class A surface

[s_praboo]

can anybody tell me what class A and class B surface in 3D modeling means and how it can be achieved in pro/E

 

[dlongmi]

It's easier to start with the definition of a Class "B" or simply just "B" surface. A very easy answer is: any surface that is not an appearance surface is a "B" surface. For example, the back side of a car fender, or the inside of an enclosure.


You will findwith yourquestion you will probably get all types of answers. But I will give you the definition that is most widely held as true. The general definition is:Any surface is an "A" surface that is physically visible to a consumer and is developed with appearance surface rules.It can be a painted, plated, clear or textured surface. It usually implies a rigid materialhowever, car seats, headliners and soft trimare also "A" surfaces. Where this subject takes an ugly turn is when people try to decribe "how" and "in what manner" an "A" surface is achieved.


Body design is where the term "A" surface is derived. It is as much art as it is science. It is not just creating G2 continuity or dihedral angles less than 0.5 across boundaries ofsurfaces or making something look cool and smooth. You will hear this from many people. It's just not that simple. "A" surface development is a holistic discipline and methodology of creating aesthetically pleasing and "theoretically perfect" surfaces and surface transitions. Body designers try to create G2 continuity in the 8-10 decimal place to ensure the starting point of the productis theoretically perfect or "sweet".Sweet is a term usedby Body Designers andIndustrial Designers to describe the "perfect" surface or faired curve."A" surface development is an aesthetic endeavor and requires visible light reflections, or highlights, to be manipulated on the surfaces to create a smooth appearance. You will hear the terms G-1, G0, G1, and G2 used in Body design to describe discontinuous, positional, tangential and curvature conditions respectively.


The "science" part of the equation requires mathematical precision that each body designer must ensure. For example, car manufacturers require a certain decimal precision of G2 continuity across the centerline of the car to be considered "symmetrical" and smooth enough for manufacturing. If not you would "read" the dip, ridge or wavy line right down the car's hood and roof. The precision also applies to boundary conditions. There are accuracy rules as well that the surfaces and transitions must achieve. Such as a minimum curvature deviation allowance to avoid kinks and buckling.


The "art" part of the equation is a Body Designers eye. Industrial and Body Designers have honed their sense of proportion and balance whereas Engineers and CAD Designers have not. I say this as a generalization. There are stars and hacks on both ends of the spectrum in any given discipline...Design is no exception. In other words, it takes a certain skill to design a good looking product that also has Class "A" surfaces. Again, just because it looks cool does not mean it has one mathematically true Class "A" surface on the part.


Now for the Pro/E issue. In theory, AGAIN IN THEORY, Pro/E cannot create a true "A" surface. I did indicate this is where the subject gets ugly. "A" surface development usually requires software manipulation ability that Pro/E simply does not have. That's partly why PTC bought ICEM/Surf, sucked it dry and have incorprated some of the cool stuff to ISDX.However ISDX is basically CDRS with some goodies added and is still not there IMO. Pro/E on the other hand can create G2 across some boundaries and can be tricked to and finessed to get Class "A" surfaces. But it is extremely difficult to do. Check out www.neutechnet.com andlook at the John Deere Telehandler and MackTruckproducts. These were developed in Pro/E (Pro/Surface)without ISDX...Pro/Surface only. They were verified Class "A" with secondary analysis.I wouldn't recommend anyone attempting is as it was "PAINFUL".

Pretty...but painful.


ISDX users canget closer easier but I still do not consider ISDX a true Class A surface tool because thereally fine tuning abilityis simply not there. If anyone reading this post uses ICEM/Surf they know what I am referring to. For example,having to the ability to manipulate highlights and centerline curvature is where "real" Class "A" surfacing is acheived.Many "experts" will disagree with me butthat's the great thing about open forum. Disagree away...


I really look forward to this thread.

Edited by: dlongmi

 

[s_praboo]

THANX A LO

T I REALLY HAD WIDE KNOWLEDGE ABOUT WHAT CLASS A SURFACE THOUGH I AM INTHE FIELD OF AUTOMOTIVE BODY DESIGN I WILL DEFNITELY LOOK INTO THAT WHATEVER I CREATE IS A CLASS A TO THE POSSIBLE EXTENT





THANX

 

[miked]

Regardless of the app or module its all about dihedral angle. Keep it less than .5


Mike

 

[james.lynch]

sorry for the newb question but, what exactly is the dihedral angle?

 

[miked]

James, I have created an image to help explain it. Send me your email address and Ill attach the file. It needs to be at least 300k for clarity so I dont think I can post it here.


You will also hear talk of g2 continuity for surface connections. There are three different kinds ie. G0, G1, G2 and I am also hearing talk of G4.


G0 = surfaces share a common boundary also referred to as "positional"


G1 = surfaces sharing a common boundary but they are also tangent at each end of the surface but not across its boundary.


G2 = surfaces tangent along entire connection and I think the tangency carries up two knots into the surface.


G3 = like G2 only tangency propagates further into the surface away from the connection


Mike
Edited by: miked

 

[dacct]

miked - you can possibly try to post the image on imageshack (www.imageshack.us) and include the link in here.

 

[james.lynch]

"G2 = surfaces tangent along entire connection and I think the tangency carries up two knots into the surface"


Mike, What is a "knot" or how is it defined?

 

[jeff4136]

"....G1 = surfaces sharing a common boundary but they are also tangent at each end of the surface but not across its boundary. ...."

Not true. G1 indicates that the surfaces are tangent anywhere along the intersection (within system tolerance).

G2 indicates a curvature continous intersection; e.g. the curvature of both surfaces is equal. Note that this is an "instantaneous" value; the curvatures need match only at the intersection to qualify but may be different some minute distance from it (high rate of change). This is the reason that G2 doesn't necessarily translate to "class A". Intersections can be mathematically G2 and ugly.

G3 indicates that the intersection is not only curvature continuous, but the rates of change are also equal.

A curvature graph of a G2 intersection will, itself, be G0 and for a G3 intersection G1 (see

http://news2.mcneel.com/scripts/dnewsweb.exe?cmd=article&gro up=rhino&item=73174&utag=

for a graphic).

Think I got that right .... 8~)

 

[jeff4136]

"...What is a "knot" or how is it defined?..."

http://www.rhino3d.com/nurbs.htm has a pretty fair explanation.

Continuity conditions (G0, G1, G2, etc.) are defined by control vertices or points, which are not the same thing as knots.

 

[james.lynch]

There's a space in the link..


[url]http://news2.mcneel.com/scripts/dnewsweb.exe?cmd=article&amp ; ;group=rhino&item=73174&utag [/url]=


Thanks Jeff, that clears up a lot!
Edited by: james.lynch

 

[miked]

Jeff is right. It isnt knots but control points, my brain is not working so well today.


Sorry for the confusion.

 

[dacct]

Alias also has some explanation (w/ visuals) at:


[url]http://www.alias.com/eng/support/studiotools/documentation/U sing/About31.html[/url]


and


[url]http://www.alias.com/eng/support/studiotools/documentation/L earningSurf/Foundations3_CreatingAndFittingCurves3.html [/url]