Curtis L. Olson writes:

 >  - respect runtime property changes (i.e. new altitude or location)
 > 
 > I would respectfully suggest that this one not be implimented or at
 > least be put at a low priority.  We tried to do this in JSBsim and it
 > can really get messy.

The problem was that the startup trimming routine and the properties
always ended up arm-wrestling.  In principle, this should be very
easy:

1. Before each iteration, copy all state variables in from the
   bus/property tree.

2. After each iteration, copy all state variables back out to the
   bus/property tree.

YASim already does #2; it simply needs to add #1.

JSBSim retrims automatically when certain values are changed outside
the FDM unless certain properties are set, etc. -- it's all fairly
confusing.

Over all, I think it would be better if none of the FDMs trimmed
automatically.  FlightGear, which provides the primary user interface,
is in a much better position to know when trimming is required; for
example, you do want to trim when an altitude and speed are selected
on the command line, but you do not want to trim when a flight is
being restored from a save file.  If the FDMs simply retrim for steady
state when requested, we can make sure the request is issued when
needed.  That should clean up the JSBSim/FlightGear interface code a
bit as well.

Tony: what do you think?

 > I think it would be much cleaner to force a reset to a new location
 > each time we warp to a new location.  This allows us to delete the FDM
 > instance and create a new one so it can be freshly inited and trimmed
 > for the new conditions.  Otherwise, it's really hard not to carry over
 > some state from the previous location which can cause obscenely large
 > forces and other wierdness.

The problem is that when we restore a saved flight or start a premade
scenario, we'll get bumps (etc.) from the trimming routine, when the
saved state was already (presumably) steady.  I think we're pretty
close -- you want to force a reset, and I want to be able to request a
retrim.


All the best,


David

-- 
David Megginson, [EMAIL PROTECTED], http://www.megginson.com/

_______________________________________________
Flightgear-devel mailing list
[EMAIL PROTECTED]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel

Reply via email to