techniques in modeling part in pics

[TWINTURBOTOM]

Bascially I am creating a surface that is to be controlled by splines. As you see I have six splines, this will increase to 8. The suface begins life as a circle, with respect to the "top datum" it is placed on, after moving along the splines it should end as a rectangle....


How woud you guys model this...keep in mind I would like the controllability of splines


The attached pics are views of a variable section sweep I attempted but failed in creatingwhat I intend on....this is a vss with the section normal to the "top" datum...which is why it fails where it fails...when I change the option to "normal to trajectory" the same "effect" occurs...i.e. the section does not lie upon the ends of all of the trajectories (because at their tips they are all not fully parallel)......


I guess i'd like to be able to pivot the normal direction of the sketch at a specified point along the vss trajectory....thats not possibly though is it?


I was thinking a swept surface blend (pivot option) and aligning the points on my sketch (spline points) to the trajectories so that I would have the controllability of the curves....it wasn't comming out properlly...although I only spend 30 seconds on it...


any ideas, tips, suggestions are always appreciated....thank you!!!

last one is normal to trajectory....I tried using the t= to fix the problem but it didn't this is close....
Edited by: TWINTURBOTOM

 

[vlad1979]

I think you could try with swept blend with select sketch sections / normal to origin traj. You have the splines, sketch sections at the end of yoursplines then use them to create the swept blent.

 

[jeff4136]

All you really have to do is add a curve to control
the section normal, but throwing a bunch of curves
at a shape is just so ... low end.

View attachment 29752007-01-04_014323_prt0001_wf2_.prt.zip

 

[TWINTURBOTOM]

i guess i'll go that route......low end on multiple curves eh?...the reason I need multiples is because I need to control all corners at different times....i.e. on yours the cirlce moves into a rectangle...i need to do this but at some points leave one side rectangular and the other side round.....your way will produce the generic form of what I am looking for; but will not have the control i need.....control being the ability to manipulate the corners of the cross section as it moves along the trajectory....Thank you for your input!!!


thanks for the part file....I'll look at it tonight (when i'm on a wf machine and not this 2001 dinosour!)...
Edited by: TWINTURBOTOM

 

[jeff4136]

Then you might try adding the traj to control normal direction or maybe Vlad's swept blend suggestion (I might prefer that but don't know enough about the driving criteria). You might also consider creating local details as seperate features.


"Low end" was a tongue in cheek jab. Adding curves (or more curve 'fit points') isthe only optionfor some user / system combinations. It's sometimes the answer but more often not the ideal solution.No one size fits all answers to be had but, in generally, overcontrolled surfaces are unhappy, seldom'sweet'surfaces.

 

[TWINTURBOTOM]

good point on the unhappy seldom sweet surfaces. I understand your point now...I will look into your technique vlad that does soundexactly like what I am looking for!!I will also check out the traj control (as I have no real experience with that)...i don't really understand what you mean by local details....i'll look into it......more research for me!!


i couldn't opentthe WF 2 prt on my machine...I need to move out of wf and up to 3.0.....I'mnot really sure what your dimx dimy and dim rho was all about...those were pre-defined parametersfor control correct?


Thanks guys!
Edited by: TWINTURBOTOM

 

[jeff4136]

> check out the traj control (as I have no real experience with that)


View attachment 2980


A trajectory doesn't have to have section geometry constrained to it.
Here the section, a constant dimension square, is constrained to
Chain_1. The Origin traj just controls the section plane normal. For
your shape I'd use a curve going from start section centroid to end
section centroid (identical to the trajectory you'd probably use for
a Swept Blend).


> don't really understand what you mean by local details


Just a rule of thumb and probably not relevant to your current task.
If you need to force a relatively small irregularity in a large
surface, you should probably do the irregularity as a seperate
feature and merge into the larger one instead.


> not really sure what your dimx dimy and dim rho was all about


They're Evalgraph features controlling VSS section dimensions and rho
(each quadrant of the shape is a conic arc). They serve a purpose
similar to your 'flow' curves but, being planar, can be easier to
control, more finely controlled (?) than curves in 3D space.


For your task the Swept Blend sounds like a good candidate. It's not
hard to 'over constrain' them as well, though. For instance, defining
end tangencies and introducing a mid section may set up unwanted
bulges, waves, etc. You may find it easier to control doing it with
two (or more) features instead of one.


Good luck with it ...

 

[TWINTURBOTOM]

awsome....your the man....you cleared that up for me entierly.....Now I undersand the traj control....i always thought the trj had to be a part of the section...stupid me!...I've modeled stuff and not had the traj touch the section but didn't realize it could be anywhere and control it by modifying....THANK YOU!!!!!!!!


i was debating two features.....i'll work on this more....thank you for everything!!!!!
Edited by: TWINTURBOTOM