×
INTELLIGENT WORK FORUMS
FOR ENGINEERING PROFESSIONALS

Log In

Come Join Us!

Are you an
Engineering professional?
Join Eng-Tips Forums!
  • Talk With Other Members
  • Be Notified Of Responses
    To Your Posts
  • Keyword Search
  • One-Click Access To Your
    Favorite Forums
  • Automated Signatures
    On Your Posts
  • Best Of All, It's Free!
  • Students Click Here

*Eng-Tips's functionality depends on members receiving e-mail. By joining you are opting in to receive e-mail.

Posting Guidelines

Promoting, selling, recruiting, coursework and thesis posting is forbidden.

Students Click Here

Jobs

Symmetric Spline problem
2

Symmetric Spline problem

Symmetric Spline problem

(OP)
Hello everyone,

After creating a spline I want to mirror it at both endpoints. In the spline dialog there is a continuity type called "Symmetric Modeling". I can apply this successfully to one endpoint, but when applying this continuity to the other endpoint as well, NX gives the message "Cannot compute bcurve". The attached image shows how the spline is mirrored at both endpoints using symmetric continuity at one endpoint and tangent at the other. As you can see, the spline is not continuous where the tangent continuity is applied.

What can be done to achieve perfect symmetry with splines at both endpoints?

NX8.5 Win7SP1 64bit i7-3770K@4.3Ghz 16GB Quadro2000

RE: Symmetric Spline problem

Draw a reference line perpendicular to the mirror "line/axis", then constrain the spline end point to have G1 or G2 continuity with the reference line.

RE: Symmetric Spline problem

(OP)
That's exactly what I've done where I've applied the tangent continuity, but as you can see with the comb analysis, the curvature of the spline at that G1 endpoint is not continuous with the mirror spline.

NX8.5 Win7SP1 64bit i7-3770K@4.3Ghz 16GB Quadro2000

RE: Symmetric Spline problem

Hi Cnsu.
How many control points do the Spline has? If you wish to have perfect G3 (G2 continuity is already there on both planes as i can see from the the image) then you must have at least 7 control points(or may be more).Actually once the G2 continuity is achieved then G3 can be achieved using the middle pole.
Or the best you can do is....when you achieve G3 at one end then use X-FORM to fiddle with the middle control point (while keeping continuity check (G2/G3 ) through BOUNDARY CONSTRAINT.This way you wont loose your contiuity at the opposite end.
Best Regards
Kapil

RE: Symmetric Spline problem

(OP)
It doesn't matter whether it's G1, G2 or G3 continuity. What matters is that the curvature is continuous between the original curve and the mirror curve. This can be checked with the comb analysis tool.

NX8.5 Win7SP1 64bit i7-3770K@4.3Ghz 16GB Quadro2000

RE: Symmetric Spline problem

(OP)
The "bump" you mentioned is what I want to avoid because it indicates that the curve is not flowing smoothly from the original to the mirrored curve, even though it has G2 continuity. That's why you have the option "Symmetric modeling" in the continuity type options, but it only works for ONE of the endpoints (not both).

NX8.5 Win7SP1 64bit i7-3770K@4.3Ghz 16GB Quadro2000

RE: Symmetric Spline problem

2
In the second post you say that you have made it tangent ( G1 or maybe G2 ?) to a line at the top , right ?
If so, - what curvature does a line have ?
Zero. And that is what i see in the image. Since a line has zero curvature, the spline matches that and the end of the spline has therefore zero curvature.
But, if you mirror that over YC in the image, the result is actually G2, - Curvature continuous Since the G2 definition only describes that the curvature on each side of the point is equal.

What you desire is that the rate of curvature change is zero at the point. - No acceleration/deceleration of curvature in/at the endpoint(s). - which is what the Symmetry condition defines.
To accomplish that I think you need at least 8 poles. ( it takes 2 poles at each end to match G1, 3 poles at each end to match G2, 4 poles at each end to match G3. Symmetry should be a special case of G3.)

Now note that "symmetry" isn't a guarantee for a good looking curve... It's very difficult to handle the shape of the spline if the 4 poles at each end are controlled by the system, -You have very little freedom to manipulate the shape left. Also 8 poles does allow quite a lot of undulation in total.

I have attached an example. Edit the spline and play around with the poles to see what i mean.

Btw, the way i set the symmetry is to select, in the dialog not in the graphics, the first or the last point, then specify symmetry + the direction. ( again dialog)

Regards,
Tomas

RE: Symmetric Spline problem

(OP)
Thank you Tomas. Changing the spline type to "by poles" (not through points) and increasing the number of poles to 8 solved the problem. Also I greatly appreciate that you explained the theory behind what's happening.

However, in your example, you use a degree 5 spline with multiple segments (there are 2 knots visible). After playing around I've discovered that checking the box "single segment" in the spline dialog increases the degree to 7, and that makes it easier to control the shape of the spline because the poles seem to be more evenly spread apart. Another advantage by making the spline a single segment is that when creating a surface based on that spline (eg swept using the spline as guide and an arc as section), that surface has fewer poles, seams and patches than a surface based on a multiple-segment spline, thus creating a cleaner surface.

NX8.5 Win7SP1 64bit i7-3770K@4.3Ghz 16GB Quadro2000

RE: Symmetric Spline problem

- I did not dare to propose degree 7 smile but since degree 7 and single segment has 8 poles by nature, it is probably what one should use from start.

Regards,
Tomas

RE: Symmetric Spline problem

Forgot to mention that in most cases, you can do well just by "eyeballing", i.e since the junction will be G2 no matter what you do, it's a matter of making the curvature comb look right at the mirror point. - Getting the shape you desire. I.e on the vertical YC the edge of the curvature comb should be horizontal ( parallel to XC) And by increasing the scale and in some cases the number of needles of the curvature comb, and then adjusting the poles ( use the microposition option) you will have reached a level which is fully ok. ( and quicker to make.) See attached example. As a comparison, a "standard degree 3 spline" can have a far uglier curvature comb due to that the interior knot points are G2 but not G3. ( Splines degree 4 and above has G3 knots.)

Regards,
Tomas

RE: Symmetric Spline problem

(OP)
You have created a very nice shape, and proof of that is the curvature comb which is smooth as a baby's ass. However, your methodology of "eyeballing" is non-parametric.

The way I would do it is first sketch a rectangle, then create a single segment 7 degree/8 pole spline with symmetry modeling continuity associated to the mid points of the rectangle, then mirroring the spline to create 4. The advantage with this method is that although the initial "setup" is more time consuming with adjusting the poles, when you've reached a satisfactory spline, later you can simply change the size of the rectangle, and the 4 spline curves will update and still retain the original curvature, including perfect symmetric curvature.

NX8.5 Win7SP1 64bit i7-3770K@4.3Ghz 16GB Quadro2000

RE: Symmetric Spline problem

Unless I'm misunderstanding you, I think a degree 7 spline would still have 2 poles that were unconstrained and you would probably have to eyeball their location once you edited the rectangles. Applying G2 to a spline locks in the last 3 poles at each end so 2 would still be floating free. If this is the case, your spline is still not fully parametric, unless you plan on adding points to a sketch and assigning the spline poles to those points (which would be severe overkill, in my opinion).

G0 (position) locks in the end pole; G1 (tangency) locks the next pole inward; G2 (curvature) locks the 3rd pole from the end; G3 (acceleration) locks the 4th pole from the end of the spline.

I'm not trying to imply that a parametric approach isn't correct, as it's mainly a preference from one designer to the next. However, it's usually easier to deal with splines by leaving them just associative and not trying to completely lock it down at each and every pole. The position, slope and magnitude will affect the majority of the spline's shape with these 5 and 7 degree splines. Eyeballing will suffice in most instances unless we're worrying about aerodynamics or matching existing surfaces.

Again, not trying to imply you're right or wrong, just pointing out some small things that are typical to spline behavior.

Hope some of this is helpful to you.

Tim Flater
NX Designer
NX 7.5.4.4 MP8
WinXP Pro x64 SP2
Intel Xeon 2.53 GHz 6GB RAM
NVIDIA Quadro 4000 2GB

RE: Symmetric Spline problem

The above is only true outside of a Sketch, unless you're not dimensionally constraining the spline points/poles. Haven't delved into splines in a sketch in years, as I've rarely ever needed to control the spline points or splines; eyeballing was enough. Sorry if that seemed to be misinformation of sorts.

Tim Flater
NX Designer
NX 7.5.4.4 MP8
WinXP Pro x64 SP2
Intel Xeon 2.53 GHz 6GB RAM
NVIDIA Quadro 4000 2GB

RE: Symmetric Spline problem

(OP)
In a way you're right. 2 poles are locked into position (G0), while the remaining 6 poles have 1 DOF. So technically it would seem as if the spline is not parametric. But testing this by changing the "skeleton" rectangle to different sizes with varying proportions, the spline does update in a parametric fashion. The poles spread out or contract proportionately according to the rectangle size. And to test how parametric the spline is, after several modifications to the rectangle size, changing back the the original size, only 3 of the outlying poles are just slightly off from the original locations. This is adequate for my purposes. And it's much better than a 5 degree/3 segment spline where changing the size will cause a complete mess.

NX8.5 Win7SP1 64bit i7-3770K@4.3Ghz 16GB Quadro2000

RE: Symmetric Spline problem

That's not really a 'parametric' behavior but rather a 'proportional' relationship. In fact, if you check you may find that there's been a 'constraint' applied to the spline called a 'Uniform Scale Constraint'. Now if you move one of the interior points it will change the shape of the spline, as you would expect, but if you move either the first or last points the entire spline will change size but the shape will be preserved since in essence moving the first/last point will 'scale' the spline rather than changing its shape.

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

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

RE: Symmetric Spline problem

(OP)
John, when will the sketcher spline become associative? There is a "slope of curve" geometric constraint which is very useful for creating a G1 constraint to a spline, but there is no option for G2,G3 or symmetric modeling. If I add these continuity types in the spline dialog, they disappear as soon as I want to modify the spline. In the meantime I have to create a spline separately outside of sketcher if I want associativity.

NX8.5 Win7SP1 64bit i7-3770K@4.3Ghz 16GB Quadro2000

RE: Symmetric Spline problem

As far as I know, the 'constraint engine' currently only supports Tangencies, or at least this is all that we've implemented in the Sketcher at this time.

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

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

Red Flag This Post

Please let us know here why this post is inappropriate. Reasons such as off-topic, duplicates, flames, illegal, vulgar, or students posting their homework.

Red Flag Submitted

Thank you for helping keep Eng-Tips Forums free from inappropriate posts.
The Eng-Tips staff will check this out and take appropriate action.

Reply To This Thread

Posting in the Eng-Tips forums is a member-only feature.

Click Here to join Eng-Tips and talk with other members!


Resources