Some surfaces behaviour makes me confused

[Jacoolo]

to jeff

well, I `ve wanted to stay with this is silence.Trying to be quiet untill everything is perfect - that was a hope for this particular chalange



so basicaly the last g1 and g2 comparison was`nt for 3 sided b.b. This appraoch was canceled by me long time ago after Bart suggestion about this - thx Bart, I wait for more like this)))

however I use another approach for this kinda situation

instead fo this



I prefer this



and if You really are interested in helping me with this design, tell me how best bite this thig



for the first view it looks pretty easy but in the second the difference between sections in front and back part of handle can be seen

 

[Jacoolo]

to skint

yea I posted something but it wasn`t design related though.

Since I am not industrial design and my background comes from mechanical engineering(4 years in truck industry pneumatic breaking related), it is hard for me preper any intresting design

however, since I am so much surface involved now - there is more interest from my personal side than company at all - I search all sources which could serve as good basis to understand good surfacing dogmas(tenets, rules)

this froum serve also as good source for chalange models

 

[jeff4136]

> really are interested in helping me with
> this design, tell me how best bite this thig


Did you try to create the final surface using
the boundaries you've created?


I think I understand what you are trying to do
and think you are going about it differently than
I would. What I presume is to be a bounding surf
(selected with the int crv on it, the one shown
in previous post?) would be my final blend surf
after extending the top surf down and trimming it
back or maybe extending the lower surface up and
trimming it back to get a boundary similar to what
you have now. The problem with doing what (I ~think~)
you are doing is using a guestimate curve to create
the indicated surface. As the curve is a freehand
guestimate it, or the surf created, may not be
compatible with the existing surfaces and just pull
things out of whack.


Again; this stuff is too complex to do from pictures.
All I can do is try one thing or the other to see what
works best. What will work best depends on subtle
differences in the bounding surfaces and conditions
they'd impose. There are no one-size-fits-all answers
and it would be a waste of time to recreate something
that looks sort of like what you have because my solution
may not work or be possible with your geometry. If you
really want me to try to help, to whatever extent I may
be able; post some geometry. It's ~possible~ a solution
might have been modeled in less time than it took to
write this and we'd both be learning something.

 

[Jacoolo]

Since I do not want to show to much - hence, I ve already shown a lot

- I gonna share some oldest stuff

back to the them:
nope, this surf I mentioned earlier is not a part of handle surfaces. deatails below



the bottom part of the model is much easier to create, still quality and overall view is not as desired though

my idea to deal with handle surfaces is to use VVS for bottom part of it. see pic



I tried many other approaches but VSS seems to be correct for this particular design, and the easiest either

the mess in pics comes from my general appraoch to make surf models - I do not pay such attention on no. of feats. and general oreder in tree. However, after experimenting, all should be in right place, and it is.

 

[Jacoolo]

The problem with doing what (I ~think~)
you are doing is using a guestimate curve to create
the indicated surface. As the curve is a freehand
guestimate it, or the surf created, may not be
compatible with the existing surfaces and just pull
things out of whack.

Jeff can You evolve this case a little further? Specialy - "questimate curve", and "indicated surface"?

Second - required - desired analisys. Which one and when exactly do You(all guys) use? What are the desired result?

I took a look at the surfacing book I still have after training I had participated 2 years ago. There is a chapter specialy for this case but in the end it tells nothing worth. Just say it shows the tools. Amen

 

[Jacoolo]

I push this a little further, but some feats had to be cut out.





you see, the big round is in the place but the small raound is missing

anyway this was achieved with surf quality droped away(

 

[Jacoolo]

hi jeff

I took a look on a model of a car You had attached in another topic sometime ago



and some intresting things catched my eyes

first - can You tell me how did You set marked tangency constraints in pic below



second - I noticed You use offset or create rounds to obtain curves for B. Blend. Why didn`t You use offset of intersected curve instead?



third - what is the role of this small B.Blend?



fourth - what is the role of this VSS?



thanks in advance Jeff

 

[Jacoolo]

I want to recreate following surface






I mine looks like this



However I wonder if it could be done better

this is my appraoch -





these conditions let me create two anther surf with tangency conditions



Is there another approach to do this?

 

[Jacoolo]

some questions:


second case - How can one find a tangent point in case below

Yeah, I`ve found the solutiuon at last

in sketcher > Analisys > Tangency point



and the result

 

[Jacoolo]

one small issue makes me stick with it for a while

I created a profile which contains simple line an spline. I converted them to spline. The result is one continues curve but with a curvature I do not desire. Take a look below



In the pic above You can see the profile itself in sketcher mode(I wrtk now with blended background in sketcher, and maybe it is not clear enough, but it is cool anyway), first curvature comes from analysis tool made in part mode, the second one comes from spline tool direct in sketcher mode.

You can see that the curvature from spline mode looks diferent, and I suppose it gives a clue how this spline looked like before I converted it with simple line to spline

the overall curvature for this converted entities has a litlle peek which is all but not the desired one by me

do You a clue how to avoid that, and how to tweak sketcher entities to obtain better curvature after converting them into spline in the end?

 

[Jacoolo]

I am really impressed with VSS

I played a little with tangency option included in VSS and with this fine geometry can be done with one tool only

very nice





tangency transition in back area of this part was intended.

 

[jeff4136]

Jacek,


Conic arc tangencies are, in my experience, best created during curve
creation (pull the curve back and forth after end point definition and
you should see the tangencies being 'proposed'). Sometimes the
constraints can be applied after creation but I find it less reliable
(returns an invalid choice error). I just as often settle for
dimensioning the tangent angle or will sometimes anchor a CenterLine
on the end of the conic and swing it over the tangent vector. You'll
see the curve grab the tangent and display the proposed constraint
glyph.


That model's from a long time back (my pre-fledgling days, now I'm a
full fledged fledgling) and I'd have to dig a lot to find the feats
from the pics ...


If the (VS)Swept sections are lines; I was using them to find a point
on one curve that is in a plane normal to the other curve.


I sometimes use BB surfs to create curves from; by intersection or
just thru points at .5 rel on each end of the surf. I use the BB
surfs because I can 'tweak' the constraint influence / weight /
stretch (which is a totally unworkable proposition on G2 Datum Curves).


Offset on Srf crvs: That's just one of those features I haven't
investigated much. Seems I may have tried once or twice and didn't
especially like the curvature graph created from it, but that's a
really premature opinion.


... If those explanations don't seem to be appropriate, give me the
FID's and I'll look closely. (Looking for children may explain, too.)


Merging curves: The curves being merged must have a very high degree
of continuity (G3+) to avoid oscillations on each side of the 'join'.
And sometimes even then slight oscillations may occur. Experiment
with different ways of merging. You'll sometimes see a difference
between (1) Convert to Spline, (2) project to new sketch and convert,
(3) Approx Copy. A thing to keep in mind about merged curves; they
will almost invariably have a higher knot density, end to end, than
their parents so there's a little bit of overhead associated.

 

[Jacoolo]

some fun with VSS

 

[Jacoolo]

and curves network

 

[Jacoolo]

almost all with VSS

 

[design-engine]

I see my smile tutorial in the thumb location of the handle. did you use conics in the product line?

 

[Jacoolo]

I see my smile tutorial in the thumb location of the handle. did you use conics in the product line?

1)yeap, you`re right, the smile it is.

The difference is the bottom
surface isn`t made with Boundary Blend but with VSS+tangency option.
This way is better

2)Product line - do You mean main curves for geometry? I use curves based on sketchers, which contain splines - exactly it is like below. Spline is set by control points. I see this way a lot better vs spline controled by interpolation points.



Guy like me is gonna learn whole life

 

[Jacoolo]

Since my lever model some posts ago, the case seems to be left forever in my head, rigin until it will be solved to the end.



I change a little basis of lever and made this bath part

The main thing which do not let me sleep well is the solving of 5 sided B.B

The composite curve I create to solve this case is a cause for additional edge which seems to be out of my control

the curve network



the case



questions :

1) is there any control from B.B level considering this tangency edge marked above(I created even an additional curve from the same place to use it in B.B. No way the gray edge it created no metter if I do it or not!)?

2) is there another better way to solve these 5 sided conditions for mentioned case?

 

[Jacoolo]

I struggled quite a bit with round case shown below. Pro\e gave up at start. I wasn`t so soft.

I walked through many soloution like these:


*sh*ty transition between rounds - not acceptable

*the same

and at last





not the best but worth the time

Hey, guys what are yours workarounds for following case?

 

[Jacoolo]

I`ve change handle model. Now it includes rounds. With this improvement some things appeared hard to resolved(to be done).



Do You have any suggestion how this design could be done different? I think bottom part of handle does not look right.

I run this problem once again, simplyfing a little overall conditions

it could be solved nicely with such appraoch



the changes are apparent(the shape simplicity also

)



I had to make overbuilded surface bu 3nd BB, while 4 nd BB genrated many problem. To make the task a little harder there is no straight line in the bottom contour. 3 conic arcs are responsible for it.

The smooth of surface is still ahead of me though



The lack of tweaking ability for curve(Curve through points - no ISDX) constrained on both ends with Curvature conditions really pisses me off.