From: woodyst <wo...@gm...> - 2007-07-04 22:01:26
|
2007/7/3, Roy Vegard Ovesen <roy...@ha...>: > On Thursday 28 June 2007 19:50, woodyst wrote: > > With my yoke stopped, with js_demo i see a lot of lines as: > > | 0000 -0.0 +0.1 +1.0 -1.0 -1.0 +0.0 +0.0 . | > > > > all of them without changes. > > Since it only shows one decimal digit, a change of 0.002 would not be shown in > js_demo. Yes, it was the problem. I changed js_demo.cxx so it puts 3 decimals and then I can see an ocasional movement of 0,010 some times. I apologize if I've confused someone with my previous posts. > > > > > So autopilot may change axis values (because my yoke is not noisy) and > > this difference (position 0 of yoke and autopilot setting) may be greater > > than a.tolerance, so it vibrates (I think, correct me if it is not > > correct). > > I think you have misunderstood how input.cxx works. It does not look at the > difference between the joystick position and the autopilot stick position. It > looks at the joystick position from the previous sample and if the joystick > hasn't moved more than tolerance, then its new position is not applied. > > You could hold the joystick in the full right position, far away from the > position that the autopilot wants the stick to be in. But if you can hold it > still there then input.cxx will not update the position and the joystick will > not fight with the autopilot. It is correct. > > > > > No, it is not noisy. I have tested it with utils and found that my yoke > > > > is very quiet. I think my previous afirmation may be correct. > > > > > > Very quiet might not be quiet enough. If the noise is more than the > > > tolerance value hardcoded into input.cxx (0.002) then you will see what > > > you are seeing. > > > > When I do not touch my yoke it does not pass any event. So there may be > > another problem with it. Please probe it by your own keeping your > > joystick stopped > > (test it with js_demo or something similar). You will see what is my > > problem. > > I have actually seen the problem. My joystick (Thrustmaster(r) Top GunTM Fox 2 > ProTM USB) is actually quiet enough that when it sits on my table untouched > in the center position, it does not fight with the autopilot. If I bump the > table, jump up and down on the floor, listen to loud crappy music, etc. then > it will vibrate, and it will fight with the autopilot. > > _I_ am still convinced that your joystick is more noisy than the tolerance > limit in input.cxx. I have a few suggestions you could try: > > - If the joystick driver has a dead zone option try that. I have put it in the fgfs file that describes my yoke. I have calculated the ideal dead zone for eliminating noise. It is about 0.01 (the deviation I saw in previous tests. I have used 0.01 because it is the max deviation from center I have observed with yoke centered. > - Try setting the tolerance limit in input.cxx to a higher value. This is a problem, because then it only reacts to my movements every +-tolerance increment, so I lose sensitivity (and this is a very good yoke). > - Remove the noise from your joystick: clean or replace the pots. The yoke is almost new. I think noise may be caused by the internal design of the yoke plus the vibration of the table, that it is very weak. I think the dead zone in the xml is the better solution for me, because I maintain the sensitivity of the device. Thank you very much for your patience. Sorry for the large discussion, but I was desperate and autopilot was unusable in my machine. Nevertheless I think this thread has contributed with some very interesting new ideas for the AP that I would like to improve. If we get any common point of view I would be grateful and I would help in implementing some changes. Best Regards. Woodyst. > > > -- > Roy Vegard Ovesen > > ------------------------------------------------------------------------- > This SF.net email is sponsored by DB2 Express > Download DB2 Express C - the FREE version of DB2 express and take > control of your XML. No limits. Just data. Click to get it now. > http://sourceforge.net/powerbar/db2/ > _______________________________________________ > Flightgear-devel mailing list > Fli...@li... > https://lists.sourceforge.net/lists/listinfo/flightgear-devel > -- Woodyst. |