Steam tower

[mrbunnyban]

24 minutes ago, DonDegow said:

@mrbunnyban

Valves do make thermo-regulators more efficient depending on how you use it; you need to loop-back the output to the input but if you don't use a valve it tends to conflict with gas incoming the other way (valves are one-way only, they're used like electronic diodes in such a setup).

Although the fine tuning is quite heavy, I didn't succeed to have a stable output of LOX, valve helps but having a perfect setup is really hard.

Now that I understand! Control the flow in one direction, got it.

[mrbunnyban]

16 minutes ago, Saturnus said:

Maybe using the gas pump feedback and liquid pump outlet is best as a final stage. And just having feedback looped on the initial stage.

However, the calculation of negative gain is the sum of the flow through the valves squared. So if you have 4 valves at 100% the same throughput as the input valve and one at 10% it's the same as 4.1^2=16.81 inline thermoregulators. EDIT: that can't be right. Let me set it up and check the numbers when I have access to my regular PC later this weekend.

Final stage? The idea is not to have to watch over the o2 being cooled in case it cools too much into a solid. This way, liquid oxygen is never sent back into the loop.

[Saturnus]

After trying a lot I can't get it to work reliably. The game physics isn't stable enough and doesn't allow for gasses to flow as natural as they should. I can get it to work using gas pipe bridges instead of gas valves for the feedback loop but only if I dial down the input flow to about 82g/s and increase the number of gas pipe bridges to 6. However, it's highly unstable. If the gasses don't flow absolutely the right way the whole system instantly seizes up. It might be a case of just getting the pipes laid out correctly but I don't think I can be bothered to error-correct it at the present stage with 5 days remaining for a major update that would invalidate the effort to a large extent.

[MoMuffins]

Even if you've gotten it to work reliably.  The output is 1/64th of the input.  You would still need to duplicate this setup many, many times to get the same throughput as 16-17 inline thermo-regulators.  I do not see how you solved anything.

[Saturnus]

4 hours ago, Fatmice said:

Even if you've gotten it to work reliably.  The output is 1/64th of the input.  You would still need to duplicate this setup many, many times to get the same throughput as 16-17 inline thermo-regulators.  I do not see how you solved anything.

No. The output is pretty much 1:1. Nothing gets lost by doing a feedback loop.

[MoMuffins]

How is the output 1:1?  You limit the input to 100g/s, then limit the final output to 10g/s.  The smallest rate dictates the overall rate.  It doesn't matter that your mass is conserved.  You rate limited yourself to 10g/s of whatever, so you would need 10x of those setup for 1 dupe.

[Saturnus]

5 minutes ago, Fatmice said:

How is the output 1:1?  You limit the input to 100g/s, then limit the final output to 10g/s.  The smallest rate dictate the overall rate.  It doesn't matter that your mass is conserved.  You are rate limited yourself to 10g/s of whatever, so you would need 10x of those setup for 1 dupe.

No. You misread the post. It's the final feedback valve that is set to 10g/s. There no output limiting valve.

Either way. It doesn't quite work in the game. Not because the theory is wrong, I mean, it's pretty basic engineering stuff really, but because the game doesn't model liquid and gas flow correctly.

I can get liquid oxygen just not reliably and sometimes the system freezes as temperatures gets too low (0K).

[MoMuffins]

I don't think I missed read your post.  Perhaps you should draw or show the build, even if it doesn't currently work?

This is my understanding of your description.

           |------------------------------------------------------------------------------|
           ||------------------------------------------------------------------|          |
           |||------------------------------------------------------|          |          |
           ||||------------------------------------------|          |          |          |
	   |||||------------------------------|          |          |          |          |
	   ||||||------------------|	      |	         |	    |          |          |
-(100g/s)->++++++[Thermo-regulator]+-(50g/s)->+-(50g/s)->+-(50g/s)->+-(50g/s)->+-(50g/s)->+-(10g/s)->

or

	   |------------------------------------------------------------------------------------------|
	   ||------------------------------------------------------------------------------|          |
           |||------------------------------------------------------------------|          |          |
           ||||------------------------------------------------------|          |          |          |
           |||||------------------------------------------|          |          |          |	      |
	   ||||||------------------------------|          |          |          |          |          | 
	   |||||||------------------|	       |	  |	     |          |          |          | 
-(100g/s)->+++++++[Thermo-regulator]+-(50g/s)->+-(50g/s)->+-(50g/s)->+-(50g/s)->+-(50g/s)->+-(10g/s)->+
                                                                                           |
                                                                                           |---------->
or

           |----------------------------------------------------------------------------|
           ||----------------------------------------------------------------|          |
           |||----------------------------------------------------|          |          |
	   ||||----------------------------------------|          |          |          |
	   |||||----------------------------|	       |	  |	     |          |
-(100g/s)->+++++[Thermo-regulator]-(50g/s)->+-(50g/s)->+-(50g/s)->+-(50g/s)->+-(50g/s)->+-(10g/s)->


or

	   |----------------------------------------------------------------------------------------|
           ||----------------------------------------------------------------------------|          |
           |||----------------------------------------------------------------|          |          |
           ||||----------------------------------------------------|          |          |          |
	   |||||----------------------------------------|          |          |          |          |
	   ||||||----------------------------|	        |	   |	      |          |          |
-(100g/s)->++++++[Thermo-regulator]-(50g/s)->+-(50g/s)->+-(50g/s)->+-(50g/s)->+-(50g/s)->+-(10g/s)->+
                                                                                         |
                                                                                         |---------->

None of them have an output rate equal to the input rate.  The overall mass is of course conserved.

Even if you had a build where feedback loops gave an output rate equal to the input rate, the amount of work on the gas is still the same for the phase change.  If it took 16 inline workers to produce 1 unit/s, then a worker can produce the same unit if given 16s.  The amount of work is the same because rate of work per worker has always been the same.  Another word, 16 workers on feedback loops would still produce 1 unit/s.  Where is the savings?

[Saturnus]

You're forgetting that the initial pressure is increased by the pressure in the feedback loop. Which feeds back and increases the pressure in the feedback loop, and so on. The very principle of a feedback loop.

I've made a very simple drawing so you can understand the principle of a feedback loop. Basically, it's an infinite series adder where after each cycle half the total input pressure is added to the input pressure so you end up with an input pressure that is double what was put into the system by the input valve.

You're also don't seem to get that in order to feed 16 inline thermoregulator with a constant maximum flow you need about 50 air pumps. And the system can supply oxygen for up to 100 dupes. It will also use about 10KWatt constantly which is basically about 40 dupes doing nothing but run on hamster wheels all day long when you count in sleep, and dinner and bathroom breaks. What I'm basically saying is that you run a feedback loop system so you have an oxygen supply that actually matches the number of dupes you have, and have consistent lower running power draw.

At the end of the day, if you have 100 dupes, there's no gain from a feedback loop over the inline system but that seems highly unlike to ever be the case.
 

[mrbunnyban]

1 hour ago, Saturnus said:

You're forgetting that the initial pressure is increased by the pressure in the feedback loop. Which feeds back and increases the pressure in the feedback loop, and so on. The very principle of a feedback loop.

I've made a very simple drawing so you can understand the principle of a feedback loop. Basically, it's an infinite series adder where after each cycle half the total input pressure is added to the input pressure so you end up with an input pressure that is double what was put into the system by the input valve.

You're also don't seem to get that in order to feed 16 inline thermoregulator with a constant maximum flow you need about 50 air pumps. And the system can supply oxygen for up to 100 dupes. It will also use about 10KWatt constantly which is basically about 40 dupes doing nothing but run on hamster wheels all day long when you count in sleep, and dinner and bathroom breaks. What I'm basically saying is that you run a feedback loop system so you have an oxygen supply that actually matches the number of dupes you have, and have consistent lower running power draw.

At the end of the day, if you have 100 dupes, there's no gain from a feedback loop over the inline system but that seems highly unlike to ever be the case.
 

Upcoming update fixes how gases and liquids combine/are not lost in pipes, so this issue may be fixed next update. Give it a try then.

[MoMuffins]

First, what you said vs what you drew are very different things.

On 3/10/2017 at 6:06 AM, Saturnus said:

Instead you have an intake valve (or several in parallel) setting the intake to the desired flow rate before the thermoregulator. 100g/s is a good example because it's the maximum setting on a air valve, and it's the exact amount of oxygen needed for a regular dupe. Then after you have a series of valve on the output that loops back to the input of the thermoregulator after the intake valve. Since valves are one-way valves that means you force half the stream at this point back to the input. For each valve you double the efficiency of thermoregulator, so after inline 4 feedback valves all set to the same value as the input valve you thermoregulator is running at 16 times higher efficiency. Just remember that the limiting factor here is the 10kg/s that the pipes can transmit as the pressure at the intake is also doubled for each valve. Now you just put on a 5th inline feed back valve and start by setting this to the minimum 10g/s which should be the same as a system running 16.8 inline thermoregulators. You need to tweak it a bit but it should fairly straight-forward to dial in so you can have a constant stream of liquefied O2.

 

1 hour ago, Saturnus said:

 

What you drew is a set of N thermo-regulators each with its own feedback valve?  This set is then connected in series?  Your description was for a single thermo-regulator, as I did not see the use of plural noun, with multiple feedback valves, which is what I drew from reading what you wrote.  If it is the case that you are using N thermo-regulators, then you still haven't solved anything.  Things are still being worked on one after another, just there are loops now, and it is more likely that the thing being worked on is finished while still in the loop, which is the reason why you crashed, which shouldn't happen but for bad code.

To be sure though, you are drawing this?

                   1st                    nth
-(100g/s)->+[Thermo-regulator]+...+[Thermo-regulator]+->
           |<-----(50g/s)-----|                      |
           |                    .                    |
           |      		.                    |
	   |		        .                    |
           |<----------------(10g/s)-----------------|

or

-(100g/s)->+[Thermo-regulator]+->
           |<-----(100g/s)-----|

Neither works due to pipe logic problems while the second one indeed does have equal input and output rate.

Quote

You're also don't seem to get that in order to feed 16 inline thermoregulator with a constant maximum flow you need about 50 air pumps. And the system can supply oxygen for up to 100 dupes. It will also use about 10KWatt constantly which is basically about 40 dupes doing nothing but run on hamster wheels all day long when you count in sleep, and dinner and bathroom breaks. What I'm basically saying is that you run a feedback loop system so you have an oxygen supply that actually matches the number of dupes you have, and have consistent lower running power draw.

No, I have 16 inline thermo-regulators built in my base and it only takes 2 pumps.  You do not need more than that ever since thermo-regulator can only process ~1000g/s while the pump will at most pull in ~500g/s.  Really, the pump average out around 350-400 g/s, so you would need 3.  But the pumps aren't the main power consumer.  The 16 inline thermo-regulators eat 3840 W or 2304 KJ per day while providing at most 1000g/s of oxygen due to their work rate.  Hamster wheels will need 9.6 dupes or 10 dupes with some breaks, which consumes 1000g/s, plus or minus since they sometimes consume more or less during the day; definitely more during the night.  They will definitely be 100% stressed.  If you have diver's lungs for all, then you can at most support 13.  In this case, you will net energy and oxygen, which you will need for the pumps.  The total energy needed is 11.4 hamster wheels, so 13 diver's lungs is really really important.

So, I do not see where you get your numbers from.

Pipes do not work using the pressure principle, which would be very calculation intensive for the game.  Pipes are like roads and the contents like cars.  The closer a branch point is to a source, the more likely the cars will go down that branch point.  Thus all of these feedback loops will do nothing as cars do not spread out on the roads.  At most two feedback loops are useful in the current game build as the content rarely visits the third branch.  One feedback loop provides at most 2x the input rate.  Adding a second will bring that to at most  2.5x the input rate.  This drops badly with a third.

If the patch changes pipe logic, maybe feedback loops will be more useful.

 

[Saturnus]

@Fatmice I'm sorry if I couldn't satisfactorily explain to you what a single feedback loop does and how you can effectively just put them in series to get the result I describe. I'm not a teacher, and English is my 3rd language so I chalk it down to my inability to describe the function in a way that you can understand. I'm only a MSc in Electrical Engineering, so I use these types of feedback loops practically every day, so the function is so natural to me that I can't even begin to imagine not grasping the principle. Again, sorry. Maybe someone else has a better way of describing it?

[Mijae]

Okay, not that I didn't believe you but I did the calculations myself to understand how the feedback loop would effect efficiency and I think I understand it now.  I initially had an error, but I agree each additional loop should double the efficiency.  So 3 would reduce temp by 112C and 4 by 224.  This is likely good enough on its own, but might need to tweak the starting temp.  To lessen the confusion, you could just put a 5th single regulator in series on its own to simulate 17 (instead of using the 10g valve to get a non-integer multiple) or loop it once for 18.  5 is the most you should need though. @FatmiceNote the difference in terminology is that the N loops are not running in series, but recursively.  This is how it gets to 16x.

Of course, this assumes temperatures and volumes should mix in the pipes as expected.

It's interesting that it should take very few loops for the efficiency to stabilize also.  However, you will need to account for that time before it does stabilize and what to do with the very cool but not liquid dirty O2 that comes out.  Of course, this might be something to deal with regardless due to input temperature fluctuations.

What's more, I think you could could get quadruple efficiency if you split the output and feed it back into itself also (so 3/4 volume instead of half).  Sending the feedback in with 3 splits (so 7/8 packets) should increase efficiency by 8x.  You could theoretically get 16x with two of these in series.  It does take a little longer to stabilize though.  Note that the pressure would also gradually increase internally, stabilizing at 8x the input inside the system.  That means 100g in still means 100g out, but the regulators are processing in 800g chunks.

Unfortunately, there are limits to the volumes that pumps and regulators support.  Because of this I think the single split for 2x regulator efficiency you proposed is most volume/power efficient using one pump without a valve.  5 regulators plus a pump is 1440W, which can be generated by ~4 dupes.  So, you just need the pump to average enough to keep those dupes alive (easily for diver's lung dupes).

 

[Mijae]

I've been trying out some configuration with limited success.  I think if volume were to actually split at junction points this could be very viable.  However, since it redirects every other packet down splits the system ends up being unstable/unpredictable.  I'm going to start a separate thread with some of my findings.

 

 

[MoMuffins]

Actually, the feedback loop would work if we have a gas filter that is temperature sensitive.  It would work very easily.