Lee Elliott said:

> Hello All,
> I'm having trouble with a couple of FG problems.
> The less serious of the two concerns the weather settings.  I'm using the 
> default weather settings, except that I've increased the visibility ranges 
> but I'm finding that the actual weather 'level' I'm getting, apart from 
> initial start-up on the ground, is pretty random.  For example, more often 
> than not, after taking off, as I pass through 500ft the weather settings 
> transition to the 6000ft or 9000ft level so I'm getting up to 30kt crosswinds 
> at a little over 500ft.
> I should point out that this is with airfields that are < 200ft alt.
> Sometimes it doesn't happen and the weather progresses smoothly through the 
> 3000ft, 6000ft and 9000ft levels, but this is pretty rare.
> Opening the weather settings gui and selecting 'Reset' seems to be pretty 
> random too - sometimes it'll correct the weather but quite often it'll do 
> nothing or just select another incorrect level - for example, I can hit reset 
> several times while at < 1000ft and I can see (and check via the property 
> browser) that I'm getting the 6000ft or 9000ft levels.  Frequently I'll get 
> the right weather level by hitting reset only for it to jump back to the 
> 9000ft level when I hit 'ok' to close the gui.
> Once I've finally got the correct weather for the level I'm flying at and 
> closed the gui without messing it up, the weather then seems to behave 
> itself.
> The more serious problem is with nav1 ils guidence.  I can get the a/c onto 
> the glide slope and following the nav1 heading nicely but at approx 1 mile 
> from the runway the nav1-heading-error-deg can suddenly jump from < 0.01 deg 
> to about 4 or 5 deg.  While smaller aircraft can cope with this, something 
> like the AN-225 can't.
> I can tune the pid controllers so it can cope with this jump but then it can't 
> get on to the nav1 heading and stay there without badly overshooting (by up 
> to a couple of miles) several times before it settles down, by which time 
> it's almost on the runway.
> I've noticed that if there's any degree of crosswind, although the 
> nav1-heading-error-deg might be showing < 0.01 deg, the hsi needle is clearly 
> showing that I'm off to one side of the GS and I suspect that the jump I'm 
> seeing is some sort of correction for this.
> Neither of these problems seem to be dependent on the a/c - I've experienced 
> them with a variety of a/c.
> Is anyone else experiencing these problems?  They aren't particularly new - 
> I've been seeing them for quite a while and they're still occurring with the 
> cvs update I did a couple of hours ago.

Hi Lee,

Sorry for taking so long to respond to this.  Have you taken a look at the 747
config?  For AP using the nav1-heading-error I need to go back to that and
smooth out the transition a little so it isn't so sudden.  Basically once we
get close enough to the ils (1 mile seems short, but maybe not at some
airports) the nav1 error output changes behavior slightly to give reasonable
performance close in.  This tightening screws up the process of intercepting
the radial (automatic turn to final), hence the change at a certain distance out.

In any case it'd be interesting to know if this method is anything like how a
real AP works,  both in the AN-225 and others, like the c172.  My guess is
that it isn't even close, and the whole heading intercept espeicially and
nav1-heading-error method we're using is wrong.  Maybe we can treat
interception and ils hold as two seprate functions. And end up with something
on ils that runs purely off the needle "deflection" value without
incorporating "distance" or converting to anything else e.g. degrees (might
have to "normalize" it based on "to/from").  Obviously if you have any cross
wind at all you don't want to be seeking radial heading which is basically
what the nav1-heading-error thing does.

That said, I think that you'll find the 747's method (using the gps flight
path output to calculate flight track error) more effective.  It manages to
plunk that pig right on the centerline even in a fairly brisk crosswind.



Flightgear-devel mailing list

Reply via email to