Overhauled Intake Air, solving airhogging and asymmetric flamout with one stone

[Kasuha]

I think it falls down to how Unity interpolates these settings (AFAIK it's Unity functionality) and I'm quite sure it does not use very high order polynomials. I did not find any info regarding exactly this, but this might be a start:

http://docs.unity3d.com/Manual/EditingCurves.html

I have also tested behavior of the curve behind the end of the graph and the turbojet softly zeroes on zero thrust as you pass 2400 m/s surface velocity so I don't see any bugs in there.

[Clockwork_werewolf]

I suppose it might look a bit weird doing a "curve" for only two points. So you add that S curve.

On the other hand if you don't end with 0 what does it do? It obviously doesn't look at previous points so...? Flat line? Assume 0 at 10,000?

TO THE MODDING MACHINE!

[Clockwork_werewolf]

I have also tested behavior of the curve behind the end of the graph and the turbojet softly zeroes on zero thrust as you pass 2400 m/s surface velocity so I don't see any bugs in there.

I think it just wasn't behaving the way I was expecting it to, by having many curves not fitting a curve to the points.

[Kasuha]

Here's what I found:

http://answers.unity3d.com/questions/464782/t-is-the-math-behind-animationcurveevaluate.html

Short answer: it's bezier curves.

[Red Iron Crown]

Interesting links, Kasuha. Looks like they're using cubic Bezier curves for the S-profile, and other curves could be made available to smooth curves with more points.

I'm not sure if you logged any data points during your testing, did you happen to notice a slower rate of thrust change as you approached the 2000m/s mark, similar to the basic jet approaching 850m/s?

[Clockwork_werewolf]

Not that surprisingly it flat lines after the last point (what ever that is).

I made a turbo jet with just two points

key = 0 0.5 0 0

key = 1000 1 0 0

It kicked with about 4 G's as I left the atmosphere

[arkie87]

Airhogging is exploiting the current intakeAir / jet-engines-that-aren't-really-jets model to produce thrust at altitude that is far beyond what should be possible there. A jet engine can only produce the same thrust at sea level and at an altitude where the pressure is nearly 1/10 that of sea level if, and only if one of these two conditions is met:

So i could be wrong, but to me, what OP meant by "airhogging" is directly related to asymmetric flameout i.e. one engine receives more air than the other (even if both engines and intakes were placed with symmetry), and causes the other to not receive enough, resulting in asymmetric flameout.

[Capt Snuggler]

So i could be wrong, but to me, what OP meant by "airhogging" is directly related to asymmetric flameout i.e. one engine receives more air than the other (even if both engines and intakes were placed with symmetry), and causes the other to not receive enough, resulting in asymmetric flameout.

I think the OP is raising the two separate but related issues of airhogging and asymmetric flame out.

personally I don't think asymmetric flame out is really something that needs to be "fixed" but instead the player needs to be given better tools to manage it.

To me, air hogging and intake spam are the same thing. I do not believe allowing the player to stick 10 - 100 air intakes per engine, all over the craft, is the intended outcome. its a side effect of a hastily assembled, place holder part set introduced earlier in development. the implications of which were not realized at the time, or were realized but were considered low priority.

each engine should have its own set of intakes. giving some parts a "air passable" tag and introducing a better user interface for setting up airflow (resource flow) through your craft would go a long way to correcting the broken intake air system.

[Wanderfound]

I think the OP is raising the two separate but related issues of airhogging and asymmetric flame out.

personally I don't think asymmetric flame out is really something that needs to be "fixed" but instead the player needs to be given better tools to manage it.

To me, air hogging and intake spam are the same thing. I do not believe allowing the player to stick 10 - 100 air intakes per engine, all over the craft, is the intended outcome. its a side effect of a hastily assembled, place holder part set introduced earlier in development. the implications of which were not realized at the time, or were realized but were considered low priority.

each engine should have its own set of intakes. giving some parts a "air passable" tag and introducing a better user interface for setting up airflow (resource flow) through your craft would go a long way to correcting the broken intake air system.

Yup.

Airhogging (in the "intake spam" sense) is, like Kraken drives, infinite thrust RCS, etc etc, a game exploit that, while occasionally amusing, is probably not something that should exist in the completed game.

(wow, that's a lot of commas for one sentence. You'd never believe that I used to be a professional writer...)

I do like the air-passable part solution, although there are several ways a decent result could be achieved. Basically, it needs to get to the point where the game mechanics don't incentivise builds that would make an aeronautical engineer fall over laughing.

There's nothing wrong with keeping the possibility of ridiculous contraptions, but they shouldn't outperform things that are constructed sensibly. The current stock-aero approach to intakes shatters immersion and willing suspension of disbelief.

As for asymmetric flameouts, they're very easy to manage already; I don't think they need to be any easier. Asymmetric flameout is an issue that real-world pilots and engineers have to consider. Having it in the game is a feature, not a bug. But, again, it should be a bit more sensibly done. Having to fiddle around with order of part placement isn't good, and always knowing which engine is going to be the one to flame out first isn't great either. And there probably should be some sort of "auto shutdown on flameout" part for those who feel the need.

[Kasuha]

I do not believe allowing the player to stick 10 - 100 air intakes per engine, all over the craft, is the intended outcome.

The game is in general designed to allow you to build your craft any way you please. Want 20 command pods in a ship? Or 20 fuel tanks? 20 engines? Or 20 intakes? You're free to do that. That's what is fun on KSP that the game lets you do all kinds of stupid things with its parts.

I don't understand the fuss around "airhogging". You can build a functional SSTO with one intake per engine. If you use more, the more you use the more comfortable it gets - but above three or four the gain is starting to vanish, you don't really get much more from your plane if you have 10 intakes per engine than if you have three. Honestly, I don't see the problem.

As far as I understand OP's post, I think his definition of airhogging is storing intake air in closed intakes. That IMO counts as exploit. But it's not like it gives you a whole lot of power.

[Jouni]

I don't understand the fuss around "airhogging". You can build a functional SSTO with one intake per engine. If you use more, the more you use the more comfortable it gets - but above three or four the gain is starting to vanish, you don't really get much more from your plane if you have 10 intakes per engine than if you have three. Honestly, I don't see the problem.

The problem is that we don't have jet engines in the game. We only have some magic technology from Star Wars that masquerades as jet engines, but works in a completely different way. Nothing we build with them feels even remotely like something powered by jet engines.

Even though KSP is mostly a game about rockets, jet engines would be a good addition to it. Especially now, as the next update is going to include many new spaceplane parts.

[Clockwork_werewolf]

I'm just going to leave Scott Manley here

.

I stick to my earlier comment of x8 intake and 50x mass for intakes, though I will add that turbo jets should have their thrust curve greatly attacked at high ends. I found an experimental Turbo jet that worked at Mach 3.2 http://en.wikipedia.org/wiki/Airbreathing_jet_engine P&W J58 Mach 3+ afterburning turbojet. Mach 3.2 = 1,088.928 m / s. While this was its cruising speed the idea that our turbo jets work at 2,399 m/s (mach 7) is just ridiculous. Instead of

key = 0 0.5 0 0

key = 1000 1 0 0

key = 2000 0.5 0 0

key = 2400 0 0 0

I think the turbojet should be more

key = 0 0.5 0 0

key = 1000 1 0 0

key = 1500 0.5 0 0

key = 1600 0 0 0

and even this way outside of the real world even in experimental designs.

[Kasuha]

Okay, you don't like how jet engines work in KSP, I get that. But that's a completely different topic. And you still did not explain to me why airhogging as intake spam is bad, or even when it is still fine and when it is already airhogging.

You won't get your plane much above 2200 m/s ground speed regardless how many intakes you use and it's almost irrelevant whether you can do it at 40 km with three intakes or at 45 km with twelve, you're efficiently in orbit in both cases and your total orbital energy is not vastly different.

Actually, intake air distribution among engines was improved by devs in a recent release (was it 0.22?) so now you can do with less intakes what needed much more intakes before. With that change, airhogging is a non-problem to me. You don't need to spam intakes. You don't get any substantial reward if you do.

[Wanderfound]

I don't want to pick a fight with you, Kasuha, however: see my post above.

Yes, KSP should allow you to stick 20 command pods on a ship if you want to. But that ship should be substantially outperformed by an equivalent build that ditches the 19 redundant pods.

At present, in stock aero, the 20 intake-per-engine plane is better than the 1 intake-per-engine plane. The game actively encourages the construction of silly unrealistic things and actively discourages the construction of sensible things. This is backwards to me. Silly is fun, silly is fine, but sensible should have value as well.

I'm not the dictator of KSP, and you're entitled to your view just as much as I am. But you asked why we think intake spam is bad, and this is why.

[Jouni]

Okay, you don't like how jet engines work in KSP, I get that. But that's a completely different topic. And you still did not explain to me why airhogging as intake spam is bad, or even when it is still fine and when it is already airhogging.

Airhogging is a part of the same general problem.

With basically every other system, having too many of it makes the craft perform worse. There is always a trade-off. If you have too many engines, your delta-v suffers. If you have too much fuel, the rocket wastes more fuel and may not even lift at all. If you have too many wings, the plane is harder to fly. Not so with intakes: if your plane has too many intakes, it just flies better.

It's neither realistic nor fun. A better intake mechanism would have you make meaningful decisions. Use too many intakes, and the plane performs badly at low altitudes due to excessive drag. Use too few intakes, and the plane can't fly high enough to reach orbit efficiently. Make a compromise between the two extremes, and the plane performs adequately everywhere, but isn't very good anywhere.

[Kasuha]

Yes, KSP should allow you to stick 20 command pods on a ship if you want to. But that ship should be substantially outperformed by an equivalent build that ditches the 19 redundant pods.

Ship with 20 command pods can bring 20 copies of the same science measurement.

Ship with 20 engines has better TWR.

Ship with 20 reaction wheels has better torque.

Ship with 20 fuel tanks has greater dv.

Ship with 20 intakes can fly higher.

Each of them outperforms their equivalent with just one copy of the part.

At present, in stock aero, the 20 intake-per-engine plane is better than the 1 intake-per-engine plane.

Not substantially. That's my point.

Airhogging is a part of the same general problem.

Not really. There's plenty of spaceplane challenges with "no airhogging" in their rules, but none of them prevents usage of jet engines. You're just evading the question.

With basically every other system, having too many of it makes the craft perform worse. There is always a trade-off. If you have too many engines, your delta-v suffers. If you have too much fuel, the rocket wastes more fuel and may not even lift at all. If you have too many wings, the plane is harder to fly. Not so with intakes: if your plane has too many intakes, it just flies better.

It still has more mass and drag and most importantly higher part count.

It's neither realistic nor fun.

Realistic, it's not. But I disagree with not fun. I have plenty of fun flying planes with current jets.

[Clockwork_werewolf]

Ship with 20 command pods can bring 20 copies of the same science measurement. But weights at least 16 tons more.

Ship with 20 engines has better TWR. But has so little Delta-V that it has to drop the stage early

Ship with 20 reaction wheels has better torque. But weights at least 2 tons more and uses 20 times the power requiring more batteries or power generation.

Ship with 20 fuel tanks has greater dv. But has lower TWR to the point it becomes inefficient or can't lift off.

Ship with 20 intakes can fly higher. but weights 0.2 tons... wait that's nothing.

"At present, in stock aero, the 20 intake-per-engine plane is better than the 1 intake-per-engine plane."

Not substantially. That's my point.

That's why I posted the Scott Manley video. Although 20 is overkill, 12 intakes is still excessive but lets you get an AP out of the atmosphere (not PE) with about 50m/s Delta-V to add in space to get orbit.

Not really. There's plenty of spaceplane challenges with "no airhogging" in their rules, but none of them prevents usage of jet engines. You're just evading the question.

By not allowing air hogging they mostly make using jet engines at high altitudes (and therefore speeds) impossible. People could be using shutdown of engines to allow the jet to work at high speeds but this is a compromise and maybe they should ban jet engines .

It still has more mass and drag and most importantly higher part count.

Its extra mass is tiny, if you use 12 (probably optimal) and 2 engines then you add 0.24 mass. The drag of intakes maxes at 2 instead of the standard 0.2 but this only means they have the drag of a 0.1 mass part. 1.2 DragMass is about the same as the engine but if this lets you put 4 tons into orbit then try and get that into orbit with just 1.2 mass extra fuel. Also when in space you still only have 0.12 of extra mass to tow around.

High part counts should not be a test of your design, that's a test of your computer and means laptop players are less efficient that MEGAMACHINE players.

I have plenty of fun flying planes with current jets.

Yea i think we all do or we wouldn't care about this. The point is that we know we can chuck stuff into orbit for almost no cost if we airhog but it looks and feels totally wrong with all the intakes.

I will say that if jet engines thrust curve was changed then they would need a reduction in cost to make them worthwhile again but we now have more factors (like cost) to get things balanced.

[Capt Snuggler]

Ship with 20 command pods can bring 20 copies of the same science measurement.

Ship with 20 engines has better TWR.

Ship with 20 reaction wheels has better torque.

Ship with 20 fuel tanks has greater dv.

Ship with 20 intakes can fly higher.

Each of them outperforms their equivalent with just one copy of the part.

yes but all have a significant penalty.

20 com pods = I have never tried but I would think you would get diminishing value of duplicated science measurements. science storage is another example of an system that will be revised.

20 engines = too many and you waste fuel.

20 reaction wheels = massive power consumption and too many - you don't fly.

20 fuel tanks = too many and you don't fly.

20 intakes = no significant penalty

they should be renamed oxygen generators.

Not really. There's plenty of spaceplane challenges with "no airhogging" in their rules, but none of them prevents usage of jet engines. You're just evading the question.

intake spam is bad because it makes jet engines stupidly overpowered and makes every other engine redundant when it comes to launch stages. I do not believe this was the desired outcome.

Also I don't think he is evading the question, the question is just quite hard to follow.

It still has more mass and drag and most importantly higher part count.

The mass and drag increase is not significant enough nor does this adequately represent the design issues and considerations involved with using jet propulsion, or any other air breathing engine for that matter, for getting in to orbit.

Also the higher part count is not relevant because the effective penalty of this is lag and dependent on your CPU GHz, instead of an in-game balance mechanic.

[Majorjim!]

Hi all,

Just my two cents. Using intake spam is kinda cheaty and the game shouldn't allow it. Surely code could be written into the game to limit the number of usable

intakes per engine. Say capped at two max per engine. Then any added after that number wouldn't give you anymore oxygen and would just add weight.

[Jouni]

Realistic, it's not. But I disagree with not fun. I have plenty of fun flying planes with current jets.

Flying spaceplanes can be fun, and designing them can also be fun, but there is nothing fun in adding the intakes. It's just a routine task of adding more and more intakes, until you get bored, the plane starts looking too ugly or too unreasonable, or your computer can't handle the part count anymore.

An easy way to make intakes more interesting would be to make their mass nontrivial. With 0.5 tonnes per intake, you would have to consider seriously, how many intakes the plane does really need. As an added bonus, intake placement would also be a serious challenge with stock aerodynamics. (Or, more likely, the intakes would need a smaller drag coefficient.)

[Kasuha]

My experiment regarding "intake spam".

Here is a "plane" with single turbojet and single RAM intake.

Up to 30 km, its speed and performance was held back by drag

Between 30 and 33 km, it was limited by speed curve

At 33 km it still runs on full throttle it is still producing 18.7 kN of thrust

Above 33 km, intake air limits kicked in (so I had to reduce throttle) together with speed curve

At 54 km on reduced throttle it is still producing 0.1 kN of thrust

It did not definitely flameout below 56 km altitude

When it left atmosphere, it was on orbital trajectory with apoapsis at 150 km and periapsis at 30 km

All in all, it flies at least as well as Scott Manley's "airhogging" design. With single intake. Of course, if you add more intakes today, you can fly on full throttle a bit higher. But that's it. Air just goes away exponentially and more intakes just give you a few meters of altitude at which you still can fly at full throttle.

Adding more intakes is just making things slightly more convenient. You just can't get significantly more from it.

Javascript is disabled. View full album

And just to convince you it works even if it is complete plane (picture taken before return to atmosphere):

Edited September 8, 2014 by Kasuha

[andrew123]

It's pretty easy to go hypersonic with one intake.

Javascript is disabled. View full album

Javascript is disabled. View full album

[Jouni]

All in all, it flies at least as well as Scott Manley's "airhogging" design. With single intake. Of course, if you add more intakes today, you can fly on full throttle a bit higher. But that's it. Air just goes away exponentially and more intakes just give you a few meters of altitude at which you still can fly at full throttle.

Now you're arguing that because jet engines are already broken in every conceivable way, it doesn't matter that the intakes are also broken. Your example demonstrates the well-known fact that engine spam makes things as ridiculous as intake spam. With extremely high TWR, even one intake per engine is enough to launch the ship to a stable orbit. Because drag is proportional to mass in stock aerodynamics, doubling the TWR is essentially the same as doubling the number of intakes per engine.

[Kasuha]

Now you're arguing that because jet engines are already broken in every conceivable way,

You just keep changing the topic. I assume it is because you're afraid to admit that I'm right.

I have already proposed a common point we can agree on: turbojets are unrealistic and powerful. I have no problem agreeing on that. Does that make you happier? But this is a game and turbojets work as designed. Turbojets are built to allow players reach space. Can you agree on that? We got over that topic several times already. I am not discussing them.

My argument is that intake spam is not an exploit since you don't get any substantial reward for it.

Edited September 9, 2014 by Kasuha

[Jouni]

My argument is that intake spam is not an exploit since you don't get any substantial reward for it.

You get substantial rewards for it. Even your example vehicle with just one ram intake exhibited serious intake spamming (engine spamming is just another name for it).

Because of how stock aerodynamics work, intakes/engine is a bad measure of intake spamming. The total amount of air the intakes provide defines an upper bound for the total thrust of all jet engines. It doesn't matter whether this thrust comes from one engine or from multiple engines. If that peak thrust is higher than drag, the plane continues to climb and accelerate at (near-)orbital velocities. Because drag is proportional to vehicle mass, intake spamming should be defined based on intakes/mass.

By my definition, the normal amount is around 0.1 ram intakes/tonne, and anything significantly above that gives you substantial rewards. If you have x times more intakes per unit of mass, your maximum cruising altitude is ln x * 5 km higher. With x = 3 (roughly 3-4 intakes/engine in typical designs), the difference is significant, as it raises the cruising altitude well past the point where terminal velocity exceeds orbital velocity. Below that point, you need a lot of thrust to keep flying at near-orbital velocities. Above that, the amount of thrust you need to sustain your velocity rapidly decreases.

Increasing x above 3 is also beneficial, as it makes flying easier. If you were flying at 1/3 throttle with x = 3 before switching to rockets, you can just keep flying at full throttle with x = 9.

Edited September 9, 2014 by Jouni