I've managed to recreate the excess G bug

[Yeffers]

I have made a scenario which for me recreates the excess G bug every single try. I have never
experienced this bug before until I tried to put 2 DGIIs into the same scenario. To recreate the
bug, load the scenario, switch to the 2nd glider (GL-02) and undock from the ISS. As soon as you
undock, the glider is pulling about 0.5 g's and if you use retro thrust for about 4 seconds, g
exceeds 10 and everyone passes out.

Here is the scenario:

http://users.bigpond.net.au/yeffs/G-Bug.scn

[DanSteph]

Sorry for late reply,

This "bug" is framerate related it seem that when you have very low
framerate Orbiter make some "jump" in the position
Position jumping logically give G value because it's an acceleration

So if you have low spec machine this bug will appear when you have a
lot of things on screen as in the scenery you give where you have two DG2,
ISS and the earth in view wich should lower the framerate.
Notice that I didn't get anormal value with your scenario because my
computer is a P4 with geforce 4. (It run at about 63fps)

The solution is to try to keep the framerate at an acceptable value
lowering the eyes candy that eat framerate or to buy a better computer....

Sorry there is stricly nothing I can do here if orbiter make you ship
jumping I cannot detect in any way if this a normal move or not.

Dan

ps: as a workaround you can disable crew death in Dg2config.exe

[Yeffers]

This is strange, as I run Orbiter on an Athlon XP2000 with a Geforce 4. Did you switch between the
ships before you undocked? I have never experienced the problem until I tried a scenario with two
DG2's. I'm not at all experienced with the code behind the program, but it seems that the 2nd DG2
is trying to calculate its acceleration based on data from the first, thus resulting in the excess g. It
definately isn't measuring acceleration anymore, the G value reads more like velocity relative to
the other ship, so when I am moving 5m/sec away from the other DG2 I get 5g's.

That said, maybe it is specific to my system, and I guess it isn't that important, but it definately
isn't frame related.

[DanSteph]

If you switch ship, In this case this is possible,
(even if you don't it may be possible but I don't recall exactly)

let me explain:

The way orbiter execute multiple ship's code (of the same class)
is rather strange and leaded for the strangest bug I saw
in my life...


The first month I coded the dgII with a scenery that contained two
DGII I was hurt by a bug that was caused by a value that was
TRUE and FALSE at the same time I never saw that before
in 15 years of coding.... I tried then all the debug things I know as writting
the value in a file and the result was strange also sometime the value showed
true and sometime false..... several hours later I did find that if you have
two ship of the same class the core code (timestep function) was working
in same time and value shared. so a value of ship1 can be read by the 2nd vessel.


I tried a workaround consisting of a flag that was initialised when the panel
is initialised (that code is unique per ship) but if you make the mistake to try to fly
with the second ship (so you initialise the panel also for this 2nd ship) the bug
was back. not satisfying so I dismissed this code.

there are old message related to this in the Orbiter's bug forum but
as the search function is broken I didn't find it.

Sharing code of same class ship is a know "behaviour??" "bug??" of Orbiter
so the solution for now is to avoid having two dgII in the same scenery
perhaps martin will fix this in the next release, if not I may find another solution
(as I write I have suddenly an idea...

Of course if you don't have two dgII in the scenery and still have a G bug with
an high framerate this is another *unknow for yet* bug !!??  let me know...

Dan