On 2013-09-23 14:35, WAYMAN John wrote:
> Sorry to resume this topic, folks, but I am continuing to learn about
> surfaces this way.
>
> I am still having trouble with my original surface blend. I can use a
> Boundary blend to join the two surface together, but Pro/E decides to
> add in some extraneous lumps & bumps along the way! If I look at the
> finished blend in wireframe, I can see squiggly lines that double
> back
> on themselves where I would expect only straight lines.
>
> Squiggly lines:
>
> Just one of the four corners of the real part, with its squiggly
> lines.
>
> As an experiment, I created a picture frame (or something similar).
> It comprises a sketched datum curve in the shape of a rectangle on
> one
> datum plane and another rectangular sketched datum curve on another
> datum plane, parallel to but offset from the first plane
>
> If I create a Boundary Blend between the two curves, tangent to each
> curve, I get the thing shown above.
>
> Again, I would expect straight lines like those in the top left and
> bottom right corners to describe the shape of the blended surface,
> but
> I see a pair of random-looking lines in each of the top right and
> bottom left corners.
>
> They are not as squiggly as those in the real problem part, but this
> part is much more simple - it's planar, whilst the real part is
> curved
> around both the TOP/SIDE axis and the TOP/FRONT axis.
>
> Can you explain to me why Pro/E puts in those two random-looking
> lines (I'm sure they are NOT random at all) instead of just joining
> all four corners of one sketch to the corresponding corners of the
> other sketch?
>
> I tried creating curves between corresponding corners and using them
> as Direction 2 curves. That works as I would expect, but I can't make
> the resulting blend tangent to the two planes.
>
> If, instead, I use the corner-to-corner curves as Influencing Curves,
> the thing will not regenerate with tangency.
>
> I thought this was a pretty simple surfacing problem! I'm glad I
> don't have a hard one to sort out…
>
> Any assistance gratefully received, as usual.
>
> Thanks,
>
> John
You will note that the 2 curves created by Pro/E start at each of the
corner points of the curves.
What Pro/E is doing, is taking the total length of each of the curves
you are using, and then create a surface by:
* taking a certain ratio of the total length of each curve
* create a point at that ratio on each curve
* create a curve connecting those two points
Of course all this is done below the hood, so you never get to see the
points and curves, *except* on certain spots like corner points in the
geometry, hence the two curves in the corner.
Half way in your surface, (ratio 0.5 on each of the curves), the two
virtual points created by Pro/E end up exactly on the corner point for
both line, so there you only get one line.
In this case, it seems like the product should be symmetric in two
directions. So the best thing to do would be to model only a quarter of
your current surface, and then mirror it in two directions. (Right now
your surface is not symmetric!)
That is my theory anyway.
Best regards,
Patrick Asselman
"Guns Are One Thing But Phone Calls Are Too Dangerous To Go Unchecked"
![[cid:image005.png@01CEB837.C5843110]<http: twitter.com=" dukaneias=">[cid:image006.png@01CEB837.C5843110]<http: www.youtube.com=" user=" dukaneultrasonics=">">http://usblog.dukane.com[cid:image004.png@01CEB837.C5843110]<http: www.facebook.com=" pages=" dukane-intelligent-assembly-solutions=" 3139592...](http://usblog.dukane.com[cid:image004.png@01CEB837.C5843110]%3Chttp://www.facebook.com/pages/Dukane-Intelligent-Assembly-Solutions/313959232413?ref=ts'%20style=){kind=link}
