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

[<js...@ha...>]

> Jon S. Berndt wrote:
> > Eqn. 5.2 in this document should also be pertinent:
> >
> > http://www.simlabs.arc.nasa.gov/library/mcfarland/Basis.pdf
> 
> Ah, here is the good voodoo.
> 
> What happened is that I mistook the meaning of "local" coordinate
> system.  It doesn't mean local-to-the-aircraft (that's what they call

I thought this would help! :-)

> But I lucked out.  The local frame is STILL an accelerated reference
> frame (if the plane is turning, then the local frame is tracing out a
> circle relative to the ground -- that's acceleration).  Thus, all the
> points about not being able to rely on Newton's laws in an accelerated
> reference frame still apply.  The McFarland paper actually tells you
> that this frame is non-inertial on page 2 (paragraph 3), it just
> (sadly) skips the discussion of the implications of this fact.

We are working on correcting this in a way that can be derived and understood 
and shared.

> As far as references, you could take a look at your college intro
> physics textbook.  I have one still on my shelf ("Physics", 2nd ed.,
> by H. Ohanian) that has a really good discussion about this stuff in
> chapter five.  You could also try a google search on "inertial

Well, I've got my orbital mechanics books, several books and papers on flight 
dynamics, etc. I had never looked at the pilot accels calcs, so this is a first 
for me. Tony is really working the problem, but I want to understand it, too. I 
want to be able to *derive* it.

> The bottom line is that the "local" frame is a non-inertial reference
> frame.  Physics in non-inertial reference frames doesn't obey Newton's
> laws.  In fact, most of the 53 page McFarland paper is dedicated to
> explaining the bizzare and counter-intuitive corrections to apply to

Point taken. There are reasons for using the frame, however, for reporting 
certain values to the instruments and calculating certain other values. 
However, using an earth centered inertial frame or something else has always 
been on my mind as a possibility for future consideration. Documentation to 
explain what we are doing in more detail would also help, I am sure.

> Really, I'd strongly suggest ditching the local frame as a calculation
> medium.  You can still _use_ it to calculate altitude and gravity and
> whatnot.  But don't do calculations in it, that way lies madness.
> Calculate your force and torque in the body frame, transform to the
> earth frame, and apply them (and integrate them, and differentiate
> them) there.  Everything will work much more simply -- no FunnyValues.

I think things are not as simple as you think they are. Also, forces and 
torques, etc. *are* calculated in body frame, with some minor variations. For 
instance, aero forces are calculated in stability axes, then transformed into 
body frame. Some ground reactions are by necessity calculated in local frame, 
then transformed into body axes.

Still, if you really feel we can do better in the EOM calculations, study the 
code and show us how we can make it better.

Jon