Lander flips despite being perfectly balanced

[Tokamak]

I have a large rover that I am trying to land horizontally. I'm an experienced player and have done this many times before. But this time I cannot get it to work properly.

I'm not using TCA because I cannot for the life of me get it to play nice with MechJeb, and have trouble getting it to work at all despite having read the manual. And when I *can* get it to work, I can't get it to land at an assigned target. So the solution without TCA is to just balance thrust perfectly, right?

The center of gravity and center of thrust are both the same, with the tanks full, partly full, and empty. But no matter what I do, it starts out fine, and then as the tanks empty it starts nosing up and eventually flips. I've checked that they are both draining equally. (I wish there was a way to view center of gravity/thrust in flight mode!)

The weirdest thing is that it only starts flipping once the tanks have emptied a bit, not when it's full. As the tank empties I can see the RCS working harder and harder to stabilize the thing until it finally flips.

Any idea what could possibly be going wrong? I've spent a lot of time on this thing. And I've done this before and always gotten it to work. I'm at a loss.

[king of nowhere]

it's not about the center of mass, i have a similar rover that is not finely balanced, and it still manages to fly using gimbaling to compensate. in fact, you should not need rcs. maybe that's part of the problem? is gimbaling active? if gimbaling is active, you should balance no problem.

wait, maybe too much gimbaling. the vector is the engine with the greatest gimbaling, and it often overcompensates, pushing so hard that it unbalances the ship more. so tuning down gimbaling to 20% or so may work.

too much thrust? seems like that rover has a huge twr - unless it's a lot heavier than it looks - and I see sometimes ships have more instability problems at higher thrust. where is that rover supposed to land, tylo?

another idea, maybe the rover is flexible? the rockets push at the edges, the rover bends, and the rockets are suddenly off-center. which is related to fuel draining and thrust, the fuel is on the edges of the rover, draining the fuel increases the acceleration of the edges compared to the inertia of the center. seems the most likely issue. and in that case, moving those rockets closer to the center of mass would help.

 

P.S. I can tell you are experienced because the first half dozen solutions I thought, i read a line later "and i checked this thing".

[Bej Kerman]

2 hours ago, king of nowhere said:

wait, maybe too much gimbaling. the vector is the engine with the greatest gimbaling, and it often overcompensates, pushing so hard that it unbalances the ship more. so tuning down gimbaling to 20% or so may work.

If you take a look at the way the engines are arranged, the gimbal is probably hindered by the engines being at the same height as the CoM. If they were shifted up a bit more, the gimballing of the engines would simply cause lateral motion instead of causing torque.

I think the gimbal should be outright disabled, but not for the same reason you think. The gimbal as I said probably just causes the vessel to drift to the sides instead of rotating it, and the vessel already has RCS anyway.

6 hours ago, Tokamak said:

The weirdest thing is that it only starts flipping once the tanks have emptied a bit, not when it's full. As the tank empties I can see the RCS working harder and harder to stabilize the thing until it finally flips.

A video of this happening would be appreciated. Really, just saying it's "flipping" is very vague and it can mean a lot of things. Does the flip happen on contact with the ground - are you moving laterally when you land? Could the means of applying torque on your vessel be fighting against SAS inputs - sometimes, like with backwards wings for example, a part can end up applying torque in reverse. It's probably obvious what it means to you, but since we weren't there when the accidents happened there's not a lot to be gathered from your description.

[Tokamak]

I already experimented with raising and lowering the engines, and having various levels of gimbaling.

What happened when I say "noses up and flips" is just that while the engines are on, the front end gradually rises as unequal forces overwhelm the RCS, until it starts spinning end over end. While up in the air, while doing a burn with a simple mechjeb SAS setting like "up" or "prograde".

 

I think it may have been something to do with drag, though. I installed hyperedit and teleported it to space...where it seems to work just fine. So it's gotta be a drag thing. But fortunately, thanks to hyperedit (I restore afterwards so as not to cheat) I found that it works in space and on Duna, so that's what matters.

 

Thank you for the suggestions and ideas!

[DennisB]

A drag issue was my first idea, but I assumed, as an experienced player you considered it already.

I would add at least more reaction wheels, it's more effective in this scenario, than RCS.

If that's not enough, you can try drogue chutes at the back to get the nose down, and cut them, as the nose starts to pitch down too much, and/or you can add some aerodynamic surfaces for pitch control.

[Spricigo]

Since you mention possible drag issue I get that you want to land it somewhere with a atmosphere.(not Duna)  We can also see that the landing apparatus is disposable, so you are not planing to take of later.

It seems to me the obvious solution is to strap some parachutes in the top of the rover, keep some much smaller engines to  steer it to the correct landing spot and decelerate the craft enough so you can open the parachutes. What I'm missing there? 

[Bej Kerman]

15 hours ago, Tokamak said:

What happened when I say "noses up and flips" is just that while the engines are on, the front end gradually rises as unequal forces overwhelm the RCS, until it starts spinning end over end. While up in the air, while doing a burn with a simple mechjeb SAS setting like "up" or "prograde".

Are you hovering, or are you accelerating? If the latter, then it probably is drag.

[Tokamak]

26 minutes ago, Bej Kerman said:

Are you hovering, or are you accelerating? If the latter, then it probably is drag.

Accelerating. So yeah, I think that's what it is. Duna's atmosphere doesn't seem to bother it enough to be a problem, though, so I guess the issue is addressed.

13 hours ago, DennisB said:

A drag issue was my first idea, but I assumed, as an experienced player you considered it already.

I would add at least more reaction wheels, it's more effective in this scenario, than RCS.

If that's not enough, you can try drogue chutes at the back to get the nose down, and cut them, as the nose starts to pitch down too much, and/or you can add some aerodynamic surfaces for pitch control.

You'd think I would have thought of that, but... XD

34 minutes ago, Spricigo said:

Since you mention possible drag issue I get that you want to land it somewhere with a atmosphere.(not Duna)  We can also see that the landing apparatus is disposable, so you are not planing to take of later.

It seems to me the obvious solution is to strap some parachutes in the top of the rover, keep some much smaller engines to  steer it to the correct landing spot and decelerate the craft enough so you can open the parachutes. What I'm missing there? 

The reason to not use parachutes is that I need to land it fairly precisely. Even though it's mobile, I don't want to have to drive it many kilometers over who knows what terrain (It's hard to find flat terrain on Duna) I've never managed to get much precision with parachutes, especially on Duna, since they have so little drag. Getting everything times just right is nearly impossible for me, in my experience. >.<