Curve by equation / motion skeleton

[PI_6131154]

Hello!

I need help to define curve with equation. I have simple loop as shown in the attachment.

I have the problem to guide a point /curve) on the loop.

Although, I use the loop curve at all as reference (and not jus a segment) and I can quide the point in the sketcher, the loop will be not  identified by the motion skeleton.

I suppose, that if the curve is generate by equation, maybe it will run.

Can somebody write the equation for this loop?

Thank you!

 

 

 

 

[pausob]

Hi @sacquarone, I have to say your presence in the forums has been a welcome surprise (forgive me for my ignorance, but to me it seems it just recently started).  I am grateful for your insight and helpful attitude and I commend you on how you respond to the issues - with an impressive amount of detail and easy to understand video snippets.

It has certainly made me rethink about my previous attitude about PTC not caring about customers (except those with large accounts).

Regrading the approximate curve issue; this technique of turning a series of sketch segments into an approximate curve and subsequently using it as the reference to the slot constraint was was one of the many new things I learned from you.  In my toy model, after I make the size of the "chain" too large, I get messages about "mechanism has become disconnected".  So I think there is more to it than just something related to zoom-factors.  In fact, there is one thing that is troubling about the "approximate curve" tool in that the function has no internal parameters exposed to its users and it seems to me that at some model size, it produces rather strange results (at least as far as mechanism connections are concerned).  Absolute accuracy or mechanism accuracy settings didn't solve the issue for me.

So the work-around of selecting each segment of a curve when specifying the slot constraint sure works, but it is also rather clumsy.  If you have a chain model with 100 links attached to it, and then decide to change the shape of this chain by adding a tensioner element - now you have 100 connections to redefine in order to rebuild the model!  I thought the approximate curve would be a neater way, but thinking more about it, using it is also a work-around.

Shouldn't the user be able to simply specify a sketch feature or an intent chain as a reference to the slot constraint?  I know there is a process for asking for such improvement to be implemented - but I don't have a maintenance account so I can't even post an idea about it...  Maybe @PI_6131154, you could post it?

[TomU]

I recently opened a case to figure out why a slot follower connection through visually identical curves performed drastically different during dynamic analysis.  One curve was constructed as a curve through 360 points, the second through 3600 points, and the third was any approximate curve created by copying one of the other curves.  One of the things that case brought to light is the disadvantage of using curve through points or composite curves.  Both of these are generated using splines and the 2nd derivative used for calculating acceleration looks horrible.  R&D recommended only using 'analytical' curves,  specifically line, arcs, ellipses, and curves through equation.  Of course this doesn't work well when you're trying to create a slot follower connection and have to manually select 360 (or 3600!) line segments.  I still don't have a good solution.

Here's some of the stuff from my case correspondence:

----------------------------------------------------------------------------------------------------------------------

I did a quick test where I created a bunch of line segments between a small subset of the points and of course the curvature is non-existent (since each one is a straight line).

Creating an ‘exact’ copy of this curve also has no curvature because it replicates all the individual segments.

Creating an ‘approximate’ copy of this curve creates a single spline, but the curvature looks just as bad as the original spline.

I also tried reduced the point count from 3600 to 360.  It’s amazing how much smoother the curvature (and probably the 2^nd derivative) is with less points, even though it’s still a spline.

Curvature of original curve with 3600 points (at the same scale):

[Patriot_1776]

Yikes!  Yeah, I gave up trying to use the "spline" command in sketcher.  What Creo needs is simply a "polyline" command like they have in AutoCAD that simply turns all the elements, arcs and lines etc., into a single entity, WITHOUT turning it into some monstrosity of a zillion spline elements.  That's just absurd.  I can think of a TON of benefits to that simple enhancement, from motion curves to drastically enhancing the function of Variable Section Sweeps, and even enhancements to Boundary Blends.  But noooooo, they'd rather waste time annoying us with new ribbon graphics and menus....  Have their code-jockey developers spent a single minute actually USING the software in the real world as a customer?  I think not.

[kdirth]

Have you used Convert to Spline in your sketches before?  This combines all of the selected entities into a single entity with all of the existing dimensions still active.

Is this similar to what you are referring to in autocad?

[Patriot_1776]

Yup, that's what I was talking about.  At least in Creo 4, it does NOT act like AutoCAD (I used that for 10 years+ starting in '86/v2.18, still miss parts of it) where it simply joins all the elements together.  What happens in sketcher in Creo 4 is that it converts all the entities into a kind of bastardized zillion-element spline, that you can't use constraints on.  I haven't tried it in a VSS but considering what happened when I tried to use it as a motion path, I think it will STILL not allow me to pick it as one entity, making it useless as a reference.

[pausob]

Interesting.  I thought curves through-points weren't so bad, and I'm surprised about the R&D suggestion for using curve from equation.  These are not well behaved in my experience!  I mentioned this a while back in this thread https://community.ptc.com/t5/Analysis/cot-cotangent-in-the-equation-for-Curves/m-p/546951 in which I show that the curvature graphs of the curve-through-points are smoother than the ones from the curve-from-equation:

Anyway, I think we can all agree that we don't have much of control over these Creo curves!