cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Showing results for 
Search instead for 
Did you mean: 

Community Tip - Need to share some code when posting a question or reply? Make sure to use the "Insert code sample" menu option. Learn more! X

Variable section sweep with multiple sketches?

pimm
14-Alexandrite

Variable section sweep with multiple sketches?

I'd like to build a locked sidewall that has varied draft.

I was under the assumption that the best way to build this feature would be a Variable Section Sweep as my sections would vary along the length of the side rails.

Can you have multiple sketches in a Variable Section Sweep?

If yes, could someone explain the method of picking up additional sketches along the length of the trajectory?

If not, what would the appropriate tool be for what I am wanting to accomplish?


This thread is inactive and closed by the PTC Community Management Team. If you would like to provide a reply and re-open this thread, please notify the moderator and reference the thread. You may also use "Start a topic" button to ask a new question. Please be sure to include what version of the PTC product you are using so another community member knowledgeable about your version may be able to assist.
28 REPLIES 28
TomD.inPDX
17-Peridot
(To:pimm)

You might look at it as additional "guides". As long as you don't "lock" the swept section, the section will adjust to the trajectory and guide sketch. When you select an additional trajectory(hold CTRL), it will be the 1st "guide" and additional sketches will be the additional guide features.

swept_VSS.JPG

Patriot_1776
22-Sapphire II
(To:TomD.inPDX)

Going further, what I'd like to see is the VSS command allow both multiple trajectories AND multiple sections. This way, you could get around a problem I have where you can't force an arc to be flat, or vice versa. also, you could do blend vertexes like in a blend/swept blend.

Of the 2, I like and use VSS's way more, for various reasons including using trajpar.

pimm
14-Alexandrite
(To:TomD.inPDX)

Hi Antonius,

From what you show in this example there are 2 trajectories. What I was hoping is that Variable Section Sweep would allow for multiple sketches applied somewhat perpendicular to just 1 trajectory. From what Frank suggests you can not add multiple sections along the trajectory.

In another respect what you have shown could be quite helpful in respects to getting a tweakable side wall.

If I needed to modify a sidewall with lock variation and varied draft; would there be a good tool in Creo for placing multiple sections along the trajectory path?

Thank you for suggesting a different approach. I will try this and see how accurate I can match the customer's sidewall geometry.

Patriot_1776
22-Sapphire II
(To:pimm)

What you CAN do, is sketch a curve at the start, and at the end, then use those to drive the multiple trajectories. Note that you CANNOT tie any geometry used in the single sketch of the VSS to any static geometry (unless you set the option to "Constant Section", or it will either fail or give weird results. Now, you can if you're REALLY careful measure things and write relations, or make things egual, but it's VERY tricky to get it right. You can also lock some dimensions in the sketch, and drive others by the reference points of the trajectory, or drive ALL the vertexes by trajectories. It's very flexible that way.

Good luck.....

pimm
14-Alexandrite
(To:Patriot_1776)

I've run into some difficulty in using Boundary Blends instead of a Variable Section Sweep.

I recalled that someone in this thread suggested that sketches could be made at the beginning and end of the VSS. This could be quite helpful in my application.

Frank: What technique do you use to have a sketch at the beginning and end of a VSS?

Of equal importance; I'd also like to ask how you can substitute your own datum positioning for the sketch instead of the default positioning that pops up as soon as you select "Create or edit Sweep Section". In what I see you are channeled into using whatever positioning that automatically gets chosen for you. For the life of me I can't see how you can substitute your own datum positioning for the sketch that I'd really like to have.

TomD.inPDX
17-Peridot
(To:pimm)

If you're talking about swept/blend, it is easiest to just pre-define the sections:

sweep-blend.JPG

pimm
14-Alexandrite
(To:TomD.inPDX)

I actually was inquiring about how to have definable datum positioning for a Variable Section Sweep, and also how you would be able to have this on both ends.

Boundary Blends do give you more flexibility in this regards, yet there are advantages to the Variable Section Sweep.

TomD.inPDX
17-Peridot
(To:pimm)

The variable section sweep will only accept a sketch plane "normal" to the trajectory.

I don't think the VSS can be defined with an end sketch. As close as you can get it to have endpoints from your trajectories where the initial sketch will align with. You will still have the issue of the sweep ending prematurely on some trajectories. However, you can extend them.

Boundary blends are finicky but can be very powerful when defined in a manner that Creo want to see it.

This is why I introduced the Swept-Blend. Now you can take sections within your existing part; define one or more trajectories again from existing geometry, and blend them all together. Now you have a variable section each as defined by your "master" model; you are not bound by "normal" orientations of the -selected- sketches; and you can blend the end conditions tangent or normal to existing geometry.

Frank, you can weigh in here but I would think that a swept blend can outperform the VSS by adding an extra level of control.

Patriot_1776
22-Sapphire II
(To:TomD.inPDX)

Well, yes, and no. With a VSS, you can't control the endpoints directly, but you can make sure your trajectories all force the end section to be what you want. As noted, there is no blend vetrtex, but yo can still force a line or curve, or even the entire sketched section to zero. What you cannot do, in a VSS is force a curve into a line, or force it to reverse it's curvature direction (i.e. convex to concave). I think a VSS is superior in that you're controlling the swept shape via the other trajectory curves along the entire length of the feature. A swept blend only controls the shape at those defined sections and interpolates things in between. so, sometimes you add more sections, and end up over-controlling it with some weird results. A boumdary blend works great, except you must totally define ALL the edges exactly, so it's more work, but you do get some extra controls not in the other features.

Like I said, I'd love to see VSS and swept blends combined, so you could add sections, reverse curvatures, and still be albe to drive all the vertexs in a sketch via other trajectories. And, to top it off, how about have a tangency, normal, and curvature continuous (REAL C2) options as in the boundary blends built into the VSS command? Ormaybe have them as separate features after the fact, like a surface extent or trim etc.? Maybe if they were's wasting so much time on ribbons we could get these REAL enhancements......

pimm
14-Alexandrite
(To:Patriot_1776)

I'm with you 100%. I really wish they could consolidate the major tools into 1 more powerful function.

From someone who really hasn't got his feet yet I find the options available to be very bewildering. I can almost reach the finish line with a specific tool and always have some reason that I can't take it in to completion.

I appreciate the dialog from this thread. It gives me ideas to go back to when it comes time to try an idea out.

Frank: So it appears with VSS you pretty much are roped into normal conditions with the trajectory; you mentioned something above. What you CAN do, is sketch a curve at the start, and at the end, then use those to drive the multiple trajectories. How do you add the end curve?

It is interesting when perusing the forum how people are constantly complaining about functionality and ease of use and what you receive is window dressing. The CAD system we are phasing out decided to take the Ribbon route as well. The last thing we would look at when reviewing CAD systems would be the Ribbon. I can get used to this, but there is no value added.

Patriot_1776
22-Sapphire II
(To:pimm)

I would first develop the curve you use for the origin trajectory, then sketch curves on planes normal to the endpoints of the trajectories. Then create your curves that you want controling the vertex(s), making sure they end up at the correct vertexes of the curves at the ends (and middle if desired. This way the other trajectory curves control the sketch the whole length of the origin trajectory, and end up exactly where you want them.

TomD.inPDX
17-Peridot
(To:pimm)

In a sense, I am using your approach in a slightly different way. You can easily make multiple sections to manage your draft variation. Think of them as sections along the way to decide on where the guide curves, or trajectories, need to go. This works even better if you need to change elevations because the trajectories could be datum curves that are not planar if you need this.

Multiple sections sweeps are also very powerful and you really need to take a careful look at what you want to accomplish.

Also give careful thought on when to use Offset to keep these sweeps simple and adding details after the sweep.

...And of course there is Swept-Blend which is "variable" by definition. It is not as rigid in following guide curves, though.

pimm
14-Alexandrite
(To:pimm)

The Variable Section Sweep "Almost" works well for me.

I am happy with how the surface rail matches up with the actual part. This looks like a tool that makes locks and varied draft easy.

Unfortunately I am finding that I can not get the start and end of the sweep surface to end up where the 2 guide rail sketches start and end. Since this segment is only a part of the side rails on the part profile I have to have the surface start and end exactly where my guide sketches end.

When I normal to trajectory or normal to projection I'm finding that there is a small gap to the end of the sketch.

If I choose a constant to normal projection the starting surface lines up perfect to the start of the guide rails, but the end of the surface falls way short on one of the ending guide rails.

I can't seem to find a way to build my own sketch that would lock to the exact end(s) of the guide rails.

Is there a way of ensuring the start and end of the surface lines up with the start and end of the 2 guide rails?

TomD.inPDX
17-Peridot
(To:pimm)

I noticed that with my quick sketch too. Obviously, you could simply go further than you need with both guides. The problem is that the sketch is normal to the trajectory although you can select another normal for the sketch.

Often, complex shapes are created from many surface operations where in the end, it becomes a closed volume or a thickness can make these solids. That remaining gap, for instance, could be filled in with a Fill feature.

Can you share a image of your challenge or even a part of the file you are trying to affect?

pimm
14-Alexandrite
(To:TomD.inPDX)

RailsWithPart.jpg

This shows a couple sweeps that I applied on opposing sides of the parting line. I might have to overbuild the sketch extents as you suggest. I just wish this would stick to the start and end of the sketches.

TomD.inPDX
17-Peridot
(To:pimm)

A boundary blend would do that. It also has options for tangency. You would be working with 3D datum curves and reference surfaces to control tangency. I would probably opt to work in this part with surfaces.

Are you doing an edit, or are you defining the whole part?

pimm
14-Alexandrite
(To:TomD.inPDX)

I replied yesterday but see that the post didn't take.

As I mentioned earlier the Variable Section Sweep almost does what I need it to. If only the ends would lock in place.

From what you say, the Boundary Blend might be an even better way of getting the results I would be looking for as it would hold to the ends and also allow sections in the secondary direction.

In comparing Variable Section Sweep to Boundary Blend, what are the drawbacks to Boundary Blend? Would it be a lack of influencing directional control? Boundary Blend does appear to be a better choice for my application but I would anticipate this would require some re-working of the upper rail curves as I would anticipate the results will be different than the VSS method.

I am doing the customer part from scratch. When I complete the model I will remove the imported customer model so there are no dependencies.

TomD.inPDX
17-Peridot
(To:pimm)

Boundary blend is a different animal. It has some very nice features least of which is control over the blending with neighboring surfaces similar to the VSS. Then again, it can become a bit confused just as easily when it encounters a pinch-point.

Here is an extensive discussion on surfaces made with various methods with some good samples posted: Fill surface with guide curves

Patriot_1776
22-Sapphire II
(To:pimm)

Sometimes you have to try all 3 different techniques to see what works best. Sometimes you can't get EXACTLY what you want, either because what you want is impossible with any software, or because of the limitations in Pro/E. You just have to choose what gets you closest and allows you the best building block for future features.

Many times, what I'll do is pehaps make a simplified model where I can experiment with ONLY the area that's giving me trouble, find the best way to do it, and only then use that technique on the actual model.

For instance, I was doing some wireing harness routing (without the wiring module we don't have), and I tried to create the wiring bundle from the sweeps where the individual wires went into one connector to the same point for the other connector. I had a bundle of 11 wires I needed to do as a swept blend, and Pro/E will NOT let you have more than one closed loop in the sketch (THIS needs to be fixed and would be a REAL enhancement, unlike wasting time on ribbons....). So, I tried to do each wire separately. I was in "insert mode" because I wanted these features at the front, and I got all the wires done, and was happy with it. But when I resumed all the wiring at the connector ends, it all failed!!! WT#???? I tried monkeying with the absolute accuracy, and then redoing the swept blends, and/or the wiring at the ends, but nothing worked. So, I ended up having to "cheat" and do it as one swept blend, using the outside of the wires (all the visible surfaces) with a solid inner core. It's obviously not 100% correct, but you'll never see the problem because it's buried inside the bundle. Now, I COULD have given maybe .001 clearance between the circles as I think the issues Pro/E was struggling with was all the tangencies for the 11 wires and possible overlapping volumes due to rounding errors, but honestly, that should NOT matter. And, in the real world there wouldn't be gaps like that anyways (especially with the zip ties we're using on the harness), so that's an error right there. You can bury a simple extrusion inside another with no issue. Strange. But, it worked, most likely makes a much smaller file, and nobody will ever see the problam area, so I'm good with it.

I'll be putting some pics of the beast in my photo album shortly.

Best of luck!

Here's a JPEG of the routing I mentioned using a swept blend for the bundle, and individual sweeps for all the wires at the connector ends. A better (.bmp) image is in my photo album.

NGA_CHASSIS_62-8276_R01_ASM.JPG

pimm
14-Alexandrite
(To:Patriot_1776)

Frank: That is a very slick looking part.

I haven't even looked at Swept Blend yet but that's an impressive technique to build this.

It is tricky knowing which tool is the right one for the job.

Variable Section Sweep is the technique I used to get the results I showed in my picture. Perhaps Boundary blend might have been the better choice. I'll just have to see how well this all comes together.

I have been doing some experimentation. I'm sure I'll get better at picking the right tool as I gain experience and understand what everything does and doesn't do.

TomD.inPDX
17-Peridot
(To:pimm)

If you really want to give your CPU a heart attack, add a twist to that bundle, Frank

(angle of section orientation in the section) relation: sd#=0+4*(trajpar*360)

Patriot_1776
22-Sapphire II
(To:TomD.inPDX)

Yeah, I thought about that and WOULD have done it if the bundle had been a "twisted pair" type.......

I actually had a lot of fun doing this cable assy. Something about all those pretty colored wires.......

Patriot_1776
22-Sapphire II
(To:pimm)

See if either of you guys can view this as a 3D PDF made from creo. There are 3 views, one with the "spin center" centered on the big connector, another for the small connector, and another centered in the model.

I too like modeling these little details when drawings call for it.

I was able to open the PDF... not that I understand how you got the spin center to move to those "views".

I did a bit of work to make trajpar work for twisted cables. With what you said and many crashes later, I came up with an almost foolproof technique ( he says sheepishly)

Since we've strayed from the subject at hand, keep an eye out for the document

Patriot_1776
22-Sapphire II
(To:TomD.inPDX)

I think it's great that there are a few of us here that want to learn how to do things like this, to push the envelope (until it crashes!). I think we should always try and learn new techniques, so we can teach the others (that are actually willing to learn). I get very frustrated with Pro/E users that refuse to do anything more than the bare minimum to get the (typically sloppy) job done. I enjoy bouncing things back and forth between us!

We had the same trouble with the spin center's here. In fact, it was so bad after trying it a few times, we didn't like to use the 3D PDF's because they came in so bad. But, being the inquisitive and stubborn b#stard I am , I wanted to see if I could fix that, especially for assemblies like the wiring harness where the default CS is so far off to the side (all our cabling is constrained to the upper level assy's default CS for consistancy). I figured there had to be a way. I detailed the steps in a Word doc I posted in the Modeling forum about it. The pics are a little blurry (dunno why), but it works.

Let me know if that helps.

I'll be eagerly waiting your document Antonius!

pimm
14-Alexandrite
(To:Patriot_1776)

There was some difficulty in getting beyond permissions and setting up the spin, but this works well once set up.

Top Tags