Sibling Rivalries, Barstools, and Kebral Space Program

[cicatrix]

Landing legs are not rigid. That's the difference from a stool. Each leg can be of different length. To make them more or less rigid there's an option 'Lock suspension' on them. The idea is that no matter how many legs you have they will all hit the ground... eventually. Sometimes, I think that just 'lock/unlock' might not be enough. The power of their 'springs' or hydraulics should also be adjustable. So, if you're landing on a slope the stiffness of your landing legs should be very weak (and you should have strong reaction wheels )

Therefore, even a tripod can be tipped over and it's hard to tip over a heavy low-built four-legged armchair.

[zarakon]

One other thing I've found with 4vs3 is that with 4, that leg pointing uphill will hit the slope first much earlier than a 3 legged design, knocking you off course, or requiring MECO sooner than a 3 legged design. The extra speed tends to cause frantic panic as I try to stop tipping over. Here's a thought - how about putting on 4 legs, but retracting the upslope leg during descent? Should give the best of all worlds (Except for the extra leg mass). Find some 20* slopes to test please!

I actually did something like that recently in order to get surface data from the slopes on Minmus. I had a 4-legged lander, put two legs downhill and retracted the other two. It actually worked really well so that the lander was nearly straight upright on a steep slope. I wouldn't trust something like that on a world with stronger gravity though.

[Dilli]

One other thing I've found with 4vs3 is that with 4, that leg pointing uphill will hit the slope first much earlier than a 3 legged design, knocking you off course, or requiring MECO sooner than a 3 legged design. The extra speed tends to cause frantic panic as I try to stop tipping over. Here's a thought - how about putting on 4 legs, but retracting the upslope leg during descent? Should give the best of all worlds (Except for the extra leg mass). Find some 20* slopes to test please!

Retracting the upslope leg does help, more so on a three legged craft v four legged ones though. (I dont spend much time looking for good landing sites so often end up on slopes)

[ArmchairGravy]

Two legs bad, four legs good, three legs best! Orwell was mostly right.

[FleshJeb]

As a surveyor, I'm well aware of the power of the tripod. However, there's a thread from a few months back where someone proved that 5 legs offers those most tip resistance (distance from center of mass to bounding polygon, as Zarakon said) for the least number of legs.

[Remkeau]

Why use 3 landing legs when you can have 18?

The super over-engineered way too heavy for the job, but (in my opinion) awesomely stable Moho lander You need to slam it pretty hard to get it to tip

[Pecan]

"Three landing-legs are used because 3-legged stools don't wobble! More legs are likely to just add unwanted mass without making the landing more stable so aren't worth it" (http://forum.kerbalspaceprogram.com/threads/79658-Exploring-The-System-A-design-tutorial-campaign-0-90-Final?p=1167684&viewfull=1#post1167684)

Later, in http://forum.kerbalspaceprogram.com/threads/79658-Exploring-The-System-A-design-tutorial-campaign-0-90-Final?p=1183116&viewfull=1#post1183116 :

Fat Sally Lander Stage

Note that the already-quoted "three points define a plane" explains all of the 3-leg stability, in that it is impossible for a design with exactly 3 legs/points of contact to wobble backwards and forwards. 1 or 2 is too few and will tend to fall over, even on a horizontal surface. More than 3 requires the relative lengths of each of the legs to match the contours of the surface exactly or it may wobble. None of this says anything about whether or not a design with 3+ legs will fall over - that's a matter of position of the centre of mass - but the 3 legged design is gauranteed not to wobble while it collapses ^^.

Edited May 18, 2015 by Pecan

[StrandedonEarth]

They stopped making 3-wheel ATV's because they were too unstable, and office chairs have 5 wheels (redundant, yes, but also more stable). Coincidence, or QED?

[zarakon]

They stopped making 3-wheel ATV's because they were too unstable

Like 3-wheeled cars?

[Starwhip]

Small landers, three legs are the better option. Large landers, upwards of four. Mine sometimes have eight.

Also, "Kebral Space Program" in the title.

[Alshain]

They stopped making 3-wheel ATV's because they were too unstable, and office chairs have 5 wheels (redundant, yes, but also more stable). Coincidence, or QED?

That is a different principle. They stopped making 3 wheel ATV's because the drive was in the rear two wheels and the center of mass was on top of those two wheels, this had a tendency to LIFT the front wheel, rolling it backward. It was not because the vehicle wouldn't sit on the ground in a stable manner, it was because it was off balanced. Obviously landing legs do not have acceleration like ATV wheels.

That principle is actually closer to what a plane does when taking off. It tilts back on a fulcrum, just like the ATV, and lifts the front wheel off the ground. Unlike the ATV, the plane keeps going up.

Edited May 18, 2015 by Alshain

[Temeter]

Like 3-wheeled cars?

Nah, 3 wheels are different. A car moves forward, not downward. Landing gear limits horizontal movement, that's the whole point of triangle thing.

[Red Iron Crown]

On the face of it, it would seem that four legs should be more stable as the CoM is further from the line between two adjacent contact points and so the slope at which the CoM is vertically above that line is steeper. However, there is suspension to consider: A four-legged lander's legs will compress only 75% as much as a three-legged lander's, which means the CoM is higher in an equivalent four-leg design once landed. I don't think it is as cut-and-dried as one approach being better than the other in all cases, a lot depends on the vessel layout and mass.

[DarkGravity]

Compared, the real lunar lander did land on a pre-determined, optimal landing site, not the accidental side of a hill like you'd do in KPS, so the counter-wobble of 3 legs wasn't needed and they went for the advantages of having 4 legs.

The real lunar lander, at least the first one, missed its landing site completely, and the crew had to eyeball a new landing site on the fly, much like we do in KSP, only they had less computer power than an IPhone 1.0 running everything. They did an awful lot of flying by the seat of the pants.

[Temeter]

The real lunar lander, at least the first one, missed its landing site completely, and the crew had to eyeball a new landing site on the fly, much like we do in KSP, only they had less computer power than an IPhone 1.0 running everything. They did an awful lot of flying by the seat of the pants.

Right, forgot how the nav comp bugged out. It still was at least an eyeballed, even site. Lets be real, noone in their right mind would land at the places we do in KSP.

[Renegrade]

On the face of it, it would seem that four legs should be more stable as the CoM is further from the line between two adjacent contact points and so the slope at which the CoM is vertically above that line is steeper. However, there is suspension to consider: A four-legged lander's legs will compress only 75% as much as a three-legged lander's, which means the CoM is higher in an equivalent four-leg design once landed. I don't think it is as cut-and-dried as one approach being better than the other in all cases, a lot depends on the vessel layout and mass.

Very much depends on the mass and surface gravity. My Mun landers experience almost no compression in stock (I go with the 'less is more' design approach, not the 'moar boosters' approach), and my Minmus ones have no visible compression unless I bounce 'em hard off the surface...

The real lunar lander, at least the first one, missed its landing site completely, and the crew had to eyeball a new landing site on the fly, much like we do in KSP, only they had less computer power than an IPhone 1.0 running everything. They did an awful lot of flying by the seat of the pants.

*sigh* Kids these days.. The Apollo Guidance Computer (4 kilobytes of RAM, 16-bit, 2mhz processor) is on the scale of a Vic-20 (although an 8-bit, 1mhz design, the 6502 can push through an instruction in 2-3 cycles, vs 12+ for the AGC) - the original iPhone has a 400mhz, 32-bit ARM CPU with 131,072 kilobytes of RAM...which is much faster than the Cray X-MP supercomputer (100mhz, 16,384kb design), released many years after the end of the Apollo program~

[jarmund]

I think it comes down to a combination of wibbliness vs intuition. Humans, me included, possibly including most other humans, have an easier time thinking about 4-point orientation, because thinking in 90 degrees is simple. This is why I always use 4 legs: It's easier for me to predict where it'll tip in the event of pilot error. On the other side, wobbly chairs always has that one leg that doesn't reach all the way to the floor.

[Tracer13]

*sigh* Kids these days.. The Apollo Guidance Computer (4 kilobytes of RAM, 16-bit, 2mhz processor) is on the scale of a Vic-20 (although an 8-bit, 1mhz design, the 6502 can push through an instruction in 2-3 cycles, vs 12+ for the AGC) - the original iPhone has a 400mhz, 32-bit ARM CPU with 131,072 kilobytes of RAM...which is much faster than the Cray X-MP supercomputer (100mhz, 16,384kb design), released many years after the end of the Apollo program~

Sorry to take this off track, but I just had to comment.... I got to meet the man who created the VIC-20 and ended up working with him for a short bit. He designed it and wrote its first kernel (if you can call it that). In the Commodore manual, it said that VIC stands for Video Interface Chip. It was actually named after Vic, the guy that created it. He was super nice and said he made it to help get young people into computing. I said "It worked! Here I am!"

On another side note, I was at the Bay Area Maker Faire this weekend and got to meet Eben Upton and Liz from the Raspberry Pi foundation. Great people!

Tracer

Edited May 18, 2015 by Tracer13

[DancesWithSquirrels]

"Like 3-wheeled cars?" - Gah. That particular "Top Gear" episode was utter rot, as they loaded the side of the cars with weights (lead?) to make them especially tippy. In short, they cheated.

Now, as to contact: 3 legs will always touch on a plane, whatever the orientation, while 4 won't touch equally on uneven ground unless rotated correctly (there is a all-4 rotation angle for planar contact, although it's not going to be clear what that is prior to examination of the ground). Given the shock travel on the standard (unlocked) gear, however, it's compliant enough that all 4 legs will generally contact with some degree of force.

There are different stability limits, however. Given a particular lander and the height of its CG, if the ground is tilted so that the vertical drop from the CG falls outside the leg footprint the lander will tip over. The smallest circle touching the boundaries described by the legs is stable (think a circle inside and touching the edges of the triangle or square), outside that and some angles will tip. Given the same radial outward placement from the CG, 3-leg designs will have the same max stability as a 4-leg when each have a leg pointing straight downhill, but less stability if the downhill slope runs directly between two legs, a lower minimum limit.

If you want the same minimum stability angle in a 3-leg as in a 4-leg design, you need to place them further out from the CG, wider, so that the minimum stability circle matches that of the 4-leg design.

Wider footprints, lower CGs -> more stable landers.

[Raptor9]

EDIT: sniped by DanceswithSquirrels...but yeah

Alright, can I simplify this argument a little? The number of landing legs you have is a secondary factor compared to where your center of mass is in relation to your landing leg base diameter. If you make a tall 3-legged design and land it on a slope it will be less stable than a short 4-legged lander with a wide leg base. And if you land a tall 4-legged lander, it will be less stable than a short 3-legged lander. If the lander center of mass is above or beyond the downslope leg, it's going to tip over, I don't care how many legs it has. It's the same concept as "dynamic rollover", which is a term used to discuss slope-landings in helicopters. Granted, helicopters have more forces acting on them than a direct-thrust lander, but the CoM over the pivot point priniciple still applies:

If you reference the above picture, imagine the center of mass (CoM) was toward the top of the helicopter. The higher up the CoM is, the closer it is to being directly above the downslope point of contact. If you make a lander with a CoM low when compared to the relative width of the landing base, the argument of how many legs you have is of no consequence. Although from an economical outlook, one less landing leg is slightly lighter and cheaper, but as a design principle it's not nearly as significant as where the CoM and landing base compare with each other.

[martinborgen]

in theory, 3 legs is more stable, but you will need them spaced further apart to enable the same ammount of tilting craft without falling over. Four legs would be susceptible to the wobbly table syndrome, but since legs have shock absorbers, it's not much an issue - same reason your car wont' wobble. (of course, many moons have to weak gravity to pull landers low on their shocks, but then your SAS will keep you stable)

[Pecan]

... The number of landing legs you have is a secondary factor ... the CoM over the pivot point priniciple still applies ... it's not nearly as significant as where the CoM and landing base compare with each other.

As said before there are 2 separate issues:

1. Will it fall over? Depends on the CoM staying inside the area described by the legs. Less than 3 legs won't work for long. More legs are, very marginally, better.

2. Will it wobble? Yes, unless it has exactly 3 legs.

The point of 3 legs is that it prevents them wobbling, not makes them less likely to fall over. This isn't really that much of an issue for landers but can be very important for bases, where you have mass shifting around and different components being docked and undocked. Build bases with 3 legs per module if you want to have any hope of keeping them aligned!

[magnemoe]

in theory, 3 legs is more stable, but you will need them spaced further apart to enable the same ammount of tilting craft without falling over. Four legs would be susceptible to the wobbly table syndrome, but since legs have shock absorbers, it's not much an issue - same reason your car wont' wobble. (of course, many moons have to weak gravity to pull landers low on their shocks, but then your SAS will keep you stable)

This, draw an line between the feet, triangle, square, pentagram or more. If you land in uneven terrain and you center of gravity is outside this you tip over.

Four is most cost effective here if your legs has to fit inside an fairing, 5 or more is for redundancy or if your lander is too heavy for 4 legs.

For an light lander who can use complex mufti joined legs or have low center of gravity you use 3 legs.

[BurningLegion]

Thanks for discussing this! I've learned a lot. My current re-usable Munar Lander design (Fourth time lucky *crosses fingers*) has three engine pods on modular girders coming out from a central lander. Each pod has two legs positioned so that if you look at it from as a whole it has two sets of triangular landing legs. I'm hoping to get all the benefits of a triangle with the redundancy of a different type of lander.

It's also the first one I've made I remembered to put photon beams on, which just *may* impact performance. It's attached to the Munar Exploration Station (MES), the big tripod thing.

I just hope it lands well enough xP