XNurbs releases a ground-breaking NURBS software

Filling/patching is a trivial application for xNURBS. We would like to show different operations/applications done by XNurbs, not just filling. (Except for very special cases, e.g., analytical surfaces, which include spheres, planes and cylinders etc., XNurbs is superior than traditional surfacing methods, i.e., it produces better results.)

Nevertheless, we took a quick look at your model. (using one single NURBS with G2 continuity - a lazy way to do the surface)

The following is the generated iges file (TT_ST_TEST_V21.IGS) and SolidWorks file (TT_ST_TEST_V21.zip. unzip it - since you have SolidWorks XNurbs addin, you can manipulate the XNurbs feature yourself.)

TT_ST_TEST_V21.IGS (214.8 KB)

TT_ST_TEST_V21.zip (274.4 KB)

However, you may need to redesign the model - your model can only be used as a quick concept design. For XNurbs, your original design is very strange.

  1. The surrounding geometries are not very good. You may redesign them.
  2. The boundary conditions do not support Curvature continuity. (we apply Curvature continuity for a test purpose.)
  3. You should split the surface into two (generate one then mirror it, which should give you a much better result.)

When providing examples, we would appreciate it if you could clean it up and only send us the related geometries, e.g., ā€œTT_ST_TEST_V21.zipā€

Are you friends with Yeti Cycles? If so, you may directly discuss with them for how to use XNurbs as they are also our user.

Did Yeti Cycles give you this model? If so, since Yeti Cycles is also our user, they may be willing to provide us with some demos for how to use XNurbs.

Pretty good.

This tool would be best used as a test-tryout modeling method where this surface later gets split and rebuilt with a logical patch. It would be really useful in that capacity, at least to me.

Could your software take a curve network and weave a single surface around all the curves?

@XNurbs Thanks for looking at the model and filling the patch. Itā€™s funny you mention that the model isnā€™t very good. I realize this and it was just a patch layout option that I tried (I have tried many). I never said it was the best solution. By the way, If it already was the perfect patch layout I wouldnā€™t have any need for XNurbs.

One of the things that is appealing about XNurbs is that not everybody is a super skilled Class A modeler. A tool like this can help people like me get better results faster and that is great.

I will start playing with this myself as I just purchased XNurbs for SolidWorks.

As an FYI, I am both a SolidWorks and a Rhino user and am probably interested in getting XNurbs for both pieces of software.

Iā€™m perfectly OK with this being for early stage concept modeling where the patch can be used to build a logical patch layout over the top.

With regards to your suggestion to split the model over the centerline, I have found that the resulting surface continuity is actually not very good in most cases. I find it very difficult to have good continuity over a mirror line and prefer to build a high quality uninterrupted surface across the centerline, but thatā€™s just me.

When do you think the Rhino version would be ready? Thereā€™s been a lot of discussion on this forum, but we have never seen a plug-in for Rhino.

Thanks,
Mark.

I would not mind pretending being a better modeller than I am. XNurbs comes to mind. :wink:

I look forward to the Rhino version.

// Rolf

I donā€™t think XNurbs can lower sales of other CAD programsā€¦
A easy one click surface patch is great, but that alone canā€™t design a product, after all
it is only a plugin.

Some want more history and checkin/checkout data-base like Cad with all points needed in be defined, for high precision future proof designing.
some want faster CAD that would visualise for rapid development like Rhino.
Plugin of XNurbs wonā€™t add feature of other CAD to Rhino, so Iā€™m not sure of XNurbsā€™ stance.

But, anyway, XNurbs hopefully would save a lot of time so thatā€™s why I personally was interested in it.
Setting up curves to make matching and fiddiling with control points is a bit of a hastle, so
couple hundred dollar investment for multiblend surface creating too was something I was looking forward to.

Anyway, if you donā€™t have plans to release plugin for Rhino that I can buy, thenā€¦ I guess well,
I just continue as beforeā€¦ not big loss for me, just thought itā€™d be a time saver.

By the time the big guys sales figure gets impacted, Iā€™m sure youā€™d feel great that you sold that many of XNurbs. Just my opinion.

Just a Degraded XNurbs Tangent/Curvature Continuous Multi Blend surface creating tool would be good for my use nowā€¦ anyway to make a ā€œXNurbs Light for Rhino?ā€

ā€œWith regards to your suggestion to split the model over the centerline, I have found that the resulting surface continuity is actually not very good in most casesā€¦ā€
I am sure if I understand this. As long as you apply G1 continuity over a mirror line, you will get an uninterrupted surface across the centerline. (I guess your comments may be based on some tools you used before.)

For the final production surfacing, if the generated surface does not fully meet your design intention, then simply add one or two internal curves. This is how the design is used for the final production surfacing. (As Kevin already stated in SolidWorks Forum, ā€œXNurbs has proved invaluable both for quick design phase surfacing for concepts and final production surfacing.ā€)

Since your original model may not support Curvature continuity, there is a simple trick to play when enforcing Curvature continuity. If you need the trick, let me know when you have a better understanding of XNurbs. (If your boundary conditions do not support Curvature continuity, do not enforce G2 blindly ā€“ in other words, the model we provided above is a bad design.)

Do you have some useful samples, e.g., a curve network etc.? We would like to produce some demos to show how to use XNurbs for different operations/applications.

While it is true that you can have an uninterrupted surface across the centerline, that doesnā€™t mean itā€™s a high quality transition.

Check out the screenshot, on the left I mirrored curves over a centerline and they are G1 continuous. When you look at the curvature graph you can see this. I find if very difficult to maintain anything better than this though. On the right you can see a single curve that is build across an imaginary centerline and you can see the curvature graph on this is completely smooth. It should be since itā€™s a degree 3 curve with 4 control points.

With regards to my original model I realize it does not support curvature continuity, but this is a reality of my world. I need to connect smooth flowing organic shapes to precisely defined analytic geometry and make it look good in the process. So while you may think itā€™s both a bad example and a bad model I donā€™t have the luxury to change this. A seat tube on a bicycle has to be round, straight and in a precise location.

I am currently working on a monocoque fork project for someone and will see how XNurbs fits in my workflow. However this is work for a client that I cannot share on a forum. If I have any workflow questions I will share the model with you privately.

p.s. still hoping to get an answer regarding when XNurbs will be available for Rhino as well. :grinning:

2 Likes

Itā€™s great Mark that you have the patient to explain to mathematicians the realities of production processes and real worldā€™s constraints.

Funny story: one time a developer of another software told me I was seeing a lots of issues with their importer because my files where really bad. I told him: ā€œgo pick a molding supplier, and ask them to let you review the last 100 customer files they received and had to work with. I guarantee you my files are better than at least 90% or those 100 files, and that your math will only work in 5% of those files. You wrote an academic product, not a commercial oneā€

4 Likes

i canā€™t believe there are over 300 posts for this ground-breaking software release with no demo. in the examples above, there seems to be some problems.

one is the representation of the surface in the zebra view. it just looks bad. If we canā€™t tell that your solution is better visually, the software isnā€™t very valuable. there needs to be useful feedback. if itā€™s really smooth, it needs to show that way on the screen.

there are no examples with a nicely designed surface. youā€™re examples all show pretty utilitarian cases.
you should be showing examples with a more sophisticated aesthetic, if your software can help achieve that more efficiently than doing it by hand.

Incorrect. The left image is G2, the right is (at least close to) G3!

Matching the curvature means to match the amplitude of your curvature.
G3 is matching the increase of your curvature graph.
Mirroring G1 transitions is always G2. But it clearly shows that G2 may not always produce nice curvature flows.

1 Like

That made my day, I love it and will print it on a god damn t-shirt! :smiley:

1 Like

XNurbs has an inborn capacity to distribute the curvature change over the whole surface. In other words, it will produce something similar to the right image automatically. With XNurbs, you may reconsider the layout and splitting the model over the centerline should be a good choiceā€¦

ā€œā€¦it does not support curvature continuity, but this is a reality of my world.ā€
I know this fact. As I said: (in this case), ā€œthere is a simple trick to play when enforcing Curvature continuity using XNurbsā€. When we have the time, we will show how to enforce Curvature continuity to the model posted by Kevin Quigley on SolidWorks Forum (his model does not support curvature continuity). If you are willing to share your real models privately, I can show you how to enforce Curvature continuity over the centerline. (I knew you did not provide a real design ā€“ your model even has two surfaces overlapped with each other.)

I can see why you guys do not want to show the production model you generated with XNurbs ā€¦

The discussions you posted on SolidWorks Forum already show our intention regarding our Rhino plugin: it just shows others that, if needed, we can easily help Rhino to overpower others for surfacing creation. Some guysā€™ attitude on this Forum did not help to speed up our progress for the release of Rhino plugin. However, we will try to make it available next month, and we will keep the price affordable for Rhino users, but it might not be available to everyone. Some Rhino users can even get it at 95 USD if we use their examples as demos.

Why would it not be available to everyone?

Follow along, Davidā€¦

I still donā€™t understand how to be eligible for itā€¦

Yes, you are correct, sorry about that. But you agree with my point that G2 does not always result in nice curvature over a centerline. This may be different for XNurbs, which I will experiment with since I purchased the SolidWorks plug-in.

1 Like

It looks to me that they want customers that are not going to be a pain in the neck. If you canā€™t follow the simple instructions or if you make it appear that you are trying to pick a fight you probably wonā€™t be eligible.

They have asked that you send them examples of what you would like to use Xnurbs for. If you canā€™t come up with a reasonable example to show then you probably wonā€™t be eligible.

What is it you think is not ā€œniceā€ about that curvature?? its a lot better than the curvature of the surfaces you supplied in your example file.

The Xnurbs guy was making two independent suggestions for improving your modeling technique. the first point was donā€™t expect to get G2 continuity from a surface that fills a hole when the surfaces surrounding the hole are not themselves matched for curvature. In the corners where the patch has to match 2 different surfaces at a single point, it canā€™t do that unless both surfaces match at that point. Your surfaces donā€™t match at the corners.

The other point was that it is a good strategy to split symmetrical parts. In general that should work better even without Xnurbs.

Personally I believe Xnurbs has already demonstrated it a lot better than VSR or Tsplines. Xnurbs has provided a reasonable solution for every example that people have posted. And most of the examples have been flawed. Thatā€™s far better than VSR or Tsplines. In all the years I never once saw anybody post a solution using VSR or Tsplines. Many times people have stated that you need VSR or Tsplines to solve some posted problem but not once did I ever see anybody actually post a concrete solution that I could open and analyze. As far as I can tell those two programs are just vaporware.

1 Like

Thatā€™s very true. I think some of the business SENA rod the Xnubs team are a bit weird and maybe working against them but what they showed itā€™s cool, useable. Iā€™d like some of that.

G

Well, part of that could be that most people make parts that are proprietary. However hereā€™s a side project that I did based on a commercial project. This was modeled in Tsplines and believe this would be very difficult to model otherwise. feel free to download the .3dm and take a look.


Head.3dm (1.9 MB)

1 Like