Ideas for controlling channel split

[Karlis]

Hi all, I have stumbled across annoying problem:
What I have modeled is a prototype of an intake port for motorcycle (some of you may have noticed it's something I do a lot). It is basically a smooth channel that splits in two somewhere down the middle, but unfortunately it is not symmetric since it has a slight turn to the left. The channels are created by using two individual mesh surfaces trimmed to each other. My problem is that I cannot figure a way to control the edge that defines the split. It comes out noisy and repulsive to the eye. What I have is a vertical surface down the middle that indicates where I want my split edge to be.
I have attached three files in a zip archive (engineering.com won't let me upload three individual files to one post):
1) A picture showing the nature of the problem
2) NX 10 part file containing my "problem", but most of you probably haven't upgraded to NX 10, so
3) STEP file containing curves that I used to "mesh surface" the channels, and the vertical surface that should define the split edge

If anyone would be willing to donate their free time to help me with some ideas, that would be well appreciated, as always

 

[NutAce]

Then why not trim both mesh surfaces to the trim surface and sew them afterwards?

Ronald van den Broek
Senior Application Engineer
Winterthur Gas & Diesel Ltd
NX8.5.3 / TC9.1.2
HPZ420 Intel(R) Xeon(R) CPU E5-1620 0 @ 3.60GHz, 32 Gb Win7 64B
Nvidea Quadro4000 2048MB DDR5

HP Zbook15
Intel(R) Core(TM) i7-4800MQ
CPU @ 2.70 GHz Win7 64b
Nvidia K1100M 2048 MB DDR5

 

[Karlis]

That is an option. But:
1) Will the sew actually modify my input surfaces to make them fit together? Or will it simply consider the gap as a tolerant edge?
2) Will I be able to run machining on that model if the sew tolerance is as large as needed to sew the sheets?
I remember John Baker explaining once, that sew does not actually modify the input surfaces, it simply considers the gaps as tolerant edges so the software ignores them.

 

[NutAce]

If you trim/extend your original surface to the trim surface there should be no gaps.
The sew will have no problem whatsoever in combining them afterwards. Even with a very tiny tollerance factor

Ronald van den Broek
Senior Application Engineer
Winterthur Gas & Diesel Ltd
NX8.5.3 / TC9.1.2
HPZ420 Intel(R) Xeon(R) CPU E5-1620 0 @ 3.60GHz, 32 Gb Win7 64B
Nvidea Quadro4000 2048MB DDR5

HP Zbook15
Intel(R) Core(TM) i7-4800MQ
CPU @ 2.70 GHz Win7 64b
Nvidia K1100M 2048 MB DDR5

 

[kr7530]

Is a different construction technique possible? The attachment would need fillets but there would be no transition issues

 

[mmauldin]

Wrong file earlier. I wouldn't try to make them match. I would leave a gap and use an intermediate surface to connect.

File attached.

 

[JohnRBaker]

Adding a Blend along the seam of your current model, as I've done in the attached part file below, will 'soften' any apparent discontinuity.

John R. Baker, P.E.
Product 'Evangelist'
Product Engineering Software
Siemens PLM Software Inc.
Digital Factory
Cypress, CA
Siemens PLM:
UG/NX Museum:

To an Engineer, the glass is twice as big as it needs to be.

 

[Toost]

1) Will the sew actually modify my input surfaces to make them fit together? Or will it simply consider the gap as a tolerant edge?

It will not modify the shape of the surface, the tolerant edge will "contain" the gap.
You can actually see the gap, but only if you : If you set the display to "Static Wireframe" and then turn on a quite dense grid ( Edit - Object Display - select all sheets - Wireframe display ( in the menu) u=20 V=20 ) you can then see the gridlines ( which are a display object only and has nothing to do with the accuracy of the surface) end up in space next to the edge.


2) Will I be able to run machining on that model if the sew tolerance is as large as needed to sew the sheets?

Yes, since NX is built on this ability ,- NX cam is ready to handle the situation.
The gap between surfaces might be visible in the milled result depending on the situation.

Regards,
Tomas