help lading a tall craft

[seyss]

Hello,

I have a tall craft that I want to land. I tried Duna and Moon, and on both after touchdown the craft tilts sideways and fall. I have a simpler version (without those horizontal structures) which I can land fine, but I wanna land this one.

Tried RCS as well.. but no luck. Maybe there's some way to put 4 small thrusters on each side and ignite them when the craft starts to tilt? just like RCS but with larger thrust.

Thanks in advance for any tips..

[Grumman]

For starters, you should take all the oxidiser out of that fuel tank if you're not going to use it. That's ten tons of dead weight perched at the top of a tall tower, which is asking for trouble.

[seyss]

I modified the LV-N script to make it use lf+oxi..

[HoloYolo]

Have a wider base. The wider the base, they less likely it's going to tip.

[windows_x_seven]

Make it wider, or add SAS units. You could also try landing slower, in a flatter region.

[Snark]

That's... quite a contraption.

Really, for anything that's tip-prone, there are only two answers (not mutually exclusive): Lower the center of mass, or widen the support base.

From the screenshot, it looks as though it would be pretty easy to widen the base-- you have your landing legs mounted on short girders. Just replace those short girders with long ones and you'll buy yourself a lot of stability.

To lower the CoM, figure out ways to move the heavy parts lower. For example, those LV-N's are really really heavy, and they're right up at the top of the craft-- anything you can do to move them downwards would help.

[Alshain]

Remove one landing leg.Tri-symmetry is more stable than Quad.

[Dalzar]

That looks like a difficult one to land! Nice looking rocket though I would advise widening the base so that there is more mass at the bottom of the ship. Take it slow when you are landing, it is important to stay steady. Also, make sure the place you are landing is flat and there are no rocks!

[Snark]

Remove one landing leg.Tri-symmetry is more stable than Quad.

Not necessarily. I started with 3-legged landers (because "tripods are stable" and "three legs have less mass than four"), and quickly shifted to 4-legged designs because my threes kept tipping over.

There's a reason why cars have four wheels.

Picture your craft's footprint on the ground. I'll define some terminology for purposes of discussion:

CoM footprint: The spot on the ground directly beneath your center of mass.

Support polygon: The shape made by connecting the feet of your lander with straight lines. For a typical three-legged lander this will be an equilateral triangle; for a typical four-legged lander, this will be a square. Ideally your CoM footprint will be directly in the center of the support polygon.

Maximum support radius: The distance from the CoM footprint to each leg (i.e. to the vertices of the support polygon).

Minimum support radius: The distance from the CoM footprint to the sides of the support polygon, i.e. the length of a line from the CoM footprint to the midpoint of one of the sides of the support polygon. This will be shorter than the maximum support radius. The more feet your lander has, the bigger the minimum support radius.

The issue with three-footed landers is that they have the smallest minimum support radius. Four-footed landers have significantly larger ones.

When your craft starts to tip, what makes it fall over is when the CoM moves outside the support polygon. Four-footed landers have a bigger "safety zone" within which they're tip-proof, because they have a bigger minimum support radius than three-footed.

Note that one of the big reasons for tripods doesn't apply much for KSP: A tripod can always have a stable, no-wobble footing on any surface, whether it's uneven or not, whether the tripod legs are exactly the same length or not. A four-legged object (like a table) can wobble, either if the legs aren't exactly alike, or if the surface it's sitting on is uneven. However, in KSP, typically "uneven surface" isn't an issue. Your landing site may be sloped, but it's typically a flat slope, unless you're unlucky and come down right on a ridge line. This is due to the relatively low level of surface detail on KSP planets. This means that one of the biggest reasons for tripods simply doesn't apply; and even if there is a slight ridge where you come down, the legs' suspension tends to make up for that anyway.

Moral of the story, it's not a slam-dunk that going three-footed will be more stable.

[AaronLS]

I landed a really tall craft on minmus, but it took a few hops.

- If it started to tip, lift off and try again. It doesn't use a great deal of fuel to for a couple hops on Mun. Would be worse on a heavier planet.

- If you try to use RCS after touchdown to counter tipping, don't use rotation(WASD), use lateral (IKJL). Rotation will have almost no effect once a leg is on the ground.

- When you touchdown and all legs touch, if you're not pointed perfectly straight up(nav ball) then you know you're not on level surface and would explain why you tip over even after perfect landing.

- I think 4 legs is fine to.

- I would try rotating legs up/out slightly to create a large base area. It looks like you have room underneath. I don't think it really matters if the feet sit flat, but might be wrong.

- Make sure center of mass is centered over the legs. Reduce fuel in VAB to check and see where the CoM might be after fuel has been consumed to the point you expect at landing.

[Foxster]

...or maybe...just a moment before touchdown, de-couple the top and fly it off for a separate landing. I'm assuming the plan is for them to work apart.

[Ripper2900]

Your craft looks like it's two rovers on top of each other. The top one has a huge horizontal wheelbase, and the lower one is mounted at its side, and Las the landing legs at the rear.

I would try to mount them the other way. The one with the larger footprint should be the lowest. That would prevent the tipover issue to a *very* large degree. Putting the smaller mining rover on top world require some means of deploying it, but that shouldn't be too much trouble.

[Alshain]

Not necessarily. I started with 3-legged landers (because "tripods are stable" and "three legs have less mass than four"), and quickly shifted to 4-legged designs because my threes kept tipping over.

There's a reason why cars have four wheels.

Picture your craft's footprint on the ground. I'll define some terminology for purposes of discussion:

CoM footprint: The spot on the ground directly beneath your center of mass.

Support polygon: The shape made by connecting the feet of your lander with straight lines. For a typical three-legged lander this will be an equilateral triangle; for a typical four-legged lander, this will be a square. Ideally your CoM footprint will be directly in the center of the support polygon.

Maximum support radius: The distance from the CoM footprint to each leg (i.e. to the vertices of the support polygon).

Minimum support radius: The distance from the CoM footprint to the sides of the support polygon, i.e. the length of a line from the CoM footprint to the midpoint of one of the sides of the support polygon. This will be shorter than the maximum support radius. The more feet your lander has, the bigger the minimum support radius.

The issue with three-footed landers is that they have the smallest minimum support radius. Four-footed landers have significantly larger ones.

When your craft starts to tip, what makes it fall over is when the CoM moves outside the support polygon. Four-footed landers have a bigger "safety zone" within which they're tip-proof, because they have a bigger minimum support radius than three-footed.

Note that one of the big reasons for tripods doesn't apply much for KSP: A tripod can always have a stable, no-wobble footing on any surface, whether it's uneven or not, whether the tripod legs are exactly the same length or not. A four-legged object (like a table) can wobble, either if the legs aren't exactly alike, or if the surface it's sitting on is uneven. However, in KSP, typically "uneven surface" isn't an issue. Your landing site may be sloped, but it's typically a flat slope, unless you're unlucky and come down right on a ridge line. This is due to the relatively low level of surface detail on KSP planets. This means that one of the biggest reasons for tripods simply doesn't apply; and even if there is a slight ridge where you come down, the legs' suspension tends to make up for that anyway.

Moral of the story, it's not a slam-dunk that going three-footed will be more stable.

Cars move forward along the ground, that isn't a fair comparison. They had 3 wheel ATV's once but the rear wheel drive caused the front wheel to jump up and the whole thing rolled over. You simply can't compare a ground vehicle to a vertical lander. Car wheels aren't 4x symmetry either btw.

With 3 landing legs, you have less opportunity to flip because whichever way the craft starts to tip, there will be a leg there. While a 4 leg lander can tip easily on two legs, a 3 leg lander will be less likely. 3 legs are still more stable than 4 for this reason. It still takes piloting skill, but if you land correctly oriented the craft won't be able to tip over unless it is just REALLY off balance in which case nothing is going work.

Edited July 13, 2015 by Alshain

[AaronLS]

With 3 landing legs, you have less opportunity to flip because whichever way the craft starts to tip, there will be a leg there.

Take 2 legs and draw a line between them. The mid point on that line is the closest tipping point to the center. It is exactly the same place that a 3 wheeler tips over.

For a four leg lander, those tipping midpoints are farther away from the center than a 3 legged lander, giving it less chance to tip.

The benefit of a 3 legged lander is not that it is less likely to tip over, it's that no matter how uneven a surface is, all 3 legs will make contact. A four legged lander will wobble on an uneven surface. IMO, for a tall lander you need to find an even surface regardless.

- - - Updated - - -

They actually considered 3 legs for the lunar lander but eventually switched to 4 for to improve stability.

[Sharpy]

IMO the ISRU rover should be mounted sideways. Just attach a docking port or a radial decoupler over its CoM and hang it under the top one.

Do you really need to plant it vertically? Wouldn't you prefer it to land in its target orientation?

It should fit under the lander-rover, and if it doesn't, you can make it shorter - attach two Ore tanks radially to the fuel tank, move the battery to the front and attach the gigantors to the battery. Maybe move the drill(s) to the fuel tank part too.

Or simply give it the BIG reaction wheel. With that thing you should be able to hold it in a vertical position even on uneven terrain.

Edited July 13, 2015 by Sharpy

[Jouni]

There are three different scenarios for landing strut configuration.

1. On rugged terrain, three legs can be more stable than four. This scenario is mostly irrelevant, because almost all landing sites in KSP are smooth.
   
2. On level ground, having more legs is always better. There is little advantage beyond 5 or 6, however.
   
3. On steep slopes, the important thing is the distance between the uphill legs and the downhill legs. If 12 is uphill and 6 is downhill, three legs in (2, 6, 10) configuration have a slight advantage over four legs in (1.5, 4.5, 7.5, 10.5) configuration (the distance is 1.5 vs. sqrt(2)). With four legs in (3, 6, 9, 12) configuration, the distance is even greater, but the middle legs may push the lander sideways, if it's not aligned exactly right with the slope. Four legs in (1, 5, 7, 11) configuration is probably the best solution, because the distance is sqrt(3) and you have two legs on the downhill side.
   

[rofltehcat]

The whole thing is designed to operate horizontally once there, right? Land it horizontally.

Move the landing legs to the lower side so you land on the landing legs (wheels break easily), and can retract them to lower yourself onto the ground. Get RCS Helper and let it display your full and empty center of mass. Move them as close together as possible (e.g. reposition fuel tanks). Then attach small engines (radial engines or small strut+terrier/spark, need enough TWR) in positions that minimize the torque. Also add a small docking port (or probe core) on top that you can use to 'control from here' (not necessary but it helps). For Duna you could even adapt parachute placement so you're guaranteed to land horizontally.

I landed this 71 t monstrosity on Minmus horizontally with just 5 Spark engines (~2.5 TWR). For the Mun you'd probably need more like 5 Terriers but that is also doable. A couple of Thuds or comparable could do the job as well. You could even attach them with decouplers and fuel lines if you want to get rid of them later on.

The important thing is balancing the thing right. Fuel tanks in front and back help because you can pump around fuel to balance if needed.

Edited July 13, 2015 by rofltehcat

[ajburges]

There are three different scenarios for landing strut configuration.

1. On rugged terrain, three legs can be more stable than four. This scenario is mostly irrelevant, because almost all landing sites in KSP are smooth.
   
2. On level ground, having more legs is always better. There is little advantage beyond 5 or 6, however.
   
3. On steep slopes, the important thing is the distance between the uphill legs and the downhill legs. If 12 is uphill and 6 is downhill, three legs in (2, 6, 10) configuration have a slight advantage over four legs in (1.5, 4.5, 7.5, 10.5) configuration (the distance is 1.5 vs. sqrt(2)). With four legs in (3, 6, 9, 12) configuration, the distance is even greater, but the middle legs may push the lander sideways, if it's not aligned exactly right with the slope. Four legs in (1, 5, 7, 11) configuration is probably the best solution, because the distance is sqrt(3) and you have two legs on the downhill side.
   

The math tells us that the apothem (shortest distance between the footprint edge and center) for a landing base that is a regular polygon with n sides/vertices and a circumradius (distance between landing foot and center) r is r*cos(À/n)

A triangle has an apothem of ½ circumradius. A square has an apothem of .71 circumradius. The square even offers a better apothem per vertex ratio (corrolating to mass efficiency) than a triangle without accounting for the extra length contributed by the craft body.

Where triangles can help is the high difference between apothem and circumradius. If you contact with the apothem first, you get less torque than you would with the circumradius. When the opposite side then makes contact, it provides a greater relative counter-torque to cancel the rotation if it is the circumradius. The torque difference is among other things proportionate to the sine of the angle at the foot between base and CoM. Tall craft wouldn't see much advantage here.

[Snark]

Where triangles can help is the high difference between apothem and circumradius. If you contact with the apothem first, you get less torque than you would with the circumradius. When the opposite side then makes contact, it provides a greater relative counter-torque to cancel the rotation if it is the circumradius.

First good argument for 3-footed landing I've seen in the thread. For the non-mathematically inclined, what ajburges just pointed out is that 3-foot can help when you land, if you're careful to set your roll correctly so that your first contact is two feet. That is, for a craft with any number of feet:

- the best-case scenario for first contact is that you touch with two feet perfectly aligned on your trailing edge, then settle down so that the "forward" feet on the opposite side touch down.

- the worst-case scenario is the opposite, where you touch down with one foot perfectly aligned on the trailing edge, then settle down.

ajburges' excellent point is that a three-footed lander maximizes the benefit of that best-case scenario.

That's a good point, though I find that in practice, I have better luck with 4 feet. Reason:

- taking advantage of this effect requires you to have the roll axis perfectly aligned, otherwise you lose the benefit. I find that juggling pitch, yaw, and velocity keep me mentally occupied enough that I prefer not to have to worry about roll, too.

- when I set up my suicide burn, I generally end up coming just about straight down to the surface, so that this benefit is minimized.

So the real answer to this thread is: 3 feet can work well for you, and so can 4. There are good arguments on both sides and it's not a slam-dunk either way, since it depends on variables such as your skill level, your landing style, the slope you're landing on, and the design of your craft.

[ajburges]

First good argument for 3-footed landing I've seen in the thread. For the non-mathematically inclined, what ajburges just pointed out is that 3-foot can help when you land, if you're careful to set your roll correctly so that your first contact is two feet. That is, for a craft with any number of feet:

- the best-case scenario for first contact is that you touch with two feet perfectly aligned on your trailing edge, then settle down so that the "forward" feet on the opposite side touch down.

- the worst-case scenario is the opposite, where you touch down with one foot perfectly aligned on the trailing edge, then settle down.

ajburges' excellent point is that a three-footed lander maximizes the benefit of that best-case scenario.

That's a good point, though I find that in practice, I have better luck with 4 feet. Reason:

- taking advantage of this effect requires you to have the roll axis perfectly aligned, otherwise you lose the benefit. I find that juggling pitch, yaw, and velocity keep me mentally occupied enough that I prefer not to have to worry about roll, too.

- when I set up my suicide burn, I generally end up coming just about straight down to the surface, so that this benefit is minimized.

So the real answer to this thread is: 3 feet can work well for you, and so can 4. There are good arguments on both sides and it's not a slam-dunk either way, since it depends on variables such as your skill level, your landing style, the slope you're landing on, and the design of your craft.

In my defense, I hardly consider any landing with enough horizontal momentum to generate torque on touchdown a best case scenario. There is not much margin between where 3 legs help and completely out of control. OTOH landing roll trends to be set and forget.

In the interests of transparency, I also prefer 4+ leg configurations. Though I do only use 3 for my ion landers to optimise mass.

[Snark]

In my defense, I hardly consider any landing with enough horizontal momentum to generate torque on touchdown a best case scenario.

Well, yeah, no argument there! But if the landing's picture-perfect so there's no touchdown torque, then it doesn't really matter how many landing legs there are, anyway, unless the craft is super top-heavy. It's a handy discussion for "how to design for when things don't always go as planned."

Also, touchdown torque doesn't have to be from horizontal momentum-- you'll get it when you land on a slope, even if heading straight down. (And landing on a slope is exactly when tipping tends to be the most dangerous.) So it's still a worthwhile topic of discussion.

There is not much margin between where 3 legs help and completely out of control. OTOH landing roll trends to be set and forget.

Yes, it's set-and-forget. Here's my problem, though:

The only time that I'm in much danger of tipping (and therefore might care about roll orientation) is if I'm landing on a slope. And personally, I find it annoyingly difficult to eyeball exactly which way the slope is going as I descend-- I can get the general gist, but until I'm really close to the surface, my visual estimate is often off by several compass points. (I can mitigate this somewhat by putting a couple of spotlights on the lander, angled down by about 45 degrees. Great for night landings, but they also help even in daytime by giving a couple of visual reference points. But it's still tricky for me.)

Maybe others are better at eyeballing slopes than I am, but for me, that makes it so that if I want to adjust roll for the optimum case, I can only do that in the last few meters of descent, when I tend to have my mental hands full. So I design for that by giving my landers 4 legs, and making them very squat, so that my roll orientation doesn't matter so much.

[Wemb]

(I can mitigate this somewhat by putting a couple of spotlights on the lander, angled down by about 45 degrees. Great for night landings, but they also help even in daytime by giving a couple of visual reference points. But it's still tricky for me.).

This,, which I shamelessly stole from 633 squadron after watching The Dambusters works a treat if you don't want to fly with mechjeb or are landing (eek) in darkness. Angling them so they converge at a point directly below your main engines can really help with judging the vertical height above the ground.

Wemb

[Snark]

Angling them so they converge at a point directly below your main engines can really help with judging the vertical height above the ground.

I actually prefer to have 'em diverge, i.e. one points off to the left and one off to the right. It creates two widely-spaced spots of light on the ground, so that if there's significant slope, one of the lights will be visibly farther away than the other and I can see the slope better. I can judge height-above-ground pretty well, too.

[Wemb]

I actually prefer to have 'em diverge, i.e. one points off to the left and one off to the right. It creates two widely-spaced spots of light on the ground, so that if there's significant slope, one of the lights will be visibly farther away than the other and I can see the slope better. I can judge height-above-ground pretty well, too.

Nice tip - not a problem 633 squadron had - they knew they were flying over level, err, ground.

OTOH, they did have fly to stay at 18m while doing 108 mps sideways. In the dark. Being shot at. And their 'reaction wheel' was spinning at 500 rpm and contained 3,000kg of high explosive.

Would put Jeb to shame.

Wemb