Surfacing Question??

[SiW979]

What would be the best method of filling the gap where the green surface is currently positioned to get a nice smooth transition between the two other surfaces. If you look at the relection analysis, you will see that at present, it'sm pretty shocking.

Many thanks

Si.

Best regards

Simon NX7.5.4.4 MP5 - TC 8

http://www.jcb.com

 

[MickyV007]

I don't know if it sufficient, but i used one bridge curve and swept it aside the two halfs, with the edges as guide curves.

Best regards,

Michaël.

NX7.5.4.4 + TC Unified 8.3
Win 7 64 bit

 

[cubalibre000]

Sorry Simon,
but you need NX training.
I never seen a bad modeling as your.

Thank you...

Using NX 8 and TC8.3

 

[desgur]

Try face blend with tangency control. You must create an extruded surface first in the middle. I got a relatively nice face at R145.5.

Even section surface>End-Slope-Rho is also very close. I hope it is acceptable to you.

In both cases, the surface is better than what you get in thru curve mesh.

AK
NX 7.5.2.5

 

[Toost]

Simon,
what is your intention behind this example ?
I assume that you are playing around with surfaces just for the fun .
Since the mirrored sides are at such a large angle at each other , you will always have a ridge in the center, there is no geometrical way of blending that away. If you desire a smoother transition you must rebuild the sides/ start earlier with the transition.

An old but very reliable feature that creates very high quality surfaces is the Section Surface - Section Fillet Bridge.

A through curve mesh surface is reliable in that it will always deliver, but not always create simple math surfaces, instead it will create as many patches necessary to pass the tolerance. In this case it has to create quite a few due to the matched shape.

 

[SiW979]

I was just playing with surfaces and some different types of problems that arrise. I would indeed model that very differently, but I get asked lots of questions from lots of people in my organisation and that was one.

Thanks

Si

Best regards

Simon NX7.5.4.4 MP5 - TC 8

http://www.jcb.com

 

[Walterke]

Sorry Simon,
but you need NX training.
I never seen a bad modeling as your.

Don't you GIVE NX training Simon?

NX 7.5
Teamcenter 8

 

[SiW979]

Yes I do lots of training and todate have trained over 800 delegates, but not free-form, this we still use external resource for. I can do free-form to some extent, but being good enough to use something and being good enough to train it are two completly different things. On the modelling side (sketcher, expressions, assmblies etc) I'll give anyone a run for thier money, but (as you can tell) not free-form. We don't have a massive need for it on earth moving equipment (thanks for pointing out the obvious BTW cubalibre) I have attached one of my better efforts (which I did a good while ago) to prove I'm not a total doofus

Cheers Si



Best regards

Simon NX7.5.4.4 MP5 - TC 8 www.jcb.com

 

[cubalibre000]

Simon,
I attach an example of a similar shape made in two minutes whit total different approach.
When I said you need NX training, because I didn't see in your file best practice in modeling.
For example you don't use 2D sketch, lot better then 3D sketch.
2D sketch are better in performance and maintenance.
You use surfaces feature that are not necessary...for example 'Studio surface', 'Styled Blend' and 'Bridge Curve'.
Make simple and use simple feature where it's possible.
Best practice are not only for produce parts with best performance or fews feature, but produce parts that colleagues can understand and manipulate in the feature.

Thank you...

Using NX 8 and TC8.3

 

[cowski]

For example you don't use 2D sketch, lot better then 3D sketch.
2D sketch are better in performance and maintenance.

Regarding freeform modeling, this is just plain wrong. Depending on the shape involved, using splines outside of a sketch may be preferred. One effective method is to make simple sketches (may be as simple as a single line) and use those to control the end location and tangency conditions of a studio spline. Sometimes it is easier to get rid of the sketches entirely.

http://www.nxjournaling.com

 

[cubalibre000]

Hi cowski,
I write my opinion.

You can produce surfaces in two manner :
- Profiles (sketches, edges)
- Applied surfaces (freeform tools, bends, trimming, etc..)

For 'Applied surfaces', it's important to use the right tool, in the right manner and in the right contest.
In this case, experience in the CAD knowledge and in this case in NX, it's fundamental.
An example is, don't use potent surfaces tools, because are used in automotive, if you don't need perfect G2 or G3 curvature.
Lot of people doesn't know the mathematical rule of G2 curvature and aren't able to verify the curvature.
Apply aesthetic blend, you have to verify at the and the result with special tool, not only use the command as is.

For surfaces created with profiles, the rule to use simple profile doesn't means split profile into single entities and 2D entities created with 3D tools.

2D sketches are more robust and predictable then 3D sketches.
Again, why split consecutive 2D entities that are tangent into more entities and use them to create surfaces that need to be in G2 curvature ?
Poor profile, produce poor surfaces.
Not G2 connected profiles doesn't produce G2 connected surfaces.

If you look at the Simon's example, connected 2D entities, split as single 3D feature is a no sense modeling procedure.

Best practice in modeling or in surfaces is not only a philosophy, but it's how a designer produce a good project.
Limit the number of feature is not only important for me that I'm modeling the part, but it's important for all.
Simple, predictable and light parts, it's one of the main goals of the designer.

Thank you...

Using NX 8 and TC8.3

 

[cowski]

it's important to use the right tool, in the right manner and in the right contest.

I agree that knowing the required/desired output is primary and you must use an appropriate tool to get there.

the rule to use simple profile doesn't means split profile into single entities and 2D entities created with 3D tools.

I know of no rules set in stone, especially when it comes to freeform surfacing. As long as you know what you are after and understand the limitations/implications of your tools and/or methodology, the sky is the limit. The simple 'trick' I referred to in my previous post doesn't necessarily apply to Simon's model, but rather is an example of using the freedom of a spline outside of a sketch that is still partially constrained by one or more sketches. It is not a rule to be used all the time, neither is it a method to be ruled out entirely.

2D sketches are more robust and predictable then 3D sketches.

This is a blanket statement that I disagree with. I use both sketches and 3D geometry (I wouldn't call them 3D sketches), and there are pros and cons of both.

Poor profile, produce poor surfaces.

Completely agree. The garbage in/garbage out principle really shows up in freeform surfacing. Do you need G2 surfaces? Then the input curves need to be G2 or better.

Limit the number of feature is not only important for me that I'm modeling the part, but it's important for all.

I've had the privilege to untangle a lot of models that I suspect were created by fans or relatives of Rube Goldberg. However, silly and unnecessary features aside, I value clarity over minimum feature count.

Simple, predictable and light parts, it's one of the main goals of the designer.

This may not be true for you, but ultimately I get paid because the company I work for sells physical products. I like working with simple and predictable parts, but in reality only me and a handful of coworkers really care about how the file was built. As long as we can export a valid solid, I don't mind if old Rube slips a file through once in a while.


For those interested in surfacing, I would recommend searching this forum on the keyword surface or surfacing with handle hudson888, he has posted some top-notch work. Also, there are posts by Tim Flater (handle: NXWheelguy/XWheelguy) that have good advice. You'll find golden nuggets from others also, but I've personally learned the most from these two.

http://www.nxjournaling.com