Curved Runways

[Rdivine]

Whenever you put a ship on a runway, chances are that i'll start rolling forwards as if there's a phantom effect acting on it.

The fact is, the runway is actually slightly curved downwards, so the vessel will roll forwards.

This diagram shows why the vessel has a tendancy to move forwards.

The runway (black) isn't curved to compensate for the curvature of Kerbin (green).

Any vessel would experience a force slightly sideways relative to the surface of the runway ( red arrows ),

because the runway is inclined to the true tangent of the surface at that point ( dotted red lines ).

Making the runway slightly curved would go a long way to solving problems of vessel having trouble landing, vessel rolling forwards and crashing etc.

[cubinator]

So that's why they always do that! Huh. I like the idea, but for me it would just be more of a 'quality of life' thing. Not exactly vital to the game, but it should definitely be considered.

[Tex]

6 hours ago, cubinator said:

So that's why they always do that! Huh. I like the idea, but for me it would just be more of a 'quality of life' thing. Not exactly vital to the game, but it should definitely be considered.

I support this view. It takes .4 seconds to hit the brake button if I want to keep my planes still, otherwise I just go full throttle pretty much directly after physics loads.

[Anquietas314]

It's a nice idea in theory, but what do you do with mods that change the size of the home planet (or the entire planet for that matter)? For example: RSS, 64K, Toy Solar System, Duna Space Program to name a few. We'd need either the runway's curvature to be configurable or a way for mods to just replace the runway mesh outright, otherwise those mods would still have these issues (or worse - an excessively curved runway is probably going to be more of a pain than an inadequately curved one).

Assuming that was taken care of though, I wholeheartedly support curved runways

[Snikersnee]

SCIENCE! It has never been a problem for me but I suppose it wouldn't hurt to fix. 

[DoToH]

10 hours ago, Rdivine said:

Making the runway slightly curved would go a long way to solving problems of vessel having trouble landing, vessel rolling forwards and crashing etc.

Given the diagram and explanation (and assuming engines off  and brakes released), the vessel would roll forward  increasing speed until the exact half leght of the runway. From that point, the runway is going upwards. The vessel would lose speed while on the 2nd half of the runway and would stop before the runway end. I think KSP doesn't handle wheel friction, but drag is handled, so the vessel would stop even before.

Your theory explains the roll forward, but not the crash. Nobody is crashing because of this. And I fail to see this as the cause of  vessel having trouble landing. If this is the origin of  a forward phantom force, it woul be nice to have it solved, but, IMO, solving this should be at the end of the priority list.

[Rithsom]

If the runway were made curved, wouldn't it end up having edges because it's not possible to have a perfectly curved mesh? What I mean is this:

While that image is exaggerated, I'd still expect this to be a problem even if the edges are spaced out 50 meters. I would expect them to be most noticeable when seeing a plane roll down the runway, pitching slightly as it rolls over each one.

Edited March 24, 2016 by Rthsom

[Anquietas314]

3 minutes ago, Rthsom said:

If the runway were made curved, wouldn't it end up having edges because it's not possible to have a perfectly curved mesh? What I mean is this:

Well, yes, technically, but it wouldn't be difficult to divide the runway into enough segments that the effect would not be noticeable. If you used 3 segments as in your diagram, obviously it's not going to work very well, but even 100+ segments isn't going to tax your GPU very much. The early game runway meshes are probably already in that ballpark since they aren't uniform rectangles.

[max_creative]

That's a great idea! And I would help with my planes because they already roll. a lot. Yeah it would be hard to change the curve for mods like RSS. But this would be something I think would be useful. And the stock brakes are terrible. So yeah. This would be useful. Not super duper important to add in, but useful.

[Hodari]

You see those space centers from Kerbin?  They've got curved runways.  Curved.  Runways.

 

Seriously though, it's only an issue at all when first launching an aircraft and would probably be easier to solve by just having the brakes turned on by default.

Edited March 24, 2016 by Hodari

[zolotiyeruki]

As an alternative, maybe wheels should have some (small) amount of minimum friction all the time. Not enough to stop you when you land, but enough to keep an already-parked craft from rolling away.

[Padishar]

On ‎23‎/‎03‎/‎2016 at 1:14 PM, Rdivine said:

Any vessel would experience a force slightly sideways relative to the surface of the runway ( red arrows ),

because the runway is inclined to the true tangent of the surface at that point ( dotted red lines ).

Actually it is even more involved than this, the whole of the KSC is defined on a flat plane and the tangent point to the planet is actually at the VAB so the entire runway actually tilts slightly to the right.  This has been brought up before:

...and the slope along the length was talked about long before that (before the great forum wipe, just after the runway was first introduced).  I suspect there will be lots of higher priority things on Squad's list for quite a long time to come...

[One-Way Films]

*snip*

Edited March 25, 2016 by One-Way Films
wrong thread *facepalm*

[Padishar]

2 minutes ago, One-Way Films said:

Hope I'm not too late to board. One 1st class ticket please

No, but you're definitely in the wrong thread...

[One-Way Films]

...ah

xD

whoops

had two open at the same time

Edited March 25, 2016 by One-Way Films

[HebaruSan]

Alternately, a stock option to have the brakes on by default would also work.

[kiwi1960]

I thought everyone knew this....

But its not universal... space planes, yes... but some runways have what you would think was a ramp .. much like a jump jet ramp...

 

And then, as I said, some runways have this effect as well...

 

[Daid]

On 23-3-2016 at 2:22 PM, cubinator said:

So that's why they always do that! Huh. I like the idea, but for me it would just be more of a 'quality of life' thing. Not exactly vital to the game, but it should definitely be considered.

Kerbin has a radius of 600km. The runway is 2526m. By the diagram, half it is "pointing outwards of Kerbin", so 1263m. The sidewards acceleration should be only the division of this length by the radius of the planet. So 1263/600000=0.002105 is the fraction of the gravity that is applied forwards.

So 0.002105*9.81=0.0193239m/s^2.

I think the fathom acceleration we see is more then that. As with this acceleration it should take about a minute to get up to 1m/s.

[Snikersnee]

16 hours ago, kiwi1960 said:

I thought everyone knew this....

But its not universal... space planes, yes... but some runways have what you would think was a ramp .. much like a jump jet ramp...

 

And then, as I said, some runways have this effect as well...

 

If you can land and take off from it, its a runway.

[Padishar]

3 hours ago, Daid said:

I think the fathom acceleration we see is more then that. As with this acceleration it should take about a minute to get up to 1m/s.

Have you actually tested this?  I have, and it takes almost exactly 50 seconds to get up to 1m/s (using the stock Aeris 3a) which I make spot on from your figures...

Edited March 25, 2016 by Padishar

[Daid]

21 hours ago, Padishar said:

Have you actually tested this?  I have, and it takes almost exactly 50 seconds to get up to 1m/s (using the stock Aeris 3a) which I make spot on from your figures...

Didn't test it, that's why I added "I think". If tests shows this amount of acceleration, then this could be the reason for the acceleration.

Kerbal engineer should also show a slight decrease in altitude then if you roll down the runway.

[Padishar]

3 hours ago, Daid said:

Didn't test it, that's why I added "I think". If tests shows this amount of acceleration, then this could be the reason for the acceleration.

There's no "if" about it.  This is definitely the cause of the slow roll along the runway.  If you take a look at the album in the thread I linked in my post above then you can clearly see the slope indicated by KER.

3 hours ago, Daid said:

Kerbal engineer should also show a slight decrease in altitude then if you roll down the runway.

Yes, it shows this too, the rover is always reported as 1.147m above the terrain but at the start of the runway it is 70.48m above sea level, around the middle it is 68.98m and at the end it is 70.08m...

[Wjolcz]

It's more of a scale problem than the runway itself. I say leave it as it is.

Having a second north-south oriented runway with a ramp at the end on the other hand... Oooooh, yes. Yes, please.

[kiwi1960]

On 25/03/2016 at 5:11 AM, Snikersnee said:

If you can land and take off from it, its a runway.

Ummm... OK, so you land a Jumbo on a highway... pretty sure the Police will not be happy with you...

and the Judge will surely argue you point with you before slapping you with a hefty fine....