David Luff wrote:
William Earnest writes:


Hello,

This morning the 0.9.8 package in WinXP, running normal defaults (KSFO, c172p) starts with all instruments normal. The current CVS compiled here in Fedora 3 is much the same, but sitting on 24R, the DG looks as expected, but the mag compass is showing North. Is some xml file dropping the bearing info?



Warning - the following is just a wild guess.  I think that this might be due 
to excessive side forces reported by JSBSim when the aircraft is stationary.  
This results in the instrument physics code 'locking' the compass as though it 
was forced to the side of the cage until the force subsides.  Once you start 
rolling down the runway the compass unlocks, but it's a pain for setting the DG 
when stationary.  That's my guess as to what's happening - I could be 
completely off base, but it doesn't seem to affect yasim models IIRC.

Cheers - Dave

_______________________________________________
Flightgear-devel mailing list
[email protected]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel
2f585eeea02e2c79d7b1d8c4963bae2d


Dave,
Yep, it sounds like that may be it. If I start a slow roll, and trace a zigzag path with the pedals, it "shakes it free" and the compass eventually settles to a proper reading. Just that I've never seen an oil-damped unit bind that way in real life. The mounting needs a bit more freedom.


While tracing the compass problem, I started using my usual .fgfsrc file, putting me on Rwy 25 at 1N9, my home field. At startup, I found the plane about 100 feet above the runway, with the nose pointed up at about 20 degrees. After about 30 seconds of freeze this way, the plane nosed to vertical and made a spectacular crash.

I traced this to using the realtime weather option. After last evening's snow and today's wind, the METAR from KABE is a bit messed up, and the parser that extracts the conditions went wild and clobbered who-knows-what in the properties. If anyone wants to trace this and add some sanity checks, the METAR used seems to be this:
KABE 231651Z AUTO 33014G22KT 10SM M10/M17 A2986 RMK A02 SLPN0 T11001167 TSNO $
I pulled the above for KABE from ADDS while the mess was happening.
The weather applet in Fedora 3 also complained about invalid and missing fields.


--
    Bill Earnest  [EMAIL PROTECTED]  Linux Powered   Allentown, PA, USA
Computers, like air conditioners, work poorly with Windows open.

_______________________________________________
Flightgear-devel mailing list
[email protected]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel
2f585eeea02e2c79d7b1d8c4963bae2d

Reply via email to