Skycrane is uncontrollable.

[Motokid600]

Ive made a standard rover/skycrane setup. The thing likes to pitch over at about 50 meters after takeoff during testing. That's with ASAS/SAS on. Without it gets up to about 500 meters then slowly starts to list. But if I try to correct it the controls are all backwards and it immediately rolls over. CoM/CoT are perfect. I control it from the cranes pod and it still rolls. Now with the rover detached the crane works fine. So idk if the rover torque is somehow interfering? Because even though im controlling another pod I can still move the rover wheels. any tips?

Edited July 23, 2013 by Motokid600

[stupid_chris]

Your CoT is over your CoM, so all the gimbals on the engines are having reverse action and the game does not take account for those situations. Every command you will give your ship will have reverse action, so ASAS makes it even worse. Just lock the gimbal on all engines and you won't have any problem anymore.

[Motokid600]

Ah I see thanks so much.

[Motokid600]

Hm. Now it can get up to about 2000m, but it still rolls over. Now obviously this thing will not be ascending, but even upon decent the craft is very hard to control and still wants to roll over. Heres a CoM/CoT shot. I tried moving the engines down.

[stupid_chris]

PRobably because you don't have much control authority. Probe torque can't always do miracles. Gimbals won't work on this design for sure. As long as your gimbals are all deactivated they won't cause you any problem. Try using RCS to control maybe. Also isn't four Mk 55 engines a bit overkill for that rover? I feel like you could do with only two, that should give you more delta V and more control authority with torque.

[Motokid600]

Well I thought four would give me more control then two, but now that gimbals are off I guess that would make sense. Thanks again ill give it a try

[lammatt]

you are thrusting at yourself...your rockets thrust are pushing on your wheels; try to have a wider span so that the rockets point outside of the footprint of your rover

you cant gain acceleration with an internal force...

Edited July 22, 2013 by lammatt

[jaredkzr]

I don't think KSP simulates the effects of the rocket gasses hitting other parts other than heat. I've never had my rocket's or lander's engine force canceled out by having it aimed at other parts of the ship.

[Mr. Scruffy]

I don't think KSP simulates the effects of the rocket gasses hitting other parts other than heat. I've never had my rocket's or lander's engine force canceled out by having it aimed at other parts of the ship.

Just tested - not true. If you thrust upon your own rocket, the thrust does get canceled out. I built a very simple rocket to test: Four radial engines with four plates put under them. The design had plenty of TWR, but would not lift off.

EDIT: For completeness sake: With a plate under only one of the four engines, the rockets lifts off and immedietaly starts to spin, just as one would expect IRL.

Edited July 22, 2013 by Mr. Scruffy

[stupid_chris]

Guys, it's just the gimbals, for helvete it's even on the bugtracker. If he trusted at himself, he wouldn't move.

[jaredkzr]

Just tested - not true. If you thrust upon your own rocket, the thrust does get canceled out. I built a very simple rocket to test: Four radial engines with four plates put under them. The design had plenty of TWR, but would not lift off.

EDIT: For completeness sake: With a plate under only one of the four engines, the rockets lifts off and immedietaly starts to spin, just as one would expect IRL.

Well you learn something new everyday! I've never gotten around to actually test that. Thanks for the info!

[DerekL1963]

Hm. Now it can get up to about 2000m, but it still rolls over. Now obviously this thing will not be ascending, but even upon decent the craft is very hard to control and still wants to roll over. Heres a CoM/CoT shot. I tried moving the engines down.

Move that parachute and the RGU to the top of a modular girder segment - give it some lever arm!

[Nao]

Well I thought four would give me more control then two, but now that gimbals are off I guess that would make sense. Thanks again ill give it a try

Remember that gimbaled engine torque is proportional to distance of the CoM to CoT (assuming vertical thrust only). Your ship has almost no control authority from gimbals. Since we cant change thrust of each engine individually (as it is done IRL for such cases) the only other way to get control power without changing the design and dimensions of the craft, would be to angle engines inwards. It won't push the CoT shown in VAB down but the intersection of lines perpendicular to thrust vector of each engine (that the gimbal moments depend on) will go down (possibly below whole craft) increasing gimbal torques. It can also allow having CoM below CoT and without reversed controls).

Edited July 23, 2013 by Nao

[NoMrBond]

If you get your engines a bit further out (those M55 radials seem like total overkill for the load your skycrane is dropping) and then use shift-WASDQE to angle them out 1 or 2 clicks (5 or 10 degrees) it was really stable (on my mockup anyway, couldn't guess exact construction though).

It ran out of fuel pretty quick on Kerbin, but I'm not sure where you're trying to land this (surface-to-surface, from space, to atmosphere planetoid etc)

Seemed fine though. Only thing I can think of is remember to turn on the brakes as the spinning wheels can do weird torque things

Edited July 23, 2013 by NoMrBond

[Skorpychan]

So, engine thrust hitting things and spinning wheels cause torque?

NO WONDER my mun base-rover kept screwing up!

[mellojoe]

I've often discovered this same problem. If you have a part in the middle of an engine's thrust, it does weird things. I haven't noticed the spinning wheel problem, but you had better believe I am hitting brakes on future crafts.

[Motokid600]

I took some advice, removed two engines and turned off gimbals. Worked like a charm. Thanks fellas : ) and turning on the brakes...really good idea.