How big can we go?

[KerbMav]

This is not my video, to say it from the start.

But I watched it and wondered: What problem did he encounter? How big can we actually build? What would stop this shaking and breaking? (Which was a long while back a problem of mine while building a much smaller station already.)

 

 

[Rayder]

10:48 - Oh my god...

Previously the problem (even if you got everything right) was navigating the station with the camera. Since 1.2 lets you orient the camera on any part you wish, I'd say limits would be:

- Computer processing horsepower
- Launch vehicles
- Structural stability

If you're thinking absurdly large, the physics render distance is the ultimate hard limit.

Edited September 25, 2016 by Rayder

[cubinator]

I'd estimate it's probably about 3+-1 orders of magnitude more parts and he could build a legit Star Destroyer.

[problemecium]

Actually several people, most famously Danny2462, have demonstrated that although 2.5km is the limit for physics simulation of separate vessels, any and all parts of the current vessel have physics applied regardless of distance - thus this is not a hard limit on vessel size at all. In fact no video I've seen has documented any hard limit on vessel size; once I saw someone (I believe Nexter's Lab) flying a USS Enterprise replica roughly the size of the Mun.

EDIT: I have now watched the whole video and now I know what I'm doing in KSP today. Or at least whenever I get back to KSP since today my grandparents are in town, but you get the point xD

Edited September 25, 2016 by parameciumkid
I lost The Game!

[swjr-swis]

A quick slap together of parts in 1.2.0.1532 show that judicious use of the new autostrut tweakable can successfully rigidize a 1km cargo beam, even once cheated into space. I imagine this will work for much larger too. Should be easier to find the limits of the system.

Proof (pics, craft file and save file) in this Dropbox folder link.

[Ultimate Steve]

I am having the shaking problem with an only 100m long mothership.

8 minutes ago, swjr-swis said:

A quick slap together of parts in 1.2.0.1532 show that judicious use of the new autostrut tweakable can successfully rigidize a 1km cargo beam

I should probably start playing my mission report in 1.2 then. I want to build bigger! (Plus my screenshots folder is at around 2000 now)

[Gaarst]

This thread is full of big things. One noteworthy rocket is @AstraNebula's 1600m tall rocket:

He made orbit in RSS with it.

[Xavven]

I wonder if this is even possible IRL. Would gravitational tidal forces cause structural failure?

[Deaf3279]

love this station but sad it dont work without hack 

[Snark]

6 hours ago, Xavven said:

I wonder if this is even possible IRL. Would gravitational tidal forces cause structural failure?

Structural failure no, maneuverability yes.

If you could build a real-life space station that was long, skinny, and a few kilometers from one end to the other, tidal forces would tend to make it want to orient itself with the long axis pointing through the planet's center.  The actual mechanical stress involved, however, would be pretty small.  Even a 6km station is under 1/1000 of the length of its orbital radius, so tidal stress would be a small fraction of 1% of one gravity.

Might be an issue if you try to keep its orientation something other than radial, though, since a structure that long is going to have poor stiffness. Tension is fine, compression not so much.

Getting kind of off-topic, though, so I'll stop there. 

[Sharpy]

[nvm]

Edited September 26, 2016 by Sharpy

[MarkusA380]

Still waiting for the space elevator.

BUT

Having the kraken select you as it's next victim is one thing,

summoning it deliberatley is another.

[Sharpy]

12 hours ago, Snark said:

If you could build a real-life space station that was long, skinny, and a few kilometers from one end to the other, tidal forces would tend to make it want to orient itself with the long axis pointing through the planet's center.  The actual mechanical stress involved, however, would be pretty small.  Even a 6km station is under 1/1000 of the length of its orbital radius, so tidal stress would be a small fraction of 1% of one gravity.

Might be an issue if you try to keep its orientation something other than radial, though, since a structure that long is going to have poor stiffness. Tension is fine, compression not so much.

Getting kind of off-topic, though, so I'll stop there. 

Strategically distributed RCS thrusters and tensometers could efficiently combat any dangerous stress build-up. Think of the station as a swarm of independent satellites staying in formation through use of RCS, instead of a single rigid structure.

[ZooNamedGames]

If you really want to test the limit whilst maintaining integrity use procedural parts and Kerbal Joint Reinforcement. Let me know what you can do with this combo  .

[problemecium]

I spent today running some experiments on the feasibility of this endeavor, with disappointing results.

At first I tried building 100m segments and docking them in orbit. I constructed a launch vehicle which, after dozens of adjustments, had the right arrangement of struts to remain stable through launch and orbital insertion. When I rendesvoused two together, though, for some reason one of them started oscillating with increasing violence despite both craft being identical.
Further tests revealed that when a long series of Mk3 cargo bays are chained together, for some reason small perturbations end up being amplified as they travel up and down the length of the craft. The effect is greatly reduced if all the cargo bays are empty such that the craft's total mass is low, but it appears that stability beyond this point depends on pure luck.
Presuming that this was a flaw in the Mk3 cargo bay, I tried again with various Mk3 fuel tanks, to the same results. Wondering if this was a problem with the entire Mk3 fuselage family, I tried again with Rockomax Jumbo-64 fuel tanks (pictured; the thermal overlay is on to enhance visibility).
Strangely enough the oscillation problem remained. The test vehicle split into two, and each half of it continued to squirm more and more violently until every fragment more than a few tanks long had exploded.
What I think I have determined via these tests is that KSP's PhysX implementation lacks a failsafe to ensure that a force caused by another force is never greater in magnitude than the original force - this would also explain the odd behavior I've observed with landing gear wherein they start bouncing violently until the craft flips over ("wheels are fixed now" my butt).

I'm given hope by @swjr-swis's example, but I remain mistrustful of the new auto-strut feature and have found that it has only partially helped in the designs I've tried so far.

[swjr-swis]

8 minutes ago, parameciumkid said:

I'm given hope by @swjr-swis's example, but I remain mistrustful of the new auto-strut feature and have found that it has only partially helped in the designs I've tried so far.

I've been able to make airplanes safely pull 99G turns at Mach 3 the past few days, making them effectively unbreakable, using nothing but a few autostruts; where in 1.1.3 and before they would've spontaneously disassembled at below Mach 2 turns even with tons of regular struts. And the cargo bay construct in my previous post here was made rigid the same way - without them it started to undulate like a giant caterpillar the second it was put in orbit. Autostruts definitely work, but they have to be placed selectively.

Use the minimum possible, try to place the root part and the heaviest part on opposing ends of your construct, and alternate autostruts between heaviest and root (and where appropriate, with grandparent). Also it helps if the heaviest part is a symmetric set, because the autostruts will connect to all of the symmetric group, essentially creating a truss-like connection.

[KerbMav]

OK, the size seems to be somewhat atainable - but what do you guys think stopped the docking ports from docking to yet one more vessel?

@parameciumkid, @swjr-swis, @Snark

[Snark]

1 hour ago, KerbMav said:

but what do you guys think stopped the docking ports from docking to yet one more vessel?

...searches thread for the phrase "docking port"... not finding anything except the above quote...

Would be a little easier to answer (without having to go through the whole thread with a fine-toothed comb) if you could quote the thing you're asking about? 

[problemecium]

1 hour ago, KerbMav said:

what do you guys think stopped the docking ports from docking to yet one more vessel?

My guess was that it was because @TapeGaming was adding single segments onto the end of the already very long station, and the docking port on the last segment was so far away from the root part of the station that Physics was partially disabled for the station, such that collisions remained functional but the docking ports did not. If this is the case then the solution would be to build two 500-meter segments and then join them together to complete the spine.