms of tracking cap at 9,000kPa) from the metre reading of 9,000 despite the oxygen tank rising by 40,000 kPa.

 

In my experience, the more pressure in a pipe the more ineffectual intake becomes. At this point in time, there was likely over 100,000kPa in the scrubbers pipe - not good. There should be an passive vent outlet into space controlled by a valve for the scrubbers pipe. This presently isn't doable because vents don't operate without being under the control of an APC, and the area outside of atmos' windows in terms of mapping is not a designated area for any room.

 

7) Add in high-volume metres.

 

Our regular metres cap at 9,000kPa, as do PDA gas scanners. Add in other metres that can detect volume above this amount, or just change the existing metres/scanners.

 

8) Speed up LINDA by capping it's calculations by the power of a thousandth, rather than ten thousandth.

 

In layman's terms, LINDA calculates atmos to .0001 decimal places - at least from what the in-game scanners and metres suggest.

 

.0001 pressure is irrelevant in mechanical terms and impact, and I assume only exists to measure the effect 'breathing' has on the atmosphere of a room.

 

Force LINDA to only calculate to .001. There would be a huge amount of net savings in terms of computational weight. Assuming our common pressure range runs from a thousandth, to a ten thousandth the savings are somewhere in the range of 1/8th, or 12.5%.

 

That is of course assuming that the ten thousandth decimal place is factored into all calculations (Which it by all means should be in the vast majority of cases, unless there's some weird voodoo going on).

 

If this affects breathing (As breathing may or may not calculate in the thousandth decimal place), simply make breathing cost more. The net effect breathing has on atmos is so incredulously small that it's irrelevant anyways - the only potential knock-on effects from this is that oxygen/nitrogen/plasma tanks may run out at a slightly faster rate.

 

This is easily offset by changing their default values.

 

Feel free to add on to this, or discuss.

 

[Mrs Dobbins]

 

Passive pumps are designed for shifting precise volumes of gas. Its for precise mixing, not shifting massive amounts of gas.

The "turbine" tank is a general mix tank which has been in atmos since the dawn of time. Hell, I used to use it in my cooling setups to tank gas so I could let it into a cooling loop slowly to not flood it and break it.

I do support making breathing use more air (and LINDA not being as shit) as it dosen't feel dramatic enough to be stuck with limited oxygen, as unless you have an emergency tank, you're set for air for the whole round.

As a long term atmos tech player, this is a thread I would've made with basically the same points so all of my support.

 

[FalseIncarnate]

 

1) Passive pumps should be... Passive pumps.

 

There is no such thing as a passive pump. You probably mean passive gates, which can be thought of as pipes designed to limit the flow of gas to a specific pressure for use in precise mixing or to decrease the pressure in a line after a certain point.

 

2) Fix digital manifolds, passive vents, and pumps.

 

If by digital manifolds you mean the digital switching valve, I have not seen any odd behavior with them. I will need to double check to make sure things weren't accidentally messed up, but as of the last time I dispensed and used one, it was the right pipe. I haven't messed with passive vents recently, so I can't say for sure what may or may not be the cause on that. As for pumps running off of environment, this is likely intended, but I am confused if you meant that once the power comes back on that they do not work again unless replaced, or simply that they stop while the power is out.

 

3) Make air injectors run off air alarms.

 

This would not be a bad one, though air injectors are meant to be used with the automation, including them on the vents tab for Air Alarms so that you can manually adjust them would be nice.

 

4) Allow 'doubling up' of pumps, or remove double up'd pumps in atmos altogether.

 

The example you gave is actually not a bug or issue, but rather just another step that you could easily repipe away. The two pumps do have valid uses, especially if you are planning to do any sort of mixing or re-use of the gases in the turbine room pipes.

 

5) Remove the Turbine tank.

 

As stated above, this is not actually the "Turbine tank", but rather the general purpose "mix tank". I have seen this used in the past for everything ranging from holding super-cooled firefighting mixes, mixing vox-safe anesthetics, plasma-flood failsafe collection, and even Buzzsky's prison. It is best to set this up mixing a bit in advance of when you begin pumping out of it, as it can hold quite a lot more gas than pipes or canisters can.

6) Fix overpressurization, and add a space line to the scrubbers network.

 

7) Add in high-volume metres.

 

This wouldn't even be necessary, we'd just need to ensure the meters can read higher values. Good suggestion though, shouldn't be too hard.

 

8) Speed up LINDA by capping it's calculations by the power of a thousandth, rather than ten thousandth.

 

Breathing actually works off of the number of moles of gas in the air you are breathing (which is why even a pure oxygen tank will suffocate you below 16kPa), so the pressure may not make a large difference. Further, I'm not sure if the advantage of capping the decimal places would really be an advantage in processing power since we'd need to then round the previously capped values to the new cap (taking a bit of processing power). Even if we assume the cost of the rounding to be 0 (negligible), the difference will likely be quite small and otherwise indiscernible to the players.

 

[Shadeykins]

 

1) Fair enough, but couldn't you use gas pumps for the /exact/ same reason?

 

2) Passive vents don't work at all presently, digital manifolds when selected at the dispenser dispense uvents, and pumps can stop working entirely if you toggle environment power. These are all things that were recently broken (expect for the digital manifolds).

 

Example:

 

I set up a turbine line, and rigged a pump connecting two pipes together. I had hacked the APC to toggle environment power to remove the air injector. The pipe before the pump had 4,500kPa. The pipe after had 0. I turned the pump on, at 4,500kPa. Gas did not move through the pump.

 

I replaced the pump with a volume pump. Nothing.

 

I replaced the pump with a passive pump. Nothing.

 

I toggled environment power again. Nothing.

 

As soon as I added a straight pipe to connect the line, gas went through. The only variable that I can really think of is the toggling of environment power before laying the pumps down. This is something recent, and started happening at the same time passive vents broke.

 

5) Alrighty, that makes a bit more sense.

 

8) The return won't quite be 12.5% as some aspects of LINDA are not going to be solely computational. The return in efficiency is likely smaller, but with that said we've got no real reason not to make the Atmos system as efficient as possible. Whether or not the degree of efficiency is negligible, we should still try to make as many savings as we can - small steps add up eventually.

 

"Rounding" of places occurs anyways regardless of the number of decimal places you have (unless the decimal places are zero), what this does is effectively remove an extra number from processing. Instead of rounding at the hundredth thousandth place, you round at the ten thousandth. Depending on how Dreamweaver spooky script works, rounding may not even occur at all - much like in C++ where nothing is calculated beyond the defined decimal place. (It simply doesn't calculate beyond the number of decimal places defined. No rounding occurs.)

 

5.574 is easier to compute than 5.5743, even if just marginally. As I said, it would drop computational weight by 12.5% at best - which isn't necessarily an uptick in performance by 12.5% overall (The actual net savings in performance are probably under 1% considering how fast a processor can do math). I just see having a ten thousandth place as being moot in terms of game mechanics, and needless computation.

 

I'm not sure how Dreamweaver spooky script works, but in most scripting languages defining the amount of decimal places you calculate is really, really simple and generally involves changing one number, or adding one line.