2005-november.txt

acis说明文档

> fundamental terms.  Your coordinate axes, if normalized, are the 
> components of the matrix m that forms the rotation of the transf.  The 
> translation is the translation vector argument.  If you have issues with 
> your rotations, try transposing.
> 
> Arlo Ames
> [e-mail]
> 
> -----Original Message-----
> From: [e-mail] 
> [_mailto:[e-mail]_] On Behalf Of Stefan 
> Schmidt
> Sent: Thursday, November 03, 2005 7:14 AM
> To: [e-mail]
> Subject: [Acis-alliance] Transforming between Coord Sys
> 
> Hi,
> 
> I have one coordsys1 defined by:
> 
> origin = (0,0,0) and axis (1,0,0), (0,1,0), (0,0,1)
> 
> And I have coordsys2 defined by:
> 
> origin = (0,0,0) and three different orth. axis.
> 
> How do I transform the geometry most effectively between both system in 
> ACIS?
> 
> So far I tried (unsuccessfully) two diff. approaches:
> 
> 1) Define and set Work Coord Sys:
> api_wcs_create(SPAposition(0,0,0), xpt, ypt, new_wcs); 
> api_wcs_set_active(new_wcs);
> 
> 2) Define and apply Rotational Transform:
> transf_decompose_data transfData;
> transfData.rotatex = AngleXAxis;
> transfData.rotatey = AngleYAxis;
> transfData.rotatez = AngleZAxis;
> // etc...
> trans.compose(transfData,TRUE);
> api_transform_entity(pSatEntity, trans); 
> api_change_body_trans(pSatEntity,NULL);
> 
> Does this make sense? And if so how do I find the rotation angles 
> (AngleXAxis,AngleYAxis, AngleZAxis) from the coordsys def above?
> 
> Any help much welcome. Thanks.
> 
> -- 
> Stefan Schmidt
> GiveMePower/Felix Support and Development
> [phone]
> 
> _______________________________________________
> Acis-alliance mailing list
> [e-mail]
> _http://lists.cs.columbia.edu/mailman/listinfo/acis-alliance_
> 
> 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Acis-alliance mailing list
> [e-mail]
> http://lists.cs.columbia.edu/mailman/listinfo/acis-alliance

From schmiddie78 at gmail.com  Thu Nov  3 12:28:35 2005
From: schmiddie78 at gmail.com (Stefan Schmidt)
Date: Thu Nov  3 12:29:01 2005
Subject: [Acis-alliance] Transforming between Coord Sys
In-Reply-To: <[e-mail]>
References: <[e-mail]>
Message-ID: <[e-mail]>

 From transf.hxx:
/**
  * @nodoc
  */
DECL_BASE SPAtransf restore_transf( SPAmatrix const &, SPAvector const 
&, double,
	logical, logical, logical);

No luck here...

Ames, Arlo wrote:
> Can't remember which each is (check the header files), but they're the
> flags saying whether the matrix includes scaling, rotation, translation.
> ACIS transfs only invoke matrix operations for cases where the matrix
> contains interesting information.  If no scaling, for instance, then a
> bunch of multiplies are avoided.
> Arlo 
> 
> -----Original Message-----
> From: Stefan Schmidt [mailto:[e-mail]] 
> Sent: Thursday, November 03, 2005 9:56 AM
> To: Ames, Arlo
> Cc: [e-mail]
> Subject: Re: [Acis-alliance] Transforming between Coord Sys
> 
> Found it, what are the three flags for???
> 
> Code:
> SPAmatrix mat(xpt,ypt,normal);
> SPAtransf trans = restore_transf(mat, SPAvector(0,0,0), 1.0,logical,
> logical, logical);
> 
> Ames, Arlo wrote:
> 
>>I have always liked the undocumented function restore_transf to 
>>accomplish this kind of thing.   Spatial might have a good official 
>>answer, but I haven't found anything that takes the place of 
>>restore_transf.  Call Spatial tech support to see what alternatives 
>>they might have to restore_transf.
>>
>>restore_transf lets you set every individual item of the transform in 
>>fundamental terms.  Your coordinate axes, if normalized, are the 
>>components of the matrix m that forms the rotation of the transf.  The
> 
> 
>>translation is the translation vector argument.  If you have issues 
>>with your rotations, try transposing.
>>
>>Arlo Ames
>>[e-mail]
>>
>>-----Original Message-----
>>From: [e-mail]
>>[_mailto:[e-mail]_] On Behalf Of Stefan 
>>Schmidt
>>Sent: Thursday, November 03, 2005 7:14 AM
>>To: [e-mail]
>>Subject: [Acis-alliance] Transforming between Coord Sys
>>
>>Hi,
>>
>>I have one coordsys1 defined by:
>>
>>origin = (0,0,0) and axis (1,0,0), (0,1,0), (0,0,1)
>>
>>And I have coordsys2 defined by:
>>
>>origin = (0,0,0) and three different orth. axis.
>>
>>How do I transform the geometry most effectively between both system 
>>in ACIS?
>>
>>So far I tried (unsuccessfully) two diff. approaches:
>>
>>1) Define and set Work Coord Sys:
>>api_wcs_create(SPAposition(0,0,0), xpt, ypt, new_wcs); 
>>api_wcs_set_active(new_wcs);
>>
>>2) Define and apply Rotational Transform:
>>transf_decompose_data transfData;
>>transfData.rotatex = AngleXAxis;
>>transfData.rotatey = AngleYAxis;
>>transfData.rotatez = AngleZAxis;
>>// etc...
>>trans.compose(transfData,TRUE);
>>api_transform_entity(pSatEntity, trans); 
>>api_change_body_trans(pSatEntity,NULL);
>>
>>Does this make sense? And if so how do I find the rotation angles 
>>(AngleXAxis,AngleYAxis, AngleZAxis) from the coordsys def above?
>>
>>Any help much welcome. Thanks.
>>
>>--
>>Stefan Schmidt
>>GiveMePower/Felix Support and Development
>>[phone]
>>
>>_______________________________________________
>>Acis-alliance mailing list
>>[e-mail]
>>_http://lists.cs.columbia.edu/mailman/listinfo/acis-alliance_
>>
>>
>>
>>
>>----------------------------------------------------------------------
>>--
>>
>>_______________________________________________
>>Acis-alliance mailing list
>>[e-mail]
>>http://lists.cs.columbia.edu/mailman/listinfo/acis-alliance
> 
> 
> 
> 
> 

From schmiddie78 at gmail.com  Fri Nov  4 05:30:21 2005
From: schmiddie78 at gmail.com (Stefan Schmidt)
Date: Fri Nov  4 05:30:50 2005
Subject: [Acis-alliance] Transforming between Coord Sys
In-Reply-To: <[e-mail]>
References: <[e-mail]>
Message-ID: <[e-mail]>

Sorry to bother you again. I tried false, true, true but that went 
completely off. At the moment I do:

SPAmatrix mat(xpt,ypt,normal);
SPAtransf trans = restore_transf(mat, SPAvector(0,0,0), 1.0,false, 
false,false);
outcome	Result = api_transform_entity(pSatEntity, trans);

... which almost works correctly. But instead of getting the correct 
values (x,y,z) after the transformation I get (-x,-y,z). Any idea why or 
which parameter I need to adjust? Thanks again.

Ames, Arlo wrote:
> Sorry.  I looked in the code I have available at this time and can't
> find a complete definition.  Most of the calls I have are
> false,true,true.
> 
> Try various settings and look at the transf structure in a debugger.
> That should cover it.
> 
> I have independent verification that this is the only known way of
> constructing a transf from scratch, which is really what you want to do
> here.
> 
> If this fails, you could always look up euler angle decomposition,
> perform it on your coordinate system matrix (by hand, not with ACIS) and
> then you would have three rotations about global axes you can use to
> construct the matrix.  My way is easier.
> 
> Good luck.
> 
> Arlo Ames
> [e-mail] 
> 
> -----Original Message-----
> From: Stefan Schmidt [mailto:[e-mail]] 
> Sent: Thursday, November 03, 2005 10:29 AM
> To: Ames, Arlo
> Cc: [e-mail]
> Subject: Re: [Acis-alliance] Transforming between Coord Sys
> 
>  From transf.hxx:
> /**
>   * @nodoc
>   */
> DECL_BASE SPAtransf restore_transf( SPAmatrix const &, SPAvector const
> &, double,
> 	logical, logical, logical);
> 
> No luck here...
> 
> Ames, Arlo wrote:
> 
>>Can't remember which each is (check the header files), but they're the
> 
> 
>>flags saying whether the matrix includes scaling, rotation,
> 
> translation.
> 
>>ACIS transfs only invoke matrix operations for cases where the matrix 
>>contains interesting information.  If no scaling, for instance, then a
> 
> 
>>bunch of multiplies are avoided.
>>Arlo
>>
>>-----Original Message-----
>>From: Stefan Schmidt [mailto:[e-mail]]
>>Sent: Thursday, November 03, 2005 9:56 AM
>>To: Ames, Arlo
>>Cc: [e-mail]
>>Subject: Re: [Acis-alliance] Transforming between Coord Sys
>>
>>Found it, what are the three flags for???
>>
>>Code:
>>SPAmatrix mat(xpt,ypt,normal);
>>SPAtransf trans = restore_transf(mat, SPAvector(0,0,0), 1.0,logical, 
>>logical, logical);
>>
>>Ames, Arlo wrote:
>>
>>
>>>I have always liked the undocumented function restore_transf to 
>>>accomplish this kind of thing.   Spatial might have a good official 
>>>answer, but I haven't found anything that takes the place of 
>>>restore_transf.  Call Spatial tech support to see what alternatives 
>>>they might have to restore_transf.
>>>
>>>restore_transf lets you set every individual item of the transform in 
>>>fundamental terms.  Your coordinate axes, if normalized, are the 
>>>components of the matrix m that forms the rotation of the transf.  The
>>
>>
>>>translation is the translation vector argument.  If you have issues 
>>>with your rotations, try transposing.
>>>
>>>Arlo Ames
>>>[e-mail]
>>>
>>>-----Original Message-----
>>>From: [e-mail]
>>>[_mailto:[e-mail]_] On Behalf Of Stefan 
>>>Schmidt
>>>Sent: Thursday, November 03, 2005 7:14 AM
>>>To: [e-mail]
>>>Subject: [Acis-alliance] Transforming between Coord Sys
>>>
>>>Hi,
>>>
>>>I have one coordsys1 defined by:
>>>
>>>origin = (0,0,0) and axis (1,0,0), (0,1,0), (0,0,1)
>>>
>>>And I have coordsys2 defined by:
>>>
>>>origin = (0,0,0) and three different orth. axis.
>>>
>>>How do I transform the geometry most effectively between both system 
>>>in ACIS?
>>>
>>>So far I tried (unsuccessfully) two diff. approaches:
>>>
>>>1) Define and set Work Coord Sys:
>>>api_wcs_create(SPAposition(0,0,0), xpt, ypt, new_wcs); 
>>>api_wcs_set_active(new_wcs);
>>>
>>>2) Define and apply Rotational Transform:
>>>transf_decompose_data transfData;
>>>transfData.rotatex = AngleXAxis;
>>>transfData.rotatey = AngleYAxis;
>>>transfData.rotatez = AngleZAxis;
>>>// etc...
>>>trans.compose(transfData,TRUE);
>>>api_transform_entity(pSatEntity, trans); 
>>>api_change_body_trans(pSatEntity,NULL);
>>>
>>>Does this make sense? And if so how do I find the rotation angles 
>>>(AngleXAxis,AngleYAxis, AngleZAxis) from the coordsys def above?
>>>
>>>Any help much welcome. Thanks.
>>>
>>>--
>>>Stefan Schmidt
>>>GiveMePower/Felix Support and Development
>>>[phone]
>>>
>>>_______________________________________________
>>>Acis-alliance mailing list
>>>[e-mail]
>>>_http://lists.cs.columbia.edu/mailman/listinfo/acis-alliance_
>>>
>>>
>>>
>>>
>>>----------------------------------------------------------------------
>>>--
>>>
>>>_______________________________________________
>>>Acis-alliance mailing list
>>>[e-mail]
>>>http://lists.cs.columbia.edu/mailman/listinfo/acis-alliance
>>
>>
>>
>>
>>
> 
> 
> 
> 

From shiqiao.li at asu.edu  Fri Nov  4 20:31:52 2005
From: shiqiao.li at asu.edu (Shiqiao Li)
Date: Fri Nov  4 20:32:28 2005
Subject: [Acis-alliance] Automatic removal of fillets and chamfers
References: <[e-mail]>
	<[e-mail]>
Message-ID: <001801c5e1a8$bf6638f0$452aa995@DALPC13>

Hi, there,

I wonder if there is any API function available in ACIS that can be used to 
remove fillets and chamfers on a solid model. BTW, I'm using ACIS 7.0. Your 
input is appreciated.

Li 

From schmiddie78 at gmail.com  Sat Nov  5 12:16:15 2005
From: schmiddie78 at gmail.com (Stefan Schmidt)
Date: Sat Nov  5 12:16:36 2005
Subject: [Acis-alliance] Transforming between Coord Sys
In-Reply-To: <[e-mail]>
References: <[e-mail]>
Message-ID: <[e-mail]>

Yes, it's working now. The problem was that one of my base vectors 
defining the rotation matrix was orientated the wrong way. Thanks again 
for your help.

Stefan

Ames, Arlo wrote:
> OOPS!  NOT false,true,true, but true, false, false
> 
> Again, I strongly suggest you look at your transf object in a debugger.
> You can make sure that all the transformations are correct.
> 
> If you get incorrect signs but correct magnitudes for your rotations,
> you might try transposing the information you feed to the matrix
> constructor.  Robotics and CAD have different conventions for how
> transforms are written down, and they're transposes of one another.
> I've seen enough of transpose errors that it's the first place I look.
> Arlo
> 
> -----Original Message-----
> From: Stefan Schmidt [mailto:[e-mail]] 
> Sent: Friday, November 04, 2005 3:30 AM
> To: Ames, Arlo
> Cc: [e-mail]
> Subject: Re: [Acis-alliance] Transforming between Coord Sys
> 
> Sorry to bother you again. I tried false, true, true but that went
> completely off. At the moment I do:
> 
> SPAmatrix mat(xpt,ypt,normal);
> SPAtransf trans = restore_transf(mat, SPAvector(0,0,0), 1.0,false,
> false,false);
> outcome	Result = api_transform_entity(pSatEntity, trans);
> 
> ... which almost works correctly. But instead of getting the correct
> values (x,y,z) after the transformation I get (-x,-y,z). Any idea why or
> which parameter I need to adjust? Thanks again.
> 
> Ames, Arlo wrote:
> 
>>Sorry.  I looked in the code I have available at this time and can't 
>>find a complete definition.  Most of the calls I have are 
>>false,true,true.
>>
>>Try various settings and look at the transf structure in a debugger.
>>That should cover it.
>>
>>I have independent verification that this is the only known way of 
>>constructing a transf from scratch, which is really what you want to 
>>do here.
>>
>>If this fails, you could always look up euler angle decomposition, 
>>perform it on your coordinate system matrix (by hand, not with ACIS) 
>>and then you would have three rotations about global axes you can use 
>>to construct the matrix.  My way is easier.
>>
>>Good luck.
>>
>>Arlo Ames
>>[e-mail]