[Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
Melchior FRANZ said: > * Jim Wilson -- Wednesday 04 August 2004 20:49: > > Can you post what you have, with the model fixed and the target-offsets in, etc? > > attached > > m. That "looks" fine (at least from pitch and roll angles). It is off a little bit because the model offsets (in the Models/bo105.xml file) don't match how far the 3D model is off of 0,0,0. Also that fdm config really should be fixed, because that will throw off your rotations as well. Just make your fuselage ax,ay,az = 0 and then increase the other x and z values by .17 and .51 respectively. If all the values are updated then the FDM will operate the same, just with a different origin. Of course it is possible to just call the origin -0.17,0,-0.51 and adjust the model to that, but now you would really be getting tricky :-) Also, in the case of the c310u3a model, I actually went into ac3d and and moved the model. Then adjust the animation entries. It took a couple minutes but then at least I didn't have to worry about using the model offsets any longer. You might have noticed that the gear numbers are all screwed up too (x,y,z mixed up). I'm not sure what effect that would have other than maybe making things on the ground a little strange. It does look like the aircraft yaws a little weird when looked at exactly as you described. But I'm not really sure what the behavior should be. And of course we'd probably never notice that type of error in a non-hovering aircraft. Where should the pivot axis be when a bo105 is exactly hovering (something I have trouble doing)? What effect would centrifugal force have on that axis once movement began? In any case, it doesn't seem to rotate exactly around the nose. Sometimes it is behind the nose and sometimes it is in front of the nose. But as you say, the axis isn't located near the main rotor shaft. Best, Jim ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
[Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
* Jon S Berndt -- Wednesday 04 August 2004 23:12: > On Wed, 4 Aug 2004 23:04:16 +0200 > Melchior FRANZ <[EMAIL PROTECTED]> wrote: > > >Good idea. I'll try that. Thanks. > > Get some sleep, first !! Hehe ... OK. It won't be a lot of fun, anyway. I should trace both origin and CG in lon/lat. Have to find that yet in YASim ... Andiiieee! m. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
On Wed, 4 Aug 2004 23:04:16 +0200 Melchior FRANZ <[EMAIL PROTECTED]> wrote: Good idea. I'll try that. Thanks. Get some sleep, first !! :-) Jon ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
[Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
* Curtis L. Olson -- Wednesday 04 August 2004 22:58: > Perhaps you could export some of the FDM data (position, orientation, > etc.) into a text file and graph it. If you approached the problem > carefully, you should be able to get to the bottom of exactly what is > going on. Good idea. I'll try that. Thanks. m. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
Melchior FRANZ wrote: * Vivian Meazza -- Wednesday 04 August 2004 22:35: I think all views are fine, although view 2 is a bit disconcerting or even a bit disorientating. This seems to be a view-based issue. I'm convinced that the aircraft is rotating appoximately around it's cofg. This is *not* a view-based issue. But I'm tired of explaining how I can prove the rotation around the nose without any doubt, and independent of the view. Perhaps you could export some of the FDM data (position, orientation, etc.) into a text file and graph it. If you approached the problem carefully, you should be able to get to the bottom of exactly what is going on. Regards, Curt. -- Curtis Olsonhttp://www.flightgear.org/~curt HumanFIRST Program http://www.humanfirst.umn.edu/ FlightGear Project http://www.flightgear.org Unique text:2f585eeea02e2c79d7b1d8c4963bae2d ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
[Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
* Vivian Meazza -- Wednesday 04 August 2004 22:35: > I think all views are fine, although view 2 is a bit disconcerting or even a > bit disorientating. This seems to be a view-based issue. I'm convinced that > the aircraft is rotating appoximately around it's cofg. This is *not* a view-based issue. But I'm tired of explaining how I can prove the rotation around the nose without any doubt, and independent of the view. m. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
RE: [Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
> Melchior FRANZ wrote > Sent: 04 August 2004 19:22 > To: [EMAIL PROTECTED] > Cc: Maik Justus > Subject: [Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh! > > > * Vivian Meazza -- Wednesday 04 August 2004 19:15: > > Yes, see what you mean, but I think we went through all this with > > fixed wing aircraft. But once fully airborne, it's perfect. > Weird. Is > > it trying to rotate about a skid? > > No, same when airborne, just harder to see. You better use > the "constant-azimuth" view (6?), then try to hover a few > meters above the ground, then turn. I think all views are fine, although view 2 is a bit disconcerting or even a bit disorientating. This seems to be a view-based issue. I'm convinced that the aircraft is rotating appoximately around it's cofg. > > Perhaps not enough tail rotor authority in the hover > though? I can't > > get more than about 30 degs out-of-wind. > > Tail rotor authority seems OK (I talked about it with a real > helicopter pilot). But even if it weren't, it wouldn't > explain the rotation around the nose. Seems very much like a > rotor bug in YASim. Maybe none of the rotor forces act on the > CG, but all on the origin. This would make the 3d model look > like behaving OK when it was (wrongly) placed with the CG at > origin, but not when correctly positioned with the nose on > the origin. (Hmm, or not??) > It's nearly 30 years since I last flew one. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
[Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
* Jim Wilson -- Wednesday 04 August 2004 20:49: > Can you post what you have, with the model fixed and the target-offsets in, etc? Was already as attached to my first reply to Curt's message as diff. I'll send it again privately, just in case. :-) m. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
Melchior FRANZ said: > > > Did you put the target offset in the right place, in the right file? > > I think so, yes. (Well, I hope so, at least. :-) > Can you post what you have, with the model fixed and the target-offsets in, etc? Best, Jim ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
[Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
* Vivian Meazza -- Wednesday 04 August 2004 19:15: > Yes, see what you mean, but I think we went through all this with fixed wing > aircraft. But once fully airborne, it's perfect. Weird. Is it trying to > rotate about a skid? No, same when airborne, just harder to see. You better use the "constant-azimuth" view (6?), then try to hover a few meters above the ground, then turn. > Perhaps not enough tail rotor authority in the hover though? I can't get > more than about 30 degs out-of-wind. Tail rotor authority seems OK (I talked about it with a real helicopter pilot). But even if it weren't, it wouldn't explain the rotation around the nose. Seems very much like a rotor bug in YASim. Maybe none of the rotor forces act on the CG, but all on the origin. This would make the 3d model look like behaving OK when it was (wrongly) placed with the CG at origin, but not when correctly positioned with the nose on the origin. (Hmm, or not??) m. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
RE: [Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
Melchior FRANZ > Sent: 04 August 2004 16:05 > To: [EMAIL PROTECTED] > Subject: [Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh! > > > * Vivian Meazza -- Wednesday 04 August 2004 16:50: > > Well, I don't know. I've just flown it using a recent CVS > download. It > > looked really good to me. The point of yaw seemed to be > just about at > > the rotor hub. > > Yes, the version from CVS *seems* to behave right. But it has > the 3d model placed wrongly. Fixing the model placement (see > the attached patch in a previous message in this thread) > makes the bug obvious (and this is needed to make the skids > act correctly). Of course we could leave everything as is and > pretend that there's no bug. :-) > Yes, see what you mean, but I think we went through all this with fixed wing aircraft. But once fully airborne, it's perfect. Weird. Is it trying to rotate about a skid? Perhaps not enough tail rotor authority in the hover though? I can't get more than about 30 degs out-of-wind. Regards, Vivian ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
[Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
* Vivian Meazza -- Wednesday 04 August 2004 16:50: > Well, I don't know. I've just flown it using a recent CVS download. It > looked really good to me. The point of yaw seemed to be just about at the > rotor hub. Yes, the version from CVS *seems* to behave right. But it has the 3d model placed wrongly. Fixing the model placement (see the attached patch in a previous message in this thread) makes the bug obvious (and this is needed to make the skids act correctly). Of course we could leave everything as is and pretend that there's no bug. :-) m. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
RE: [Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
Melchior FRANZ > Sent: 04 August 2004 14:49 > To: [EMAIL PROTECTED] > Subject: [Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh! > > > * Jim Wilson -- Wednesday 04 August 2004 15:28: > > Melchior FRANZ said: > > The only difference is JSBSim has the extra feature of > specifying the > > VRP which is an offset from where 0,0,0 ends up being > specified. It > > saved changing all the entries for existing flight model > configs that > > had 0,0,0 at the firewall, etc. > > Yes, I know. For the bo105 it's the nose, though. Not > exactly, actually (see below). > > > > > It sounds like you pretty much understand what is going on, > but just > > keep in mind this FDM side is only about where the lon/lat/alt gets > > reported in FlightGear. > > Thanks, and, yes, I know. :-) > > > > > BTW I noticed that the bo105 fuselage in CVS is not > anchored exactly > > to 0,0,0 but it is offset (I think it is 0.17 forward and > 0.51 down). > > This should probably be fixed. > > I will fix it later. It has to do with the fact that Maik > "fdm-modeled" a Bo105, while I "3d-modeled" a Bo105CBS (which > is 10 inches longer). > > > > > Keep in mind that exactly straight down isn't possible, and > remember > > the camera is moving too. So yeah that "illusion" can work > from any > > direction, even with what appears to be a good fixed reference. > > OK, but this wouldn't have such a dramatic effect. > > > > > The solver places the bo105 CG as follows on my copy: > > -2.491, -0.000, 1.091 > > Sounds good. > > > > > I'm not really sure how the aircraft should turn and what relation > > rotation would have to CG. Note that your FDM thinks the > nose is 1.5 > > meters (add the offset I mentioned above) below the CG. > > I would expect it to turn pretty much around the main rotor > axis/CG, or maybe slightly forward on ground (because of a > possible forward pitch and the skid friction). But never > would I expect a rotation around the nose, which seems > physically impossible to me. > > > > > If there is a bug in YASim this would be the first I've seen of it. > > Not necessarily. Maybe the tail rotor doesn't apply the force > correctly. It could be a pure helicopter/rotor problem. > > > > > Is it possible that you never had the model positioned with > the target > > offset correct all at the same time? > > That's what I'm talking about. The 3d model was always placed > wrongly (not by me, by the way. :-) The 3d model was placed > with the main rotor axis aligned with what YASim thinks is > the nose. That's why I thing the bug was masked. > > > > > Did you put the target offset in the right place, in the right file? > > I think so, yes. (Well, I hope so, at least. :-) > Well, I don't know. I've just flown it using a recent CVS download. It looked really good to me. The point of yaw seemed to be just about at the rotor hub. Hmm ... Regards, Vivian ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
[Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
* Curtis L. Olson -- Wednesday 04 August 2004 16:37: > Could this be a bug in the helicopter specific dynamics of YAsim? Possibly. It could also be a bug in the bo105 yasim config (one that isn't obvious to me). > That is a relatively new addition and wasn't written by Andy. Can you > reproduce the problem with any aircraft? I've never seen this anywhere else. Or maybe it's only more obvious for a helicopter, because I can turn a helicopter on the spot on ground. > That might motivate Andy to take a look into the code. Yes. I'll try to find out more. If only we had another helicopter without the problem ... m. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
Melchior FRANZ wrote: * Curtis L. Olson -- Wednesday 04 August 2004 13:58: [...] Where you might be getting confused is with the view system. If (0,0,0) is the nose, then all the external views will center around the nose. [...] You wrote pretty much the same as Jim, but you obviously didn't read my response to his email, where I tried to explain why he (and now you) do not explain my problem. You both seem to have decided that I am yet another victim of the viewpoint illusion, something that I do well understand and did so before. I *did* align YASim's and the bo's origin, and I *did* adjust the view point. It doesn't help, because the bo is *really* *yawing* around the nose. (It isn't pitching around the nose, only yawing/turning!) Could this be a bug in the helicopter specific dynamics of YAsim? That is a relatively new addition and wasn't written by Andy. Can you reproduce the problem with any aircraft? That might motivate Andy to take a look into the code. Regards, Curt. How could my method create the described illusion? I'm looking at the bo from a few meters above. The bo stands on the ground with it's nose over an easily spottable point (e.g. a taxiway corner). Then I increase the collective (without lifting off), and turn the bo to the left. And the bo turns around the nose. There's no room for an illusion here. The distance from nose to ground is much too small. Try yourself: the attached patch aligns the 3d model correctly with YASim (which isn't the case now!), and sets the viewpoint to the CG. m. PS: I may still have missed something, but not what you and Jim think. :-( Index: bo105-set.xml === RCS file: /var/cvs/FlightGear-0.9/data/Aircraft/bo105/bo105-set.xml,v retrieving revision 1.9 diff -u -p -r1.9 bo105-set.xml --- bo105-set.xml 28 May 2004 09:45:02 - 1.9 +++ bo105-set.xml 4 Aug 2004 12:37:49 - @@ -44,16 +44,47 @@ Bo 105 simulation config. - + true 0.38 0.60 - -1.50 + 1.2 -8 + + + 2.5 + + + + + + 2.5 + + + + + + 2.5 + + + + + + 2.5 + + + + + + + 2.5 + + + false Index: Models/bo105.xml === RCS file: /var/cvs/FlightGear-0.9/data/Aircraft/bo105/Models/bo105.xml,v retrieving revision 1.13 diff -u -p -r1.13 bo105.xml --- Models/bo105.xml28 May 2004 09:45:02 - 1.13 +++ Models/bo105.xml4 Aug 2004 12:37:49 - @@ -3,6 +3,7 @@ bo105.ac + 2.67 -.2 ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d -- Curtis Olsonhttp://www.flightgear.org/~curt HumanFIRST Program http://www.humanfirst.umn.edu/ FlightGear Project http://www.flightgear.org Unique text:2f585eeea02e2c79d7b1d8c4963bae2d ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
[Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
* Jim Wilson -- Wednesday 04 August 2004 15:28: > Melchior FRANZ said: > The only difference is JSBSim has the extra feature of specifying the VRP > which is an offset from where 0,0,0 ends up being specified. It saved > changing all the entries for existing flight model configs that had 0,0,0 at > the firewall, etc. Yes, I know. For the bo105 it's the nose, though. Not exactly, actually (see below). > It sounds like you pretty much understand what is going on, but just keep in > mind this FDM side is only about where the lon/lat/alt gets reported in > FlightGear. Thanks, and, yes, I know. :-) > BTW I noticed that the bo105 fuselage in CVS is not anchored exactly to 0,0,0 > but it is offset (I think it is 0.17 forward and 0.51 down). This should > probably be fixed. I will fix it later. It has to do with the fact that Maik "fdm-modeled" a Bo105, while I "3d-modeled" a Bo105CBS (which is 10 inches longer). > Keep in mind that exactly straight down isn't possible, and remember the > camera is moving too. So yeah that "illusion" can work from any direction, > even with what appears to be a good fixed reference. OK, but this wouldn't have such a dramatic effect. > The solver places the bo105 CG as follows on my copy: > -2.491, -0.000, 1.091 Sounds good. > I'm not really sure how the aircraft should turn and what relation rotation > would have to CG. Note that your FDM thinks the nose is 1.5 meters (add the > offset I mentioned above) below the CG. I would expect it to turn pretty much around the main rotor axis/CG, or maybe slightly forward on ground (because of a possible forward pitch and the skid friction). But never would I expect a rotation around the nose, which seems physically impossible to me. > If there is a bug in YASim this would be the first I've seen of it. Not necessarily. Maybe the tail rotor doesn't apply the force correctly. It could be a pure helicopter/rotor problem. > Is it possible that you never had the model positioned with the target offset > correct all at the same time? That's what I'm talking about. The 3d model was always placed wrongly (not by me, by the way. :-) The 3d model was placed with the main rotor axis aligned with what YASim thinks is the nose. That's why I thing the bug was masked. > Did you put the target offset in the right place, in the right file? I think so, yes. (Well, I hope so, at least. :-) m. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
RE: [Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
Jon Berndt said: > Curt quipped: > > > If you want to use the same 3d model for multiple FDM configurations > > which choose different reference points, then that makes life a lot > > harder. I think that is what the JSBsim VRP is for. > > I had to laugh when reading these two sentences together. It sort of read ( to me ) like: > > "The JSBSim VRP is for making life a lot harder." > > Without the VRP, JSBSim would simply report the location of the CG to FlightGear. With the > VRP specified in the JSBSim config file, instead of reporting the CG, JSBSim will report > the lat/lon/alt (the position) of the VRP instead. By convention, we say that the nose of > the aircraft is what we want the VRP to represent. This can be advantageous (I think) > because everyone knows where the nose of the aircraft is. This should make 3D model > placement in the scene easy. > Ah, that I don't know. So AC_VRP is not just an offset, it is a switch as well? If it isn't specified, then you are reporting lat/lon/alt the old way (at the current center of gravity)? Best, Jim ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
Melchior FRANZ said: > I know. But still, YASim's origin (the nose) is the (unchangeable) VRP, or what > JSBSim calls "VRP", right? And the 3D model would also have to be moved such > that the nose is at YASim's 0/0/0, too. But the bo105 does only turn around the > CG if the main *rotor axis* wrongly is at YASim's *nose*. The only difference is JSBSim has the extra feature of specifying the VRP which is an offset from where 0,0,0 ends up being specified. It saved changing all the entries for existing flight model configs that had 0,0,0 at the firewall, etc. You could say the YASim VRP is unchangable, but that's partly a misstatement as there is no VRP defined in YASim. It sounds like you pretty much understand what is going on, but just keep in mind this FDM side is only about where the lon/lat/alt gets reported in FlightGear. We talk about "Nose" but that's just a convenient way for the 3D modeler to know how to position the Model so it represents where the FDM thinks the aircraft is. You may as well put the bo105 with the origin at the nose and then address any other issues after, because that is where it ought to be. BTW I noticed that the bo105 fuselage in CVS is not anchored exactly to 0,0,0 but it is offset (I think it is 0.17 forward and 0.51 down). This should probably be fixed. > Not so quick! I knew about the target offset and had considered it, but still > the model rotated around the nose: I had the bo105 parked on the ground with > the nose over a taxiway corner, then I slightly increased the collective, turned > the bo with pedal input, and looked at it from above. The bo clearly rotated > around the nose, if I had shifted the 3d model correctly in place. That was > hardly and optical illusion. > Keep in mind that exactly straight down isn't possible, and remember the camera is moving too. So yeah that "illusion" can work from any direction, even with what appears to be a good fixed reference. The solver places the bo105 CG as follows on my copy: -2.491, -0.000, 1.091 I'm not really sure how the aircraft should turn and what relation rotation would have to CG. Note that your FDM thinks the nose is 1.5 meters (add the offset I mentioned above) below the CG. If there is a bug in YASim this would be the first I've seen of it. Is it possible that you never had the model positioned with the target offset correct all at the same time? Did you put the target offset in the right place, in the right file? Best, Jim ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
[Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
* Curtis L. Olson -- Wednesday 04 August 2004 13:58: [...] > Where you might be getting confused is with the view system. If (0,0,0) > is the nose, then all the external views will center around the nose. [...] You wrote pretty much the same as Jim, but you obviously didn't read my response to his email, where I tried to explain why he (and now you) do not explain my problem. You both seem to have decided that I am yet another victim of the viewpoint illusion, something that I do well understand and did so before. I *did* align YASim's and the bo's origin, and I *did* adjust the view point. It doesn't help, because the bo is *really* *yawing* around the nose. (It isn't pitching around the nose, only yawing/turning!) How could my method create the described illusion? I'm looking at the bo from a few meters above. The bo stands on the ground with it's nose over an easily spottable point (e.g. a taxiway corner). Then I increase the collective (without lifting off), and turn the bo to the left. And the bo turns around the nose. There's no room for an illusion here. The distance from nose to ground is much too small. Try yourself: the attached patch aligns the 3d model correctly with YASim (which isn't the case now!), and sets the viewpoint to the CG. m. PS: I may still have missed something, but not what you and Jim think. :-( Index: bo105-set.xml === RCS file: /var/cvs/FlightGear-0.9/data/Aircraft/bo105/bo105-set.xml,v retrieving revision 1.9 diff -u -p -r1.9 bo105-set.xml --- bo105-set.xml 28 May 2004 09:45:02 - 1.9 +++ bo105-set.xml 4 Aug 2004 12:37:49 - @@ -44,16 +44,47 @@ Bo 105 simulation config. - + true 0.38 0.60 --1.50 +1.2 -8 + + +2.5 + + + + + +2.5 + + + + + +2.5 + + + + + +2.5 + + + + + + +2.5 + + + false Index: Models/bo105.xml === RCS file: /var/cvs/FlightGear-0.9/data/Aircraft/bo105/Models/bo105.xml,v retrieving revision 1.13 diff -u -p -r1.13 bo105.xml --- Models/bo105.xml 28 May 2004 09:45:02 - 1.13 +++ Models/bo105.xml 4 Aug 2004 12:37:49 - @@ -3,6 +3,7 @@ bo105.ac + 2.67 -.2 ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
RE: [Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
Curt quipped: > If you want to use the same 3d model for multiple FDM configurations > which choose different reference points, then that makes life a lot > harder. I think that is what the JSBsim VRP is for. I had to laugh when reading these two sentences together. It sort of read ( to me ) like: "The JSBSim VRP is for making life a lot harder." Ha! Some might argue that that is exactly what the VRP does. I wrote up an article on the VRP for the first JSBSim newsletter (on our site). [BTW, the second newsletter is 99% ready, I just need to tie up some loose ends, but conference preps have got me busier than usual.] Without the VRP, JSBSim would simply report the location of the CG to FlightGear. With the VRP specified in the JSBSim config file, instead of reporting the CG, JSBSim will report the lat/lon/alt (the position) of the VRP instead. By convention, we say that the nose of the aircraft is what we want the VRP to represent. This can be advantageous (I think) because everyone knows where the nose of the aircraft is. This should make 3D model placement in the scene easy. Note that I am always open to suggestions on a better way to do this. While I feel that it should not be so confusing to people (i.e. maybe there is a better way), the simple fact is that given the multiple FDMs that FlightGear supports, 3D models built by many different people, etc. maybe we would be silly to expect the simulation to do all the work - i.e. there is some thinking one has to do for all of this. In the end, all the FDM can do is to report lat/lon/alt. What this represents can be [re-]opened for discussion if need be. Jon ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
Curtis L. Olson wrote: Melchior, This stuff can get confusing, and I hate to reopen a big can of worms, and I didn't really pay attention to the JSBsim VRP discussion a while back. With YAsim you can pick any point as the (0,0,0) point and reference all your dimensions and positions from there. YAsim handles this just fine and has no bugs that I am aware of in this area. Most likely where you are getting confused is that YASim reports the (0,0,0) position as the model position to FG. Again, this is no big deal if as an aircraft designer, you make the (0,0,0) position of the model correspond to the same (0,0,0) position you use in the FDM configuration. Where you might be getting confused is with the view system. If (0,0,0) is the nose, then all the external views will center around the nose. The visual effect is quite compelling and makes the model *appear* to be rotating around the nose. But it's not! YAsim is handling everything correctly, the model is getting placed correctly, the model is getting rotated correctly. It only looks wrong because the view is always forced to center on the nose. When the nose pitches up, you don't see it because the view stays perfectly centered on the nose. All you see is the entire aircraft pitching down. That *looks* very wrong, but it's an optical illusion. The quickest way to fix this is to offset the view center location in the aircraft-set.xml file (see the p51 or pa28-161 config files for examples.) If you want to use the same 3d model for multiple FDM configurations which choose different reference points, then that makes life a lot harder. I think that is what the JSBsim VRF is for. Regards, Curt. Melchior FRANZ wrote: * Jim Wilson -- Wednesday 04 August 2004 02:55: Melchior FRANZ said: [...] When I noticed that the skids didn't work correctly, I found this VRP bug. I shifted the model so that the VRP would agree with the YASim origin and changed the view look-at point accordingly. This fixed the skid problem. But now it became obvious that yasim turns the bo around the origin (nose tip) rather than the CG! This was masked by the VRP bug. Actually it wasn't, and has nothing to do with VRP. VRP is just JSBSim's way of shifting the location where longitude, latitude, and altitude are reported at. It has nothing to do with YASim. This actually enables JSBSim to report location at something other than 0,0,0. I know. But still, YASim's origin (the nose) is the (unchangeable) VRP, or what JSBSim calls "VRP", right? And the 3D model would also have to be moved such that the nose is at YASim's 0/0/0, too. But the bo105 does only turn around the CG if the main *rotor axis* wrongly is at YASim's *nose*. YASim on the otherhand always reports at 0,0,0. We've often talked about standardizing with the "Nose" at 0,0,0 and reporting the location (lon,lat,alt) there. I know. I read most of these discussions (and understood some of them. :-) And now I'm out of ideas. Is it a YASim bug? A bug in YASim's rotor parts? A bug in my brain? Yes, it is a bug in your brain. Actually everyone's brain. I've added the "solution" to the Flight Gear wiki. Not so quick! I knew about the target offset and had considered it, but still the model rotated around the nose: I had the bo105 parked on the ground with the nose over a taxiway corner, then I slightly increased the collective, turned the bo with pedal input, and looked at it from above. The bo clearly rotated around the nose, if I had shifted the 3d model correctly in place. That was hardly and optical illusion. m. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d I still think this sort of thing could be straightened out by having an option to draw a three axis cursor through the the cm. Turn it on and it should become obvious what is going on. I suppose you could do it for the model origin as well. Of course, I don't know how to do it, so I can't volunteer, but it occurs to me that it might be less work than continually explaining this. Josh ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
Melchior, This stuff can get confusing, and I hate to reopen a big can of worms, and I didn't really pay attention to the JSBsim VRP discussion a while back. With YAsim you can pick any point as the (0,0,0) point and reference all your dimensions and positions from there. YAsim handles this just fine and has no bugs that I am aware of in this area. Most likely where you are getting confused is that YASim reports the (0,0,0) position as the model position to FG. Again, this is no big deal if as an aircraft designer, you make the (0,0,0) position of the model correspond to the same (0,0,0) position you use in the FDM configuration. Where you might be getting confused is with the view system. If (0,0,0) is the nose, then all the external views will center around the nose. The visual effect is quite compelling and makes the model *appear* to be rotating around the nose. But it's not! YAsim is handling everything correctly, the model is getting placed correctly, the model is getting rotated correctly. It only looks wrong because the view is always forced to center on the nose. When the nose pitches up, you don't see it because the view stays perfectly centered on the nose. All you see is the entire aircraft pitching down. That *looks* very wrong, but it's an optical illusion. The quickest way to fix this is to offset the view center location in the aircraft-set.xml file (see the p51 or pa28-161 config files for examples.) If you want to use the same 3d model for multiple FDM configurations which choose different reference points, then that makes life a lot harder. I think that is what the JSBsim VRF is for. Regards, Curt. Melchior FRANZ wrote: * Jim Wilson -- Wednesday 04 August 2004 02:55: Melchior FRANZ said: [...] When I noticed that the skids didn't work correctly, I found this VRP bug. I shifted the model so that the VRP would agree with the YASim origin and changed the view look-at point accordingly. This fixed the skid problem. But now it became obvious that yasim turns the bo around the origin (nose tip) rather than the CG! This was masked by the VRP bug. Actually it wasn't, and has nothing to do with VRP. VRP is just JSBSim's way of shifting the location where longitude, latitude, and altitude are reported at. It has nothing to do with YASim. This actually enables JSBSim to report location at something other than 0,0,0. I know. But still, YASim's origin (the nose) is the (unchangeable) VRP, or what JSBSim calls "VRP", right? And the 3D model would also have to be moved such that the nose is at YASim's 0/0/0, too. But the bo105 does only turn around the CG if the main *rotor axis* wrongly is at YASim's *nose*. YASim on the otherhand always reports at 0,0,0. We've often talked about standardizing with the "Nose" at 0,0,0 and reporting the location (lon,lat,alt) there. I know. I read most of these discussions (and understood some of them. :-) And now I'm out of ideas. Is it a YASim bug? A bug in YASim's rotor parts? A bug in my brain? Yes, it is a bug in your brain. Actually everyone's brain. I've added the "solution" to the Flight Gear wiki. Not so quick! I knew about the target offset and had considered it, but still the model rotated around the nose: I had the bo105 parked on the ground with the nose over a taxiway corner, then I slightly increased the collective, turned the bo with pedal input, and looked at it from above. The bo clearly rotated around the nose, if I had shifted the 3d model correctly in place. That was hardly and optical illusion. m. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d -- Curtis Olsonhttp://www.flightgear.org/~curt HumanFIRST Program http://www.humanfirst.umn.edu/ FlightGear Project http://www.flightgear.org Unique text:2f585eeea02e2c79d7b1d8c4963bae2d ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
[Flightgear-devel] Re: yasim + bo105 + vrp + @#%$#@ == argh!
* Jim Wilson -- Wednesday 04 August 2004 02:55: > Melchior FRANZ said: [...] > > When I noticed that the skids didn't work correctly, I found this VRP bug. > > I shifted the model so that the VRP would agree with the YASim origin > > and changed the view look-at point accordingly. This fixed the skid problem. > > But now it became obvious that yasim turns the bo around the origin (nose > > tip) rather than the CG! This was masked by the VRP bug. > > Actually it wasn't, and has nothing to do with VRP. VRP is just JSBSim's way > of shifting the location where longitude, latitude, and altitude are reported > at. It has nothing to do with YASim. This actually enables JSBSim to report > location at something other than 0,0,0. I know. But still, YASim's origin (the nose) is the (unchangeable) VRP, or what JSBSim calls "VRP", right? And the 3D model would also have to be moved such that the nose is at YASim's 0/0/0, too. But the bo105 does only turn around the CG if the main *rotor axis* wrongly is at YASim's *nose*. > YASim on the otherhand always reports at 0,0,0. We've often talked about > standardizing with the "Nose" at 0,0,0 and reporting the location > (lon,lat,alt) there. I know. I read most of these discussions (and understood some of them. :-) > > And now I'm out of ideas. Is it a YASim bug? A bug in YASim's rotor parts? > > A bug in my brain? > > Yes, it is a bug in your brain. Actually everyone's brain. I've added the > "solution" to the Flight Gear wiki. Not so quick! I knew about the target offset and had considered it, but still the model rotated around the nose: I had the bo105 parked on the ground with the nose over a taxiway corner, then I slightly increased the collective, turned the bo with pedal input, and looked at it from above. The bo clearly rotated around the nose, if I had shifted the 3d model correctly in place. That was hardly and optical illusion. m. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d