How should rockets flex?

[Kerbart]

23 minutes ago, darthgently said:

Does the physics model take triangulation of manually placed struts into account?  I do it as it makes sense IRL 

As far as I can tell, no. I tend to do it too, because it looks good, and I feel that "making things look like I think they would in real life" is a sound design principle even when you know it doesn't matter. Based on experience you can consider a strut more like a very stiff beam that handles pressure as well as stress and they just tend to keep points on two parts in the same position relative to each other.

But it does work the other way around as well. Surface mounting tanks to other tanks? (To get like a Sat 5 layout with the engines "sticking out")? I usually place a pair of struts at the "engine mount" tanks because it makes the connection that much stiffer, even if it wouldn't be that effective in real life. The surface mount is really just a point-to-point contact, and adding the strut gives you that triangular relation for stiffness.

Consider the Kickback SRB. I usually put the radial decoupler somewhere in the middle, have two struts (one on each side) at the top and and two struts at the bottom. Assuming you have two Jumbo 64 tanks, attaching those SRB's that way will result in a very stiff connection in between (the Kickback is a single part with no internal flex). If it connects at the top with a second stage... well, now you're immobilizing three joints (tank - coupler - engine (plate) - tank) or even four if there's a fairing involved. Of course you lose the benefit once the SRB's are jettisoned but by then you should be beyond Q_max.

[RocketRockington]

I really wish KSP2 had tried to do something else with the part physics.   It's not just that noodle rocket suck.  Though they do and even with PhysX there were other options - for instance having all parts rigidly attached and computing stress tensors for breakage bespoke.

But it's also the PhysX is too much of a general purpose physics engine oriented toward a wide variety of standard game uses, rather than one what's specialized for what KSP really needs.

KSP needs a physics system that is optimized for mechanical linkages.  And one that tries to be more accurate/realistic about how forces worked rather than just the verisimilitude that PhysX offers.  Conservation of energy/momentum is just not a thing PhysX really cares about, other than making it look right, for instance.

KSP2 could have done this in a couple of ways.  Unity allows you to change out the physics engine, though it is a bit more work to swap it out, a game that is so physics dependent should have at least investigated doing this.

They could have used a more open-source engine that allowed them to tinker under the hood.  Bullet  for example. Stormworks: Build and rescue uses Bullet and has thousand-part ships that are rigidly bound together, and physically simulated, very similar to what you'd what in Kerbal

They could also have licenses an engine that is more optimized for mechanical simulation, like MuJoCo or BeamNG

Unfortunately, I don't think this is a decision that will ever make the dollars and cents logic at this point for them to retract and redo - I feel bad that when their version of breaking ground robotics comes about, it's going to be just as floppy as KSP1's - maybe more so since the KSP2 team seems to currently want the floppiness.

[RCgothic]

I still don't understand why "don't build a noodle" is something you have to teach.

There's no good reason for long rockets not to be a valid design.

You shouldn't have to strut side boosters up like a Christmas tree just so the wibbling thrust vectors don't cause the rocket to flip despite appropriate CoM/CoA positions.

These are all artificial workarounds that fail to address the root flaw: unstable rockets are frustrating, creatively limiting, and unfun.

Edited March 13, 2023 by RCgothic

[dvrabel]

2 hours ago, RCgothic said:

I still don't understand why "don't build a noodle" is something you have to teach.

There's no good reason for long rockets not to be a valid design.

Bending stresses are proportional to the length / width ratio. In real life, long, thin rockets would need to be stiffer, and thus (much) heavier.

3 hours ago, RocketRockington said:

They could have used a more open-source engine that allowed them to tinker under the hood.  Bullet  for example. Stormworks: Build and rescue uses Bullet and has thousand-part ships that are rigidly bound together, and physically simulated, very similar to what you'd what in Kerbal

They could also have licenses an engine that is more optimized for mechanical simulation, like MuJoCo or BeamNG

Unity's physics engine, Bullet, and MuJoCo all seem to model rigid body physics in the same way, as far as I can tell from their documentation. Any difference will be related to how the joint constraints and other properties are configured.

[cocoscacao]

I was hoping for something like raw spaghetti as the bare acceptable minimum, but we're stuck with al-dente for now.

[mcwaffles2003]

15 hours ago, Kerbart said:

The problem with realistic flex is that it's so little that you don't see it. Metal is elastic; everything flexes under stress. But the displacement it causes is so little that we rarely see it in daily life. Apply that to rockets and they would just break. Visible flex is a bit like an early warning system that you're asking too much of the construction. The problem with KSP is that parts don't flex, only joints. Not only do joints have to be overly flexible to compensate for the lack of flexing in the parts themselves, we also have a lot of them. A real-life booster is a single mechanical entity and not cobbled together from two or three tanks.

One of the problems is that most people are, truth be told, lousy mechanical engineers. Everybody has their mouth full about the need for more realism with n-body physics, Lagrange points and decaying orbits, when when it comes to unstable contraptions the response is not "let's build something better" but rather "We NeEd AuToStRuTs."

I've never used autostruts. I use regular struts where it makes sense (Medium - Extra Snall - Medium? You bet I connect the medium parts with struts). I don't launch a 5m payload on top of a tall 2.5m rocket. My rockets don't noodle.

I'm not saying the current setup is satisfying—far from that. Yes, there is too much noodling but the correct solution is make the rocket break apart, not "make the joints stiffer" or "we need autostruts." It means that you can't just build anything you want and expect it to make it to orbit. That's not always easy. But that's why it's rocket science.

I feel like instead of flexing so much they should just highlight the joint yellow/red. Huge ask but having 1 trailing 5 s video you can review like a kill cam in COD would be cool too...

[Fullmetal Analyst]

they should flex like starship

[Superfluous J]

I've not played KSP2 so I must keep my opinions based on how I think it should work, not how it works.

That said, I think rockets should bend/flex to reflect how well (or poorly) they are built.

Which is how they bend/flex in KSP1.

[TheOtherDave]

On 3/13/2023 at 10:22 AM, dvrabel said:

Visibly flexing rockets gives a good visual indicator that the rocket is over-stressed and may be about to fail. Without some sort of indication like this, rockets that spontaneously break apart with no warning will be very frustrating , particularly for less experienced players.

Visibly flexing rockets make that problem worse since they’re so hard to control.

[EvelynThe Dragon]

Patch 1 put everything back to stupid noodles. PLEASE FIX THIS. It is impossible to build anything that works like it should.

[ArmchairGravy]

I'm ambivalent about this. On one hand, I shouldn't be able to get my 2-deep 3-way onion booster to space without struts as the game currently stands. On the other hand, struts add to part count, and I saved 12 parts by not having to strut. The gripping hand is performance impact in a game with performance issues.

[InterstellarDrifter]

On 3/13/2023 at 9:32 AM, Kerbart said:

The problem with realistic flex is that it's so little that you don't see it. Metal is elastic; everything flexes under stress. But the displacement it causes is so little that we rarely see it in daily life. Apply that to rockets and they would just break. Visible flex is a bit like an early warning system that you're asking too much of the construction. The problem with KSP is that parts don't flex, only joints. Not only do joints have to be overly flexible to compensate for the lack of flexing in the parts themselves, we also have a lot of them. A real-life booster is a single mechanical entity and not cobbled together from two or three tanks.

One of the problems is that most people are, truth be told, lousy mechanical engineers. Everybody has their mouth full about the need for more realism with n-body physics, Lagrange points and decaying orbits, when when it comes to unstable contraptions the response is not "let's build something better" but rather "We NeEd AuToStRuTs."

I've never used autostruts. I use regular struts where it makes sense (Medium - Extra Snall - Medium? You bet I connect the medium parts with struts). I don't launch a 5m payload on top of a tall 2.5m rocket. My rockets don't noodle.

I'm not saying the current setup is satisfying—far from that. Yes, there is too much noodling but the correct solution is make the rocket break apart, not "make the joints stiffer" or "we need autostruts." It means that you can't just build anything you want and expect it to make it to orbit. That's not always easy. But that's why it's rocket science.

THIS^^^

IMO this is where KSP2 should land. Soooo much of the "noodliness" of rockets is boiled down to a poorly designed rocket(unreal fineness, too much mass per diameter, unstable aerodynamics, poor choice of connections, etc.) and/or a suboptimal ascent. 

Does the current model work/feel right? I would answer no. But the right answer probably lies somewhere between better player designed rockets and a better physical simulation of mechanical properties of metal cylinders in atmospheric flight, not "auto-strut" or perfect rigidity.  The poll shows an obvious desire by the majority of the KSP community for more SIM vs. ARCADE style play. 

Also, I feel like this balance is difficult to strike. I think most people like playing KSP because we feel like we're learning, understanding, and accomplishing vs. just interacting/playing with a game. However, you want it approachable and fun for a younger crowd with less understanding a vast and complex field of physics. But, I gather that those drawn to trying KSP are typical of being ok with hard, challenging, unknown, different, new, etc.. in a game. I feel the nature/core purpose of KSP is to have fun while accomplishing and learning from experimentation and our mistakes.  So, you have to be able to make the mistakes in the first place to learn from them, so, ideally, I would suppose the balance should be found closer to simulation(how stuff works IRL, i.e. rockets going the speed of sound, break up, when flipping out in high atmospheric pressures). 

[dr.phees]

Autostrut was introduced to work around the limitations of Unity's physics model. This should have been solved before building KSP 2. I was really expecting KSP 2 to mainly deliver a solid physics platform where not every single part had to be defined separately, but using a physics framework which would allow much better parts development, especially would support procedural parts. I am really disappointed that we have the same parts and are again in early EA limited again by Unity's - already known - unfitting physics systems.

Edited March 18, 2023 by dr.phees

[mrjsonkerbal]

Hey, youtuber Matt Lowne just posted this, and it speaks to exactly my problem with the game at the moment, both the game-ruining issue that wobbly rockets represent, and also his perspective on Kerbals.

 

[RocketRockington]

3 hours ago, mrjsonkerbal said:

Hey, youtuber Matt Lowne just posted this, and it speaks to exactly my problem with the game at the moment, both the game-ruining issue that wobbly rockets represent, and also his perspective on Kerbals.

 

Hopefully when other features of the game are released that needed design work vs just being copies of KSP1 or a couple of high-profile mods like waterful and textures unlimited, they won't be as out of touch with what fans would have wanted as the UI and the wobbly rocket decisions.  Cross your fingers for science mode.

[PDCWolf]

Imma just quote myself to not type again:

I'll disagree on there being any case for wobble: there isn't any single viable case for wobble to apply, as what's a "dumb design" for a rocket, could be a proper design for a space station, or some other vehicle in whatever common or strange circumstances. Further on, we are on a tighter part budget than on KSP1, thus having to waste parts on bandaids like struts (which have always ended up adding more problems than solving) is an unjustified punishment to creativity for no reason other than the game's shortcomings. Lastly, wobble is not realistic nor analogous of any structural phenomena that goes on in real life rockets, as those are made of individual, specially designed parts, created to be integrated in a single order, and work with specifically with each other, whilst the player is forced to assemble their rockets off generalist parts, thus punishing them for a structural soundness the game doesn't simulate, nor can the player imitate.

I've also been agreeing for almost a decade with what Bac9 wrote about the Kerbals and Matt quotes. "Kerbal" as a synonymous for bad, disastrous, or janky is the worst thing that could've ever happened to this game.

[Alexoff]

Wobbling rockets are very easy to fix, just increase the parameter JOINT_RIGIDITY in the file PhysicsSettings.json by a 1000. Perhaps someday the developers will make such a simple hotfix. This will definitely make the fans happy.

[Spicat]

2 minutes ago, Alexoff said:

Wobbling rockets are very easy to fix, just increase the parameter JOINT_RIGIDITY in the file PhysicsSettings.json by a 1000. Perhaps someday the developers will make such a simple hotfix. This will definitely make the fans happy.

This doesn't fix wobbly rocket at all, just make it less bad in some situation. I'm glad they are not going with that "fix".

[Alexoff]

30 minutes ago, Spicat said:

This doesn't fix wobbly rocket at all, just make it less bad in some situation. I'm glad they are not going with that "fix".

I think doing something less bad is better than doing nothing.

[Lisias]

2 hours ago, Alexoff said:

Wobbling rockets are very easy to fix, just increase the parameter JOINT_RIGIDITY in the file PhysicsSettings.json by a 1000.

The problem we are facing here is that "wobbliness" happens when we abuse the part's strength. I don't mind replacing wobbliness with something else, as long the limits and constrains are there, forcing me to rethink the designs to cope with them.

But what would be a good substitute for it?

 

2 hours ago, PDCWolf said:

Lastly, wobble is not realistic nor analogous of any structural phenomena that goes on in real life rockets, as those are made of individual, specially designed parts, created to be integrated in a single order, and work with specifically with each other, whilst the player is forced to assemble their rockets off generalist parts, thus punishing them for a structural soundness the game doesn't simulate, nor can the player imitate.

But that's the point. There's no wobble in real life, but we don't build rockets and places in real life as we do in KSP neither.

The problem remains: how to simulate and punish bad, structurally unsound designs?

 

2 hours ago, PDCWolf said:

I've also been agreeing for almost a decade with what Bac9 wrote about the Kerbals and Matt quotes. "Kerbal" as a synonymous for bad, disastrous, or janky is the worst thing that could've ever happened to this game.

Kerbal is meant to mean whatever the player wants them to mean.

Players liking janky designs will see Kerbals as janky. Players liking "real life" designs will see Kerbals are realistic. Players liking alien style designs will see them as a highly advanced civilisation. And so go on.

Most people like to have a good laugh when playing, and so a janky Kerbal tech is usually the way most people go. There's a reason the lowest tech batteries resemble D Batteries - because it's fun this way.

Edited July 1, 2023 by Lisias
kinda of tyops.

[PDCWolf]

Just now, Lisias said:

But that's the point. There's no wobble in real life, but we don't build rockets and places in real life as we do in KSP neither.

The problem remains: how to simulate and punish bad, structurally unsound designs?

We don't have to punish "bad, structurally unsound designs" because those are impossible to define under the current building system, as a lot of it is left to imagination (using fuel tanks as modules of a space station, for example). There's no way a wobble mechanic fits any and all uses a part can have. Further on, do we really have to punish a problem the same software is creating? Again, we're on a limited part budget, and we're punishing players for not wasting that part budget on struts. Makes no sense.

1 minute ago, Lisias said:

Kerbal is meant to mean whatever the player wants them to mean.

Players liking janky designs will see Kerbals as janky. Players liking "real life" designs will see Kerbals are realistic. Players liking alien style designs will see them as a highly advanced civilisation. And so go on.

Most people like to have a good laugh when playing, and so a janky Kerbal tech is usually the way most people go. There's a reason the lowest tech batteries resemble AAA Batteries - because it's fun this way.

No.  As Bac9, and now Matt Lowne have exposed, Kerbals, through their buildings and parts, exhibit technology that requires much more intellect than humans. Yes, their battery tech is not much more advanced than ours, but their engines are decades ahead of what we have, for example. They might like to have fun, and they definitely don't have the regards we have for safety and established procedures, but they do not build jank, and they're not eager for death, that's just headcannons and the world and its components don't reflect it.

[Alexoff]

12 minutes ago, Lisias said:

The problem we are facing here is that "wobbliness" happens when we abuse the part's strength.

Let it be better to have some kind of game convention that the details in the game are made as if from cast iron, than we will see iron cylinders over and over again that behave like a bunch of sausages.

15 minutes ago, Lisias said:

Most people like to have a good laugh when playing, and so a janky Kerbal tech is usually the way most people go.

For some reason, it's very hard to find a video on YouTube where someone happily builds crazy crafts in KSP2 and laughs along with their subscribers. KSP1 was a semi-cartoon indie game where the hangar was insane place for drifting. KSP2 looks more realistic, and nothing happens in the hangar. Somewhere that unbridled fun disappeared ...

[Lisias]

7 hours ago, PDCWolf said:

We don't have to punish "bad, structurally unsound designs" because those are impossible to define under the current building system, as a lot of it is left to imagination (using fuel tanks as modules of a space station, for example).

Every single game I ever played has a set of rules and constrains, and you are punished somehow by not respecting them.

Granted, every single one of them so have settings or hacks so you can cheat your way on the game.

I don't have the slightest problem on having these rules and constraints being deactivated for people not willing to cope with them. I just don't want them permanently removed from the game.

Doing what I intend by compromising with such rules and constraints are part of the fun. As a matter of fact, my only complain about the autostruts (other than a bug on it since 1.2.2) is that it's dirty cheap. It should cost you both money and mass by using them, as it happens when you reinforce a structure (as using bigger spars on an airplane wings).

 

7 hours ago, PDCWolf said:

.No. 

Yes. My game, my rules. I don't try to tell you how you play your game,  I don't see why I should let you tell me how to play mine.

We will need to agree on disagree on this one. And let the users vote with their money.

Hint: check the majority of screenshorts and videos around. 

Edited July 2, 2023 by Lisias
Tyops gallore!!!

[PDCWolf]

4 minutes ago, Lisias said:

Every single game I ever played has a set of rules and constrains, and you are punished somehow by not respecting them.

Granted, every single one of them so have settings or hacks so you can cheat your way on the game.

I don't have the slighted problem on having these rules and constraints being deactivated for people not willing to cope with them. I just don't want them permanently removed from the game.

Doing what I intend by compromising with such rules and constraints are part of the fun. As a matter of fact, my only complain about the autostruts (other than a bug on it since 1.2.2) is that it's diety cheap. It should cost you both money and mass by using them, as it happens when you reinforce a structure (as using bigger spars on an airplane wings).

Well, once the constraints are clearly defined, we might have a discussion to partake in, but so far, there's no clear indicators of structural integrity, part strength (other than max collision speed), or any other form to tell the player how to "correctly" assemble things so they don't wobble.

6 minutes ago, Lisias said:

Yes. My game, my rules. I don't try to tell you how you play your game,  I don't see why I should let you tell me how to play mine.

We will need to agree on disagree on this one. And let the users vote with their money.

Hint: check the majority of screenshorts and videos around. 

You are invited and encourage to play however you want, however the worldbuilding and what's canonical in game sets a precedent. The fact you can drive on the sidewalk and run over people in GTA doesn't mean the game doesn't have a set of traffic rules imitating real life, just that you're free to disrespect them (and suffer the consequences).

[Lisias]

7 hours ago, Alexoff said:

Let it be better to have some kind of game convention that the details in the game are made as if from cast iron, than we will see iron cylinders over and over again that behave like a bunch of sausages.

I don't necessarily disagree with that. I want something, being the wobbliness better than nothing.  

 

7 hours ago, Alexoff said:

[…] KSP2 looks more realistic, and nothing happens in the hangar. Somewhere that unbridled fun disappeared ...

You see, I think this is the Elephant in the Room.

I had bough Juno Origins. Pretty nice game, I play it now and then, but… By some reason, I come back to KSP when I want to have some good hours of fun: doing something stupidity awesome, or most of that time just awesomely stupid - as long it's something that I could curse while laughing, or laugh while cursing.  This is the reason I spend so much time playing and modding this game.

I don't have absolutely any objections on going "serious" (it's the reason I have so many aviation mods on my rig, from ILS to VOR, not to mention GPWS as an example), but even when I'm playing "serious", there's a ludic background on the play (like, hey, let's distribute some diapers to the PAX instead of vomit bags - #JebediahFeelings).

I see these KSP2 videos and I like what I see as long it's someone else playing. But I'm not seeing me playing it for more than an hour. Hell, I have already Juno (with its astronauts) and Orbiter (not even that) for "serious" play and it's days since I fired up Juno, and months since I fired up Orbiter by the last time.

The Kerbal's "goofiness" is what makes the harsh path to the success palatable for me.

 

7 hours ago, PDCWolf said:

Well, once the constraints are clearly defined, we might have a discussion to partake in, but so far, there's no clear indicators of structural integrity, part strength (other than max collision speed), or any other form to tell the player how to "correctly" assemble things so they don't wobble.

Suggestions are welcome. How about something like this?

Not to mention aircrafts, how something like this? (the craft had a chute, no traumatising events on the video)

And how you suggest the game should support the user's diagnosing for such problems, once they happen?

I'm not being picky, I'm trying to open a brainstorming session.

 

7 hours ago, PDCWolf said:

You are invited and encourage to play however you want, however the worldbuilding and what's canonical in game sets a precedent. <…>

Yep, the Elephant in the Room I mentioned above.

Edited July 2, 2023 by Lisias
Moar tyops!