The Speed of Ground - A study into how KSP terrain colliders handle land vehicles with high magnitudes of velocity.

[Poodmund]

Planet meshes are still PQS based. Check out this presentation from around 30mins 30secs in, it'll explain a lot about how KSP renders it's LOD.

 

[suicidejunkie]

The maximum level for Kerbin varies depending on your terrain quality settings, but it can be as high as 11 subdivisions, which is indeed a pretty dense mesh for the surface, but that level of depth only exists for a few quads around you. As you move about, quads that you approach will subdivide to max level, while quads you leave behind will collapse back into a lower level.

This sounds like fodder for an Experiment!

Can you get a higher land speed by reducing the terrain quality setting so that the quads that are maximum level are physically larger, and thus have bigger colliders loading further away?

[Waxing_Kibbous]

I'll put this here, which may or may not be relevant to the conversation:

 

[magnemoe]

On 17.5.2017 at 8:39 PM, monophonic said:

Yes, you have managed to outrun KSP's ability to keep ground under you. You are right that it only considers a rectangular area around your craft for physics. I found some word of god about it from an old reddit thread.

 

This is pretty common, the elders scroll games and fallout does the same, however here its more about loading AI for npc and enemies. 
You can see this in the game, in skyrim you get an quest who track an khajiit caravan, in Obvion and Fallout 3 you also get quest markers to moving npc. 
They will move but not be attacked by enemies outside of the load zone, once inside they will get attacked. 
This also give some weird results 

[Stratzenblitz75]

7 hours ago, suicidejunkie said:

 

 

This sounds like fodder for an Experiment!

Can you get a higher land speed by reducing the terrain quality setting so that the quads that are maximum level are physically larger, and thus have bigger colliders loading further away?

Interesting idea!

I will give it a shot and let you know how it goes.

3 hours ago, Waxing_Kibbous said:

I'll put this here, which may or may not be relevant to the conversation:

 

Indeed, it is relevant. Its another instance where a mod changes how the stock game loads the colliders. In this case, Bradley used the better time warp mod's slo mo feature to get around the stock limit. But why does this slow-mo work? Stock time-dilation does not have this effect; during my low FPS runs, the time dilation slowed time to a fraction of normal speed, but this had no effect on the top speed reached.

Perhaps the mod maintainer, @linuxgurugamer, could give us some insight?

[konstantinua00]

Since there is chunk loading...
Have anyone tried different start points?

If chunk loading disappearence starts at the opposite edges of a chunk, you can get up to chunk diagonal length difference

[Guest]

Is this the general relativity theory of KSP?

[Darth Jeb]

Wow, that was truly a scientific approach. 

[Jas0n]

Have you tried testing whether physics delta time changes the max speeds?

[nhnifong]

Very clever!

If the terrain collider doesn't exist past that speed though, why does the craft sink a little bit and then explode? It clearly hit some other (possibly lower resolution) collider right? Or else it would have sunk about 10 meters to the ocean under the ice cap.

[suicidejunkie]

54 minutes ago, nhnifong said:

Very clever!

If the terrain collider doesn't exist past that speed though, why does the craft sink a little bit and then explode? It clearly hit some other (possibly lower resolution) collider right? Or else it would have sunk about 10 meters to the ocean under the ice cap.

My interpretation was that the next collider doesn't exist when you reach the edge of your current one.  You then start to fall.  Just a moment later, the collider loads in, and rips the wheels off.

[Waterlimon]

Could you keep ksp pre-loading these terrain chunks by switching through an array of pre-positioned ground stations parallel to the race track, ahead of the speeder?

[Stratzenblitz75]

On 8/2/2017 at 10:08 PM, Jas0n said:

Have you tried testing whether physics delta time changes the max speeds?

Yes I did. Turns out, the physics delta makes no difference to the max speed. The only time it does make a difference is when the FPS is very low (10 ~fps), and the physics delta is high (0.12). In this case, you will experience the bug much sooner.

 

On 8/4/2017 at 9:34 AM, nhnifong said:

Very clever!

If the terrain collider doesn't exist past that speed though, why does the craft sink a little bit and then explode? It clearly hit some other (possibly lower resolution) collider right? Or else it would have sunk about 10 meters to the ocean under the ice cap.

I believe there is some collider below the surface that instantly destroys anything that touches it. That said, I've managed to get around it sometimes. On a few runs, I've had the craft fall through this collider into the ocean. Its very inconsistent though.

 

2 hours ago, Waterlimon said:

Could you keep ksp pre-loading these terrain chunks by switching through an array of pre-positioned ground stations parallel to the race track, ahead of the speeder?

I don't think this is possible. In stock KSP you can't switch vessels while "moving over the surface". 

 

On 5/19/2017 at 11:12 AM, suicidejunkie said:

 

 

This sounds like fodder for an Experiment!

Can you get a higher land speed by reducing the terrain quality setting so that the quads that are maximum level are physically larger, and thus have bigger colliders loading further away?

So, I just got around to testing this and YES. A lower terrain quality does increase the maximum speed reached significantly. With this knowledge, it should now be possible to break escape velocity over the surface.

[cubinator]

22 hours ago, Stratzenblitz75 said:

On 8/5/2017 at 8:48 AM, Waterlimon said:

Could you keep ksp pre-loading these terrain chunks by switching through an array of pre-positioned ground stations parallel to the race track, ahead of the speeder?

I don't think this is possible. In stock KSP you can't switch vessels while "moving over the surface". 

I don't think you would have to switch to them. As long as they are within physics range the game should load the terrain under them.

[Stratzenblitz75]

Hey everyone! I want to thank you all for giving your inputs in this thread. I've certainly learned a lot from the information you've brought to the table, and the ideas you have given me.

It is because of your inputs (@suicidejunkie 's suggestion in particular) that I've managed to break escape velocity with a stock land-speeder:

You made this video possible!

-Stratzenblitz75

[kjack1111]

I look at your land speed records and i raise my "Yall crazy" sign. Good work, stratzen.

[cantab]

That was epic.

But remember, there's another escape velocity ...

[Kosmonaut]

Science done right.

Good job

[F3ARSTUNTS]

Im working on a car of my own, but i mainly challenge myself and see how light of a craft i can get to where ever, anyway my car either spins out, bounces up, or spontaneously explodes, help?

[Stratzenblitz75]

36 minutes ago, F3ARSTUNTS said:

Im working on a car of my own, but i mainly challenge myself and see how light of a craft i can get to where ever, anyway my car either spins out, bounces up, or spontaneously explodes, help?

There's two main techniques I use to make my craft more stable. For one, reduce the friction control on the wheels. In KSP, the friction control is wholly incapable of handling high-speeds so you need to set it to as low a value you find tolerable. For all of my runs, I had the wheel friction control set to 0 so that it wouldn't interfere with the performance of the craft. Of course, without friction control, your wheels don't get any grip so you have to raise it again when going slow. 

The other technique is downforce. When going fast, you will need some force to keep you pressed against the ground so that your car doesn't become an airplane. For this, I find flaps very effective because you can fine tune their deployment angle.

If you're willing to share craft pics, I can give you more specific advice.

Hope this helps!

[F3ARSTUNTS]

14 minutes ago, Stratzenblitz75 said:

There's two main techniques I use to make my craft more stable. For one, reduce the friction control on the wheels. In KSP, the friction control is wholly incapable of handling high-speeds so you need to set it to as low a value you find tolerable. For all of my runs, I had the wheel friction control set to 0 so that it wouldn't interfere with the performance of the craft. Of course, without friction control, your wheels don't get any grip so you have to raise it again when going slow. 

The other technique is downforce. When going fast, you will need some force to keep you pressed against the ground so that your car doesn't become an airplane. For this, I find flaps very effective because you can fine tune their deployment angle.

If you're willing to share craft pics, I can give you more specific advice.

Hope this helps!

Thanks man, once i get finished fine tuning my speeder, ill show you some pics and maybe you can give me some advice.

[F3ARSTUNTS]

@Stratzenblitz75 here are the pics i promised, the first few stages make the craft drife side to side, and the sepratron stage either flys into the sky, blows up, or does a full 180. http://imgur.com/a/7iwbN

 

[selfish_meme]

@Stratzenblitz75 does the mod physicsrangeextender affect the collider loading range? Harvester said collider loading is slow, so you may in fact reach an edge before the next collider loads, which leads me to my  theory that Kopernicus loads more detailed terrain, but also extends the higher quality terrain further, so more plane division and more colliders, further.

Edited August 16, 2017 by selfish_meme

[Stratzenblitz75]

2 hours ago, F3ARSTUNTS said:

@Stratzenblitz75 here are the pics i promised, the first few stages make the craft drife side to side, and the sepratron stage either flys into the sky, blows up, or does a full 180. http://imgur.com/a/7iwbN

 

Alright, after reviewing your design, I see some areas where you can make improvements.

For your first stage, it doesn't look like you have enough wheels. Evenly supporting your craft is very important for stability, so try to distribute your wheels evenly to balance the mass of your speeder. In the case of your first stage, add some more wheels to the back to support the mass of those three vectors. Also, you'll need some fins on your first stage to keep your craft going straight. This is because the ablative fairing at the front creates a lot of drag and will tend to cause the craft to flip. This true of all of your stages. These two adjustments should get rid of the side-to-side wobbling you see at the start. Also, giving your vectors more gimbal can help with stability.

That said, right now, your final stage needs the most work. First, I suggest you opt for using 3 or 4 wheels instead of 2. You can get 2 or even 1 wheeled designs to work, but they take far more effort to balance properly. Having at least 3 wheels will ensure your craft's stability without sinking a ton of time into balancing its aerodynamics. Additionally, you need a tail to keep your final stage going straight. Once again, this is because of the drag the front fairing produces. I'd suggest using a rearward facing fairing plus some control surfaces to keep it pointed straight. If you want to use seperatrons for the final stage, you need to clip them into a fairing or otherwise they produce an insane amount of drag. My favorite way of doing this is to angle them by 15 degrees, and attach them to an octagonal strut to create a ring of outward angled seperatrons. You can then place this assembly inside a fairing so that just their nozzles poke out. This system is especially useful because you can simply add more rings if you need more thrust. If you want to avoid clipping of any kind, then don't bother with seperatrons; use only vectors because they have the best combination of low-profile and high thrust.

One more thing for your final stage; make sure the flap you have mounted to the bottom is deployed upward slightly, but not too much. At high speeds, you will need a down force to keep your craft on the ground. If its too low, your craft will easily bounce off the ground. If its too high, your craft will be smashed into the surface. It is a fine balance so you'll have to do several runs to find the right angle of deployment.

In general, make the "show aerodynamic forces" option your best friend. This handy stock feature allows you to see what parts are creating drag and lift, and how much. I found it invaluable for locating high drag parts, and either replacing them or adding fins to counteract their drag.You can enable it with F12, or through the debug menu.

Hope this helps!

BTW, what was the highest velocity you achieved with your design?

Edited August 17, 2017 by Stratzenblitz75

[F3ARSTUNTS]

1 hour ago, Stratzenblitz75 said:

Alright, after reviewing your design, I see some areas where you can make improvements.

For your first stage, it doesn't look like you have enough wheels. Evenly supporting your craft is very important for stability, so try to distribute your wheels evenly to balance the mass of your speeder. In the case of your first stage, add some more wheels to the back to support the mass of those three vectors. Also, you'll need some fins on your first stage to keep your craft going straight. This is because the ablative fairing at the front creates a lot of drag and will tend to cause the craft to flip. This true of all of your stages. These two adjustments should get rid of the side-to-side wobbling you see at the start. Also, giving your vectors more gimbal can help with stability.

That said, right now, your final stage needs the most work. First, I suggest you opt for using 3 or 4 wheels instead of 2. You can get 2 or even 1 wheeled designs to work, but they take far more effort to balance properly. Having at least 3 wheels will ensure your craft's stability without sinking a ton of time into balancing its aerodynamics. Additionally, you need a tail to keep your final stage going straight. Once again, this is because of the drag the front fairing produces. I'd suggest using a rearward facing fairing plus some control surfaces to keep it pointed straight. If you want to use seperatrons for the final stage, you need to clip them into a fairing or otherwise they produce an insane amount of drag. My favorite way of doing this is to angle them by 15 degrees, and attach them to an octagonal strut to create a ring of outward angled seperatrons. You can then place this assembly inside a fairing so that just their nozzles poke out. This system is especially useful because you can simply add more rings if you need more thrust. If you want to avoid clipping of any kind, then don't bother with seperatrons; use only vectors because they have the best combination of low-profile and high thrust.

One more thing for your final stage; make sure the flap you have mounted to the bottom is deployed upward slightly, but not too much. At high speeds, you will need a down force to keep your craft on the ground. If its too low, your craft will easily bounce off the ground. If its too high, your craft will be smashed into the surface. It is a fine balance so you'll have to do several runs to find the right angle of deployment.

In general, make the "show aerodynamic forces" option your best friend. This handy stock feature allows you to see what parts are creating drag and lift, and how much. I found it invaluable for locating high drag parts, and either replacing them or adding fins to counteract their drag.You can enable it with F12, or through the debug menu.

Hope this helps!

BTW, what was the highest velocity you achieved with your design?

Thanks a million man, I'd say the highest speed I've reached so far is 700-800m/s