Unexpected change of orbit

[kurja]

Look at the screenshots, in particular, orbit of that station. I expect it to merrily orbit Jool, while my lander ship visits Laythe. But just a little later, the station has decided it should make a vertical dive to Laythe instead. I'd dismiss this as a rounding error or such flakyness in the conics system but I went back to an earlier save a couple times and even if I lower the station's orbit by a dozen kilometers or so, this still happens. A known bug? A new bug? Something else? And how to avoid in the future

Edited March 15, 2014 by kurja

[Jouni]

The way gravity is implemented in KSP doesn't work that well in the Jool system.

The spheres of influence of the inner moons of Jool are very small compared to their gravity. When the station touched Laythe's sphere of influence, Jool's gravity ceased to be a factor in its orbit. (Similarly, Laythe's gravity wasn't a factor until the transition happened.) Because the station was at a lower orbit around Jool than Laythe, it had some velocity relative to Laythe after the transition. But due to the small sphere of influence, Laythe's gravity is quite significant at its edge of its SoI, so that velocity wasn't enough to escape Laythe again.

Laythe is at a 27184 km circular orbit around Jool, and the radius of its sphere of influence is around 3724 km. This means that orbits below 23460 km should be safe from Laythe's influence, while higher orbits are not. Add some margin of error, because orbits tend to change a bit, when objects transit between rails and physics simulation.

[Supernovy]

Your station was in a very close orbit to Laythe, and not synchronous with it. It would fall into Laythe's SoI eventually.

I suggest looking up Laythe's SoI radius on the wiki and putting the station in an orbit at least that much different from Laythe's.

As to why it fell straight down, you can tell from the very similar orbits that the station and Laythe had a very low relative velocity. Once the station got into the SoI, it didn't have much horizontal velocity and gravity pulled it directly down. This didn't show up in map view because SoI changes aren't shown for on-rails vessels.

[Skorpychan]

Move your station WELL away from Laythe. I've had repeated problems with this around Duna, with the station and tugs slamming randomly into Ike during timewarp, or while mechjebbing.

Give muns a wide berth.

[SRV Ron]

Looking at the data, your satellite will dip into Laythe's atmosphere resulting in sufficient aerobraking to give it the final fatal flight path.

[kurja]

Laythe is at a 27184 km circular orbit around Jool, and the radius of its sphere of influence is around 3724 km. This means that orbits below 23460 km should be safe from Laythe's influence, while higher orbits are not. Add some margin of error, because orbits tend to change a bit, when objects transit between rails and physics simulation.

Station is far lower than that, at about 18 thousand kilometers. Even though, my beef here is primarily that the soi change appears out of thin air; at first it isn't there, and a minute later, I'm going straight down.

[kurja]

Looking at the data, your satellite will dip into Laythe's atmosphere resulting in sufficient aerobraking to give it the final fatal flight path.

Sorry, what? The ship with a Laythe Pe? Nothing strange going on with it's orbit as far as I can tell.

[SRV Ron]

Station is far lower than that, at about 18 thousand kilometers. Even though, my beef here is primarily that the soi change appears out of thin air; at first it isn't there, and a minute later, I'm going straight down.

Laythe has an atmosphere starting somewhere around 60,000 meters.

Aerobraking doesn't appear to be factored into the maneuver. Had this been any other moon in the Jool system, the maneuver would have taken placed as planned.

[SRV Ron]

Sorry, what? The ship with a Laythe Pe? Nothing strange going on with it's orbit as far as I can tell.

Update. Noticed the probe along with your capsule.

Apparently, the probe you dropped off prior to the capsule being captured by Laythe is also being capture by Laythe, but that capture path didn't show up until later in the second screenshot. Since its path is similar, atmosphere drag is causing it to crash to the surface.

Edited March 14, 2014 by SRV Ron

[kurja]

Pe is at 27,380 meters, well inside Laythe's atmosphere. It's drag has not been calculated in the maneuver displayed on the map mode. You are far enough into the atmosphere that the probe will crash to the surface.

Yes, I know. What about it?

[kurja]

Laythe has an atmosphere starting somewhere around 60,000 meters.

Aerobraking doesn't appear to be factored into the maneuver. Had this been any other moon in the Jool system, the maneuver would have taken placed as planned.

What maneuver?! I've no idea what you two are talking about?!?

edit - oops, it was just you with two posts, not two separate users, sorry

[Jouni]

Station is far lower than that, at about 18 thousand kilometers. Even though, my beef here is primarily that the soi change appears out of thin air; at first it isn't there, and a minute later, I'm going straight down.

Can you check that orbit? Visually it appears to be a lot higher than 18000 km.

The problem is in the way KSP implements gravity. When you are in a moon's sphere of influence, the planet's gravity affects you in the same way as it affects the moon's center. On the other hand, if you are just outside the sphere of influence of the moon, the moon's gravity doesn't affect you at all. Normally this approximation works pretty well, but if you have a large moon orbiting a massive planet at a low orbit, all kinds of strange things can happen.

[kurja]

Update. Noticed the probe along with your capsule.

Apparently, the probe you dropped off prior to the capsule being captured by Laythe is also being capture by Laythe, but that capture path didn't show up until later in the second screenshot. Since its path is similar, atmosphere drag is causing it to crash to the surface.

There is no probe. Take another look at the first post; there's a station, that appears to be in steady orbit around jool, and then the next moment it's crashing down instead. The ship, the one with a Laythe Pe in both screenshots, is heading for an aerobrake like it should, nothing strange about that.

[SRV Ron]

Can you check that orbit? Visually it appears to be a lot higher than 18000 km.

While under Jools SOI, it is fine. Then, it comes under Laythe's SOI which then alters the station's path to the surface via aerobraking. That is clearly shown for the capsule where its Pe is well inside Laythe's atmosphere.

[Jouni]

Ok, my mistake. The orbital characteristics listed in the wiki are based on Jool's center, not its surface. This means that safe orbits around Jool are below 17460 km instead of below 23460 km.

[SRV Ron]

There is no probe. Take another look at the first post; there's a station, that appears to be in steady orbit around jool, and then the next moment it's crashing down instead. The ship, the one with a Laythe Pe in both screenshots, is heading for an aerobrake like it should, nothing strange about that.

Sorry about the initial confusion.

Your station has come under Laythe's SOI which is not apparent in the first screenshot.

[kurja]

The orbit is almost perfectly circular 18 thousand kilometers.

Maybe I should clarify what the issue actually is. Forget aerobraking, forget the Periapsis, those have nothing to do with the issue. I'll add another screenie from an earlier moment; the station is happily orbiting at 18Mm, untill, boom. There is no indication of an impending encounter, no nothing, it just goes from a steady orbit to a crash. Same happens if I lower the orbit by some 50km to rule out rounding errors or such.

[Jouni]

Maybe I should clarify what the issue actually is. Forget aerobraking, forget the Periapsis, those have nothing to do with the issue. I'll add another screenie from an earlier moment; the station is happily orbiting at 18Mm, untill, boom. There is no indication of an impending encounter, no nothing, it just goes from a steady orbit to a crash. Same happens if I lower the orbit by some 50km to rule out rounding errors or such.

Essentially the 18000 km orbit is going deep into Laythe's sphere of influence, when there is a close encounter. To get out of Laythe's way, lower the orbit by at least 600 km.

[kurja]

Essentially the 18000 km orbit is going deep into Laythe's sphere of influence, when there is a close encounter. To get out of Laythe's way, lower the orbit by at least 600 km.

well okay, my present orbit entering Laythe's soi I should get an encounter in map view, right?

[Jouni]

well okay, my present orbit entering Laythe's soi I should get an encounter in map view, right?

You should see the encounter on map when you are less than one orbit around Jool from the encounter. Does that happen?

[kurja]

You should see the encounter on map when you are less than one orbit around Jool from the encounter. Does that happen?

No.

ten char.

[Jouni]

So you don't see the encounter on map, when the space station is the active ship, and it's going to happen during the current orbit around Jool? If so, then you should file a bug report, describing the situation in detail.

[KerikBalm]

I've had similar "encounter not showing up" things happen before... even with minmus, but especially with Duna, it also happened with moho.

I will set up my trajectory, get an intercept, and then at some point along the journey, the encounter dissappears... yet if I continue on, the encounter actually still happens.

This is a case of a planned encounter that is hard to plan for, I hadn't even considered an unintended encounter.

As near as I can figure, the encounter predictions are unreliable

[kurja]

I've had similar "encounter not showing up" things happen before... even with minmus, but especially with Duna, it also happened with moho.

I will set up my trajectory, get an intercept, and then at some point along the journey, the encounter dissappears... yet if I continue on, the encounter actually still happens.

This is a case of a planned encounter that is hard to plan for, I hadn't even considered an unintended encounter.

As near as I can figure, the encounter predictions are unreliable

I've had that happen when trying to get an encounter with a distant body, I've just always blamed rounding errors but this case is a little different - craft is rushing straight at the body but soi change indication won't appear at all, soi just changes at one point and that's it.

[Kryxal]

I can verify the numbers for the station's orbit being too high ... remember that your 18,000 km is above the SURFACE, so 24,000 km from the center, center-to-center for Laythe is 27,184 km, SoI is 3,723 km (and change), so you end up about 540 km inside the SoI.

No clue why it didn't actually show it coming, though...