Odd top-heavy lander issue--looks like a bug?

[tomchaps]

It's been a few updates since I've played, but I was piloting a trusty, simple lander towards Minimus with my son today. A few times while doing a simple retrograde burn, the craft suddenly spun wildly out of control. It acted as if there was a ton of weight in the nose, which wanted to point down no matter what. SAS did nothing. I repeatedly regained control, lost it... It was a crazy unstable equilibrium.

Our landing was perfect... 0.0 m/s at about 2m above the surface, engines off--and then the thing flipped and started cartwheeling end over end. I landed it after a few bounces, but when the Kerbal got out for an EVA, the simple act of letting go made the lander tip over--and it ended nose in the ground, with the legs bobbing gently up and down at a 60 degree angle. Walking nearby makes the craft tip over flat, and then it magically tips back up, legs in the air, balanced on its nose. It's like antigravity or something.

We've landed on Minimus a hundred times by now, but I've never seen anything like this, or understand the physics.

Anyone have any idea what's going on?

[Xannari Ferrows]

What does your craft look like, and how is it built? A small asymmetric displacement of parts on a small lander can make piloting a real chore. Not so much for a massive lander.

[Deutherius]

The flipping when landing issue might be caused by SAS - are you sure you used the correct mode?

If you click on radial in, the ship will try to point its nose to the body of the SOI you are in (which would explain what you described. From the physics standpoint, being top heavy would not influence the passive rotation of the craft in any way while in space - at least not in KSP (as gravity gradient is not modelled) - but it's a useful analogy).

If you tell the ship to hold retrograde while landing (which is useful), whenever you overburn and start rising again (even by a miniscule ammount, which would be shown as 0.0 m/s), the retrograde marker flips towards the ground - and SAS will vigorously try to follow it (which is not useful - but it would explain the landing problem).

What you describe about the EVA sounds like phantom forces - basically (as far as I understand it), sometimes parts (including kerbals) get... Entangled, so to speak. Controlling one exerts the same forces on the other. Was the ship following the same movements you were performing by the kerbal? If so, this might be your answer.

If not, the answer might be Kraken.

Actually, the answer is always Kraken. Hail Kraken.

[celem]

Deutherius has it I suspect, this one gets me when im tired. You set retrograde SAS for suicide burn, but if you hit the hover it inverts you and you die. You try to correct but the craft doesnt steer right, you dont realise that you are actually fighting the Kerbal. Always go to general stability lock once vertical.

You also have to watch out for the fact that EVA will disable the crafts SAS as the pilot is leaving, so you cant park it on a severe slope thanks to torque, then get out and expect it not to fall on you. Which i've also done. Ive started putting probes on everything just for sas when empty. Your Eva does sound a bit glitchy though.

Edited March 21, 2015 by celem

[TheMoonRover]

Deutherius has it I suspect, this one gets me when im tired. You set retrograde SAS for suicide burn, but if you hit the hover it inverts you and you die. You try to correct but the craft doesnt steer right, you dont realise that you are actually fighting the Kerbal. Always go to general stability lock once vertical.

You also have to watch out for the fact that EVA will disable the crafts SAS as the pilot is leaving, so you cant park it on a slope, get out and expect it not to fall on you. Which i've also done. Ive started putting probes on everything just for sas when empty. Your Eva does sound a bit glitchy though.

^This, except use radial out once vertical; it will try to keep the ship the right way up, and you can more easily correct horizontal velocity without overcompensating. Once landed, turn SAS off and allow the ship to settle (assuming you think its safe and that it won't fall over)

EDIT: Also, welcome to the forums.

Edited March 21, 2015 by TheMoonRover

[Bev7787]

Does it have any massless parts?

[celem]

I didnt actually think about it much when it was introduced with 0.90, because i've used SmartSAS previously. But my buddy who plays vanilla was thrown by this for days when he first got access to vector holds.

Good call on radial out, hadnt thought of that and it counters a need for me to juggle a bit at touchdown, which I do with the static lock.

[eddiew]

If it's been a few updates, do you actually have a pilot and/or SAS capable probe core onboard? A lot of people found their old saves were a nightmare because all of a sudden their ships didn't have SAS.

Or as others have said, having the SAS locked to a direction can go mad at the point of touchdown when 'retrograde' can mean anywhere.

[helaeon]

How is the fuel flow? Is it drawing from the top tanks first or last?

If it's drawing from the top tanks last then you're possibly having COM movement issues, especially if you have some radial tanks and they are not being drained evenly.

[impyre]

I was beat to the punch, but I was going to say SAS. I have had this happen... the biggest issue is when having your pilot hold retrograde (when your velocity gets near zero, it moves about wildly). The easy solution is to just switch back to stability assist before velocity gets too low.