Rocket noodling? Oh gods, not again...

[J.Random]

Oh wow. Especially seeing how the SAS controls regressed to before KSP 1.1 (maybe further, can't recall when planes stopped imitating flappy bird), that's a recipe for anything slightly complex assembled in orbit to shake itself apart. So far it seems that KSP2 follows the development of KSP1, making all the same mistakes. Expecting broken wheels and collapsing landing legs now. And terrain colliders not following the visuals (or vice-versa), making rovers drive under terrain.

(inb4 infinite acceleration powered by control surfaces' twitching is back as well)

[Tweeker]

8 hours ago, razark said:

I saw one video where it was pointed out that it did not exist.

Progress!

 

[cfds]

Yeah, the fact that they still use the weird "Indian rope trick" physics, coupled with the displayed performance and the absurd price tag, I won't touch the game until it is really released. And steeply discounted.

[theJesuit]

8 hours ago, razark said:

I saw one video where it was pointed out that it did not exist.

The way I read this was that the video was stating the video did not exist.  Like writing on a whiteboard with a marker - "this whiteboard is empty"

[Columbo]

Auto-Strut seems like such a basic function of KSP 1 at this point that i can't believe its not part of EA from day 1. This and reentry heating. Those 2 things are deal breakers for me. I won't buy into EA until those are in the game.

[Superfluous J]

I 100% agree. Poorly built rockets should not wobble about like noodles.

They should explode, and you should have to go back to the VAB and fix your poorly designed rocket until it's structurally sound enough to fly.

Or you could just do that in the first place and let the people who like to over-engineer their rockets and have some wobbliness, have their fun too.

Yeah I changed my mind they should keep it the way it is.

25 minutes ago, Columbo said:

Auto-Strut seems like such a basic function of KSP 1 at this point that i can't believe its not part of EA from day 1. This and reentry heating. Those 2 things are deal breakers for me. I won't buy into EA until those are in the game.

[AtomicTech]

17 hours ago, Jarin said:

"Slaying the Kraken"

Seeing Matt Lowne, SWDennis, and Reid Captain's footage was really disappointing, especially when they're trying to make the EA just be a replacement for KSP 1. I'm aware that it's still an early build but I'd appreciate it if we at the bare minimum got autostrut. If we can get that on release day, I'll happily shut up

[AlphaMensae]

EJ_SA during his stream yesterday did say that he asked about autostruts to the devs. The response was "We'll think about it", so they know about it at least.  Best thing to do is express your interest in autostruts here and on the KSP2 discord.

And KSP1 didn't have autostruts for a long time, and we lived with it (or installed KJR). How spoiled we have become.  I'm sure KJR will be made for KSP2 soon enough.

Edited February 21, 2023 by AlphaMensae

[GigFiz]

This is one reason why I think EA will be helpful: as far as struts, wobble, and joint rigidity, and the like go, they can only test it so much on different craft (cuz, y'know, they are still working on developing the game and only have so many man hours available for test launching lots of different designs under countless different launch conditions); opening the game up for this kind of public beta testing allows the the sample size, and variety of designs that are tested to go up by several orders of magnitude. This will give much more info about the effects and unforeseen quirks of the joint settings, and make it much easier to tune it to the place where they want it

To me, that's what the "we'll think about it" probably means: if they can get to a good place with the physics and joint systems they have in place (presumably a nice, and realistic, middle ground between kraken attacks and and rigid indestructability), that would be ideal, but if, for whatever reason, it proves more difficult than they are hoping it will be, then autostruts will re-enter the conversation.

[Jacke]

The wobbly rockets of KSP v0.25 almost killed my desire to play the game.  Fought and fought and kept trying and finally got a rocket to orbit and was completely exhausted.  The only thing that saved me was MechJeb's autopilot that could fly the rockets without them falling apart.  Which is how real rockets are flown.

I'm not pleased there's so much flex in even this version of KSP 2.  What alarmed me from the Scott Manley video was how janky controlling the rocket appeared to be.  Like there's too much control authority and it would just lead to pilot-induced oscillations.  As the control is coming from the engines gimbling, there's a control setting that's not right.

[regex]

19 minutes ago, Jacke said:

What alarmed me from the Scott Manley video was how janky controlling the rocket appeared to be.  Like there's too much control authority and it would just lead to pilot-induced oscillations.  As the control is coming from the engines gimbling, there's a control setting that's not right.

IIRC in KSP1 the engine gimbal was severely tuned back due to the earlier SAS version (like 0.20 and before?) being particularly aggressive. IRL atmospheric rocket engines have much wider gimbal ranges. It could be that the KSP2 team has used higher ranges than KSP1.

Either way, that's a target for a first initial mod, SAS tuning.

[EnderKid2]

IIRC the devs said during the event that the SAS's PID controllers need tuning, especially for aircraft

[Drakenred65]

16 hours ago, Columbo said:

Auto-Strut seems like such a basic function of KSP 1 at this point that i can't believe its not part of EA from day 1. This and reentry heating. Those 2 things are deal breakers for me. I won't buy into EA until those are in the game.

The odds are reentry heating was disabled because they expected the press at least to end up trying to reenter at speeds exceeding solar system escape velocity. Not Kerbol system EV. . Solar system EV.

[TheOtherDave]

2 minutes ago, Drakenred65 said:

The odds are reentry heating was disabled because they expected the press at least to end up trying to reenter at speeds exceeding solar system escape velocity. Not Kerbol system EV. . Solar system EV.

Hah! Now that you say it, I wouldn’t be surprised. I’m not even that good, but I know I’ve been playing long enough to forget what it’s like to not know the basics.

[mcwaffles2003]

A big problem no one seems to mention is that wobbly rockets also killed game performance.... The entire reason I got KJRwhen I did was for the performance improvements over the stock wobbly rocket. I think joints should be tighter dependent on the smallest connecting radius... its ridiculous for a 5m part to demonstrate the same flex as a 0.625m part.

[garithmar]

a rocket that sized shouldn't wobble almost at all so the game needs some serious tweaking when it comes to rigidity.  This problem becomes problematic when you include ships the size and scope of interstellar ships with all the mods for life support, radiation shielding etc.  I would highly recommend a rigidity pass on the game and make it so even large rockets don't wobble at all unless badly built.  This could include things like putting a smaller sized decoupler below a much larger rocket tank and pod without struts thus causing a weak point that will wobble till it breaks.

[tstein]

8 hours ago, mcwaffles2003 said:

A big problem no one seems to mention is that wobbly rockets also killed game performance.... The entire reason I got KJRwhen I did was for the performance improvements over the stock wobbly rocket. I think joints should be tighter dependent on the smallest connecting radius... its ridiculous for a 5m part to demonstrate the same flex as a 0.625m part.

IN fact stacks of rockets shoudl behave as a single piece up to some X size.

[mcwaffles2003]

1 hour ago, tstein said:

IN fact stacks of rockets shoudl behave as a single piece up to some X size.

i think some flexing would be okay, but nothing close to what we are used to. I think the issue is a difficult one to resolve though as making better joints would probably require more points of attachment, 3 at least, and that would take extra processing to simulate

[tstein]

4 minutes ago, mcwaffles2003 said:

i think some flexing would be okay, but nothing close to what we are used to. I think the issue is a difficult one to resolve though as making better joints would probably require more points of attachment, 3 at least, and that would take extra processing to simulate

Not at all. THe contrary in fact. The  game coudl for example REMOVE the joints from the simulation of any stack of fuel tanks that has less than X times the width. That  means we would  have some bending on larger rockets but none at all in  compact rockets.

[mcwaffles2003]

Just now, tstein said:

Not at all. THe contrary in fact. The  game coudl for example REMOVE the joints from the simulation of any stack of fuel tanks that has less than X times the width. That  means we would  have some bending on larger rockets but none at all in  compact rockets.

And just fuse them into a single piece? If thats the case then you wouldnt be able to simulate stress on the joint and it would be simple rockets

[tstein]

Just now, mcwaffles2003 said:

And just fuse them into a single piece? If thats the case then you wouldnt be able to simulate stress on the joint and it would be simple rockets

You fuse them  up to a certain height. Real rockets are closer to simple rockets than to KSP in the way they behave regarding that.  Engineers are not idiots to make a simple 4m tube bent at less than 1 G acceleration.

[CastleCustodian]

The problem is that the simulation is the opposite of reality. In the real world long solid objects flex. While joints break when stressed too much. In kerbal the simulation is doing the opposite. joints flex and if they flex too much the straight section goes poof. Is it possible to make the model closer to reality without killing computers?

[cfds]

44 minutes ago, CastleCustodian said:

The problem is that the simulation is the opposite of reality. In the real world long solid objects flex. While joints break when stressed too much. In kerbal the simulation is doing the opposite. joints flex and if they flex too much the straight section goes poof. Is it possible to make the model closer to reality without killing computers?

Since there were bridge building games in the late 90s that simulated exactly this and worked fine on the computers of their time: Yes.
But if your only effort is to tweak parameters provided by the Unity implementation of a chained rigid body like some hobbyist you will probably not get there.

[tstein]

1 hour ago, CastleCustodian said:

The problem is that the simulation is the opposite of reality. In the real world long solid objects flex. While joints break when stressed too much. In kerbal the simulation is doing the opposite. joints flex and if they flex too much the straight section goes poof. Is it possible to make the model closer to reality without killing computers?

Non rigid body simulations tend to be more CPU intensive unless you  can  create a very very simplified model based on a simple torque function. But it woudl create some strange  things.

[CastleCustodian]

Just now, tstein said:

Non rigid body simulations tend to be more CPU intensive unless you  can  create a very very simplified model based on a simple torque function. But it woudl create some strange  things.

To be fair the current noodle is really strange. Also it is responsible for some aspects of kraken attacks.