Re: [Flightgear-devel] Turn ball fix & JSB notes

[<an...@pl...>]

Tony Peden wrote:
> Andy Ross wrote:
> > The UVWdot value comes from the sum of the forces PLUS another
> > factor that I was (rather uncharitably -- sorry) calling FunnyValue.
> 
> Yes, and that term is what corrects our values derived from an accelerated 
> and rotating reference frame back to a so-called "inertial" one.  

Bingo.  But the force felt by the pilot and other objects (like, say,
the ball) MUST be defined relative to the global reference frame.  If
you don't do this, you won't get the right answers for acceleration.
This is just the Way Things Work, I don't know any better way to
explain it.  Some transforms* to the coordinate frame are legal and
preserve the laws of nature, others don't.  Accelerated motion
doesn't.

So when calculating the acceleration on the pilot, you need to look at
the UVWdot value _before_ the correction term is applied.  This is
exactly what my patch does.

But again, don't let my inability to explain the theory dissuade you
from examining the experimental evidence.  Before the patch, the z
component of the pilot acceleration drops to zero in a sustained turn.
Afterwards, it maintains the proper value.  Try it. :)

Andy

* For Newtonian mechanics, translations, constant rotations and
  uniform/unaccelerated motion are kosher.  If you're talking special
  relativity, then Lorenz transforms are the deal.  Particle
  physicists talk about gauge symmetries to mean exactly the same
  thing.  But the point is you can't just do _anything_ to the
  coordinates and expect things to work the way they did in the old
  frame.  The universe has rules.

-- 
Andrew J. Ross                NextBus Information Systems
Senior Software Engineer      Emeryville, CA
an...@ne...              http://www.nextbus.com
(510)420-3126                 Why Wait?