Strange Physics

[BBQ]

NOTE: Apparently this bug is already known -- so I deleted attached files to free up my attachment space.

 

Bug description:

 

Upon an emergency landing (crash) and coming to a complete stop, stopping engines/cut-off valves and finally the rotor brake, the Ka-50 continued to move around though there weren't any apparent forces acting on the helicopter.

 

I would expect, upon the rotor coming to a complete stop, as well as the helicopter itself, that any motion would cease.

DCS version 1.2.3

 

Steps to reproduce / Conditions of occurrence

 

Crash landing with landing gear not extended

Fast-forward to 06:45 in the mission track and view behavior

 

PC Configuration

 

See attachment

 

Attachments

 

ALIEN.txt (PC specs)

emergency.trk

Log files

 

***NOTE: zip file is password protected as mission is copyrighted -- from the Vergeev Group Campaign. I didn't want to make it available to the public obviously -- so please PM me for password.***

 

 

Edited April 18, 2013 by BBQ
bug already known so I deleted attaments

[msalama]

Are we talking about the "rubber ground" syndrome? If so, it's been like that as long as I remember. Alas, not a lot one can do about it either AFAIK...

[BBQ]

Not sure what the "rubber ground" syndrome is -- it's more like "Tourette's Syndrome" actually -- and I can say that, since I have it!

 

What does the "rubber ground" syndrome look like?

[BaD CrC]

Kinda like what you are describing. Exists since BS1 1.00. Strange but not a biggie compared to other slightly more annoying bugs that will eventually be addressed in 1.2.5 to the earliest.

[BBQ]

Apparently this issue is already known, so I'm going to delete file attachments.

[msalama]

What does the "rubber ground" syndrome look like?

 

A belly-landed and unpowered chopper shakes, moves around and sometimes jumps slightly, as if the ground was made of extremely bouncy and reactive rubber or something. Been there from the start AFAIR and there's nothing we can do about it unfortunately :(

[Devrim]

Yes this is already known bug.

 

I thought it was front wheel. Because when I did an emerg.landing, front wheel had gone crazy. Until the that day I lost all wheels. One day I did an emerg.landing whithout wheels. And that shaking was still present.

 

Then I understood, it's not front wheel. :P

[BBQ]

Then I understood, it's not front wheel. :P

 

 

:lol:

[-MadCat-]

The thing causing this is likely to be the numerical calculation and computing of contact forces between 2 (mostly) solid objects (ground vs. helicopter frame (reduced to these 2 for sake of simplicity)).

 

In a numerical simulation model it's rather difficult to reach a state of absolute 100% equilibrium, so force A equals precisely force B and all movement (actually acceleration) come to a halt.

What I think happens, is that the calculation is fluctuating between very small positive and very small negative values (A - B = +-0,## ) and doesn't come to settle at exactly 0.

 

One could introduce damping of the calculations, but that would introduce delay as well and result in weird "slow motion" crash behavior, more exaggerated the higher the damping factor is chosen.

 

We may have to deal with how it is now.

[Oznerol256]

The thing causing this is likely to be the numerical calculation and computing of contact forces between 2 (mostly) solid objects (ground vs. helicopter frame (reduced to these 2 for sake of simplicity)).

 

In a numerical simulation model it's rather difficult to reach a state of absolute 100% equilibrium, so force A equals precisely force B and all movement (actually acceleration) come to a halt.

What I think happens, is that the calculation is fluctuating between very small positive and very small negative values (A - B = +-0,## ) and doesn't come to settle at exactly 0.

 

One could introduce damping of the calculations, but that would introduce delay as well and result in weird "slow motion" crash behavior, more exaggerated the higher the damping factor is chosen.

 

We may have to deal with how it is now.

 

One could monitor the results of the calculation. This way you are able to detect when the results are almost 0 for a long time (e.g. a second). Then you could set the forces to zero. Keep the numeric calculation running and if it returns something that is not almost zero, you can apply the forces again.

This way it is possible to stop crazy bumping without inducing a "damper".

[sobek]

Yes, this is an unwanted side effect of the collision detection.

[sobek]

One could monitor the results of the calculation. This way you are able to detect when the results are almost 0 for a long time (e.g. a second). Then you could set the forces to zero. Keep the numeric calculation running and if it returns something that is not almost zero, you can apply the forces again.

This way it is possible to stop crazy bumping without inducing a "damper".

 

If you set the forces caused by the interaction with the ground to zero, it will start to fall through the surface.

[Oznerol256]

If you set the forces caused by the interaction with the ground to zero, it will start to fall through the surface.

 

In this case, set them to a value that causes the aircraft to stand on the surface.

[sobek]

In this case, set them to a value that causes the aircraft to stand on the surface.

 

That is highly orientation dependant, how do you get the values that cause it to be in an equilibrium? (I'm playing devils advocate right now, if you didn't notice. ;))

 

Anyways, without intricate knowledge of how the system works, trying to come up with solutions to this problem is a futile exercise.

[Oznerol256]

That is highly orientation dependant, how do you get the values that cause it to be in an equilibrium? (I'm playing devils advocate right now, if you didn't notice. ;))

 

Anyways, without intricate knowledge of how the system works, trying to come up with solutions to this problem is a futile exercise.

 

It was just an idea of me. If the system works in some strange way, i cant help.

[Pikey]

How about a governor module to detect if two items are oscillating and set them to zero? It can be spawned on collision detection to reduce CPU load, but would ultimately reduce CPU load. You see this behavior in Source games like HL2 and garrys mod where items start bouncing around at low energy.

[sobek]

How about a governor module to detect if two items are oscillating and set them to zero? It can be spawned on collision detection to reduce CPU load, but would ultimately reduce CPU load. You see this behavior in Source games like HL2 and garrys mod where items start bouncing around at low energy.

 

As i know neither the collision code of source nor of TFCSE, i really don't know. However, setting something to "zero" (don't know exactly what you mean, the velocities possibly?) might have undesired sideeffects (think of a helicopter in ground oscillation, which is an effect that can occur and should not be filtered).

[TimeKilla]

Noticed this but not a big deal as it's time to quit anyway when you end up like this/ :lol: