Kerbal Space Program

[Asbo Zaprudder]

Anyone know if the developers plan to implement a multibody gravity simulation? From what I can tell, they currently use 2-body solutions and bounded domains of gravitation influence for calculating spacecraft trajectories - so there are no gravitational slingshots and no Lagrange points in the Kerbol system. One might have thought that they could have implemented a restricted 3-body or a perturbation-based algorithm.

[TheGodBen]

Asbo Zaprudder wrote:

---

There's a wiki page somewhere out there that lists a whole lot more Easter eggs - so I'm boldly going where everyone seems to have gone before - sort of like the 10,000th person to climb Everest has to scramble over all the dead bodies (well, until they removed them a year of two ago).

Don't forget the monoliths, I don't think the MapSat mod locates those. The closest is 2km north of the space centre. I've seen it sometimes during launch, but never bothered to actually visit it until today.

Asbo Zaprudder wrote:

---

Anyone know if the developers plan to implement a multibody gravity simulation? From what I can tell, they currently use 2-body solutions and bounded domains of gravitation influence for calculating spacecraft trajectories - so there are no gravitational slingshots and no Lagrange points in the Kerbol system. One might have thought that they could have implemented a restricted 3-body or a perturbation-based algorithm.

They've said that the physics engine they're using can't handle n-body calculations, so that's never going to be implemented and we're not likely to see any sort of Lagrange points simulated. But gravitational slingshots are possible in the game, just do a close flyby of Mun and you can find yourself flung out into interplanetary space if you time it right. The Joolian system is an awesome playground for slingshot manoeuvres because of all its moons. During my expedition there in 0.17, I dumped two empty fuel tanks before heading back to Kerbin, and those fuel tanks bounced around gravity-wells in such a way that they picked up enough speed to escape the solar system.

The game also simulates the Oberth effect, so you'll get far better fuel efficiency if you perform your interplanetary burns while in Low Kerbin Orbit than you will if you escape Kerbin's gravity-well and then perform the burn. Last I heard, Mechjeb doesn't know how to do this yet, so that's one advantage to flying manually.

[Asbo Zaprudder]

I suspect the gravitational slingshot is an artifact of the domain of influence approximation that they're making and the timing of the crossing of the boundary. If you approach a planet retrograde to its orbital motion rather than direct, you should lose rather than gain momentum relative to the system. I haven't tried to see if this is the case for the Kerbol system. I think a simple 2-body algorithm should conserve the energy of the each interacting body although their momentum vectors will change although the sum will remain unchanged. So long since I thought about this stuff that I'm probably talking bollocks.

BTW if a Kerbie touches a monolith, does Also Sprach Zarathustra ring out and the little guy get a bit brainier? I suppose the Mun and Kerbol have to be in the right alignment. That's another thing - I wounder if there are eclipses.

[sojourner]

Yeah, the N-body limitation eliminates lagrange points, but gravity boosting works just fine.

There is a sun, there are planets, there are moons. Various ones have day/night cycles. How can you not have eclipses? Or do you mean eclipses where the sun just happens to have the same visual size of the eclipsing body like with our sun and moon?

[Asbo Zaprudder]

Snowjourner wrote:

---

There is a sun, there are planets, there are moons. Various ones have day/night cycles. How can you not have eclipses? Or do you mean eclipses where the sun just happens to have the same visual size of the eclipsing body like with our sun and moon?

I meant eclipses from the surface of Kerbin. I have the figures:

Kerbol radius R = 261,600 km
Mun radius r = 200 km
Kerbin - Mun mean distance d = 12,000 km
Kerbin - Kerbol mean distance D = 13,600,000 km

As R/D = 0.019 radians, and r/d = 0.016 radians, eclipses would probably be annular only. From the figures I have I believe the Mun's orbit is aligned with the plane of the ecliptic so there should be such an eclipse every orbit.

Objects do appear to cast shadows, but I'm now wondering if, say, distant objects such as the Mun do. The shadow for an annular eclipse wouldn't be as pronounced as for a total eclipse, of course.



ETA regarding the slingshot after thinking about it some more. For a 2-body encounter in the frame of reference (FoR) of the planet, the magnitude of the non-propelled spacecraft's entry and exit velocities are the same, although the direction vectors are different, of course. However, seen from the FoR of a third body, such as Kerbol, the magnitude of the entry and exit velocities would appear to differ by up to twice the speed of the planet relative to the third body - the sign of the difference and its magnitude depending on the direction of the encounter relative to the planet's orbital velocity vector. The planet also loses or gains a tiny fraction of its momentum seen from this FoR. So, multibody calculations aren't required for this aspect of the game as Sojourner states, although they wouldn't be sufficiently precise for real probes.

[TheGodBen]

Asbo Zaprudder wrote:

---

BTW if a Kerbie touches a monolith, does Also Sprach Zarathustra ring out and the little guy get a bit brainier?

Sadly, I couldn't reach the monolith as it was floating and the gravity on Kerbin is too high for jet packs. But something magical does happen if you attempt to touch the monolith on the magic boulder: You explode.

Asbo Zaprudder wrote:

---

As R/D = 0.019 radians, and r/d = 0.016 radians, eclipses would probably be annular only.

Yup, annular eclipses happen once every Munar orbit. On Mun, Kerbin completely eclipses the sun once every orbit.

Objects do appear to cast shadows, but I'm now wondering if, say, distant objects such as the Mun do. The shadow for an annular eclipse wouldn't be as pronounced as for a total eclipse, of course.

Planets effectively do cast shadows, just not visually, which is why solar panels don't work on the dark side of planets but you can still see the ships. During an eclipse on Kerbin, the game knows that Mun is blocking light and solar panels don't work, but it remains as bright as normal.

They were working to implement planetary shadows for 0.18, but it was a bit buggy and was held back for the next major release. They showed a spacecraft on the dark side of Kerbin that was barely visible due to the lack of illumination, and said that the effect should work for eclipses as well. So that's something to look forward to in 0.19.

[Asbo Zaprudder]

I'd like a night-vision or IR feature - the spotlamps are somewhat crappy. Even the illumination in the VAB can be pretty bad.

ETA - Oh yes, and I'd like a long f spy camera. I think there are some camera modules, but these don't work so well at the moment. A camera and docking target would also be very useful for docking.

[sojourner]

I'd like more precise placement on struts. Using them now is the easiest way to make a rocket look put together by rednecks.

[Asbo Zaprudder]

Yeah, also a finite element analysis and flight simulator for testing your design before you waste all those Kerbal dollars seeing it blow up before it gets to orbit, bringing a poor Kerman's life to a premature close.

ETA latest tally of anomalies after orbital surveys:

Kerbin 8
Mun 7
Minmus 1
Moho 1
Eve 0 (under mercury?)
Gilly 0
Duna 3
Ike 1
Jool 0
Lanthe 0 (under water?)
Vall 1
Bop 1
Tylo 2
Pol 0
Dres 0

[TheGodBen]

What I really want them to focus on in the next update is a new aerodynamic model (including payload shrouds), and re-entry heat. The new mining and colonisation features the devs have talked about sound great, but right now critical elements of the simulation are substandard or missing entirely and I wish for those to be addressed before new things are added.

[Asbo Zaprudder]

I agree - improving the aerodynamic and re-entry models is sorely needed. There are iffy things like the model for solar-power generation drop off not being inverse square law and the multibody problems discussed earlier, but they're not big issues given the somewhat cartoonish nature of the Kerbal universe.

I'm currently thinking about how I can modify existing parts to simulate an Orion project nuclear pulse engine with an Isp of 10,000s and a thrust of 4 MN. For comparison, each F1 engine on the Saturn V S-1C stage had an Isp of 263s and a thrust of 6.8 MN. So I guess it would be like a _very_ powerful ion drive. Perhaps I can modify a decoupler to keep going bang. For the Earth, I'd need 800 140-tonne yield bombs each weighing 0.5 tonnes to put 1,600 tonnes into orbit. Going to need to recalculate for Kerbin...

ETA: As I don't know how to use Blender, I scaled up the mini Oscar fuel tank (as the nuke dispenser), the xenon tank (as a large gas bag shock absorber), and the ion engine (as the pusher plate and mechanical shock absorbers) by a factor of 16 in each linear dimension, made a new resource "Nukes", and set the engine parameters to Isp = 10,000s and thrust variable between 0 and 6 MN. The engine is tuned to destroy itself if you go higher than 4 MN for too long. Goes like a bat out of hell (pulling max 7Gs with a tiny payload), but as its mass is several hundred tonnes, it's a bugger to steer. I think I'll have to use large chemical rockets as vernier thrusters.

[sojourner]

Man, I understand the reason that "wobble" is in the game, but they need to reduce it's effect. Makes building a large, graceful craft neigh impossible.

[Asbo Zaprudder]

Perhaps what I'd really prefer is an retooling of Buzz Aldrin's Race into Space - maybe with a longer time span and a more diverse tech tree. As you might guess, I'd have a strong interest in seeing Orion nuclear pulse propulsion or Truax's Sea Dragon added as possible paths. KSP is fun enough, but I'm not a fan of the interface - too much button jabbing and potential for RSI.

[TheGodBen]

Snowjourner wrote:

---

Man, I understand the reason that "wobble" is in the game, but they need to reduce it's effect. Makes building a large, graceful craft neigh impossible.

One weird thing about the latest update is that engine fairings make rockets look solid and cohesive, yet provide no structural support. This is particularly bad for the wide fuel tanks and engines, they wobble like crazy, and the only solution is to place struts across the fairings. In the old versions, it made some visual sense for rockets to wobble like that because it was literally just an engine on top of a decoupler, but in the new version it's just really dumb. Hopefully they'll fix it in the next version.

I'm really liking the unmanned probes, they're light and there's no expectation to plan for a return journey, I can see why NASA likes them so much more than manned missions. My current goal is to get a satellite to orbit and a lander to touch down on every world in the game. The difficult part is keeping track of all the launch-windows/orbital adjustments/intercepts, it's like rocket juggling.

[sojourner]

^It's worse than you think. Engine fairings and nose cones actually make your rocket worse. They increase drag, weight and complexity while giving no benefit beyond looking cool in the current version of the game.