A solution for the wobbly rocket "problem"..

[dvrabel]

... and it's not autostrut, nor is it fixing some physics "bug".

Wobbly rockets are a deliberate game design decision that higlights a real engineering problem that real rocket engineers have to design for. Nothing in real life is perfectly rigid and engineers need to design rockets to handle flexing and bending during flight. KSP exaggerated the bending to make it an obvious problem to the player -- a wobbly rocket is much more obvious than an seemingly rigid rocket suddenly and unexpectedly cracking in half as internal stresses from bending moments become too great.

So, wobbly rockets need in-game, real world solutions. and not a magic autostruct button, so what are they?

1) Struts - Additional cross bracing to stiffen a structure as a real world solution to excessive flex.

2) Truss structures - You will have seen in the parts list a good number of new truss structures. Trusses are lighter and more rigid than tubes (fuel tanks) and should make excellent central "spines" for larger craft. However, trusses are terrible aerodynamically so we've swapped one engineering problem for a different one. How do we solve this one?

3) In orbit construction - Avoid the atmospheric induced forces and drag entirely! This is in the road map for interstellar craft but I think this would work equally well for large intra-kerbol craft.

Having said all that, I would also be in favour of a difficulty option that would make all joints perfectly rigid to support the play of those who don't want to tackle this particular problem.

[whatsEJstandfor]

It may be that real-world rockets aren't fully rigid, but you also don't see struts holding the boosters to the upper stage of SLS, or struts holding the Apollo CSM to the core stage of the Saturn V. Currently (and in KSP1 without struts or autostrut), joints between parts feel like they're made of thick rubber instead of welded steel. If wobbly rockets, at least with the properties that they have now, are legitimately a design choice, I just disagree with it. It's one of the charming things about KSP at first, but I think it's antithetical to the usual mid-game goal of building big-ass crafts

Edited February 25, 2023 by whatsEJstandfor

[J.Random]

39 minutes ago, dvrabel said:

2) Truss structures - You will have seen in the parts list a good number of new truss structures. Trusses are lighter and more rigid than tubes (fuel tanks) and should make excellent central "spines" for larger craft. However, trusses are terrible aerodynamically so we've swapped one engineering problem for a different one. How do we solve this one?

You don't seem to understand how Unity fizzicks works. Stack of lightweight parts connecting heavy structures (like interstellar ships from promotional materials, with engine/fuel tank block connected to long truss stack connected to hab/command assembly) will flex and wobble like there's no tomorrow, and SAS will help it to flex and wobble even more.

[dr.phees]

That was an issue with Unity physics in KSP 1 and is now again in KSP 2. Why? I thought it was obvious, that Unity's engine is a bad choice.

[dvrabel]

You will recall that the joint strength currently scales with part diameter.

Joints are modelled as damped springs and thus can be configured to be weak or strong as the game designer likes. Note also that the mass of the connected parts does not affect their strength (but mass would affect the natural frequency of any resulting oscillation).

I can see an argument for increasing the strength (or damping, or both) of the current joints, but I would not like to see them be trivially be made fully rigid as I find dealing with the wobble an interesting aspect of the game play. If you're not interested in this aspect, then sure go ahead and turn on that rigid joint difficultly option I proposed.

[obney kerman]

Idea: When you close a fairing by connecting it to another part, let that fairing act as a strut and hold the whole thing together, instead of having the top part of the craft wobble anyway and clip through the fairing! Something to make Saturn V replicas and interstage adapters a lot easier.

[Pyne]

Ahhhh yeah, wet noodle rockets and planes.  Thought it was just me overcooking my designes.  Flying around enjoying the new graphics and wow did my plane wobble like a wet noodle.  Rockets bending in half - must rethink my designs... no strut to heaviest part trick until it's added.   

I am tempted to build something large in orbit, but the kraken killed my 5 comm sat network deploy craft so I am a bit shy about the orbital kraken.  

I will have to experiment with the truss thing you mentioned dvrabel.

[xProphetx]

From my observations... the problem seems to be that the 'anchor' is the very center and the actual area of the object is not considered.  There is no collision detection at the edges of the 'cylinders'.

Honestly I was hoping with KSP2 this kind of stuff was fixed.  I'm very disappointed that this specific game mechanic persists.  I don't do game coding and whatever, but I do work on other complex systems, this just doesn't seem like a problem that should have existed in 1 considering the popularity and the amount of mods and the time it lasted before 2.  And for 2 to still.... 

Really disappointed.

[Vl3d]

 

[Periple]

I’m fairly certain that in KSP2 only decouplers and docking ports flex. Stacks of tanks, planes, etc. are rigid. 

Good move IMO. Nobody seems to have noticed and it massively simplifies the physics. Although TBH I’d rather get rid of the noodles altogether, it doesn’t really add meaningful gameplay. If you want to stop huge masses from being held up on tiny joints just make them easier to break.

[J.Random]

13 minutes ago, Periple said:

Stacks of tanks ... are rigid. 

Nope.

Extreme example but proves that the flex is there anyway.

Edited February 26, 2023 by J.Random

[Periple]

I stand corrected.

[Wheehaw Kerman]

It’s nothing that we haven’t seen back in the janky old days before autostrutting in KSP1.  Conservatively designed craft with reasonably placed strutting fly just fine.

[The Aziz]

On 2/25/2023 at 2:17 AM, obney kerman said:

Idea: When you close a fairing by connecting it to another part, let that fairing act as a strut and hold the whole thing together, instead of having the top part of the craft wobble anyway and clip through the fairing! Something to make Saturn V replicas and interstage adapters a lot easier.

THAAAAAAAT. ↑

Interstages should connect to both sides. 

With that said, I wonder if they protect from the aero forces at all..