I’m modelling a custom end-effector holder for a robot arm. The form is intentionally quite organic — I wanted a certain flowiness to it while still having enough lateral rigidity in the silhouette to actually hold a device securely. Hence the shape.
I’m trying to surface it with NetworkSrf but I just can’t get it to work. I know my curve network isn’t perfectly clean everywhere — some intersections are a bit rough — but as far as I can tell the curves do touch at all crossing points. Still, the command either fails or gives me nothing usable.
If anyone has tips or tricks for working with NetworkSrf on geometry like this, I’d really appreciate it. And if you spot any obvious mistakes I’m making beyond the curve cleanliness issue, please feel free to point them out — happy to learn. Also open to suggestions on how to best check whether a curve network is actually clean enough before running the command.
You need to actually learn NURBS modeling, start with the Level 1 and 2 training. “Make a bunch of curves and somehow network them” is not any sort of proper workflow.
NetworkSrf is a polarizing tool - it seems like it should produce all kinds of incredible organic smooth surfaces - and it can, but only when wielded in a very specific and constrained set of ways that ironically have some very counterintuitive requirements. I’m not an expert on it why NetworkSrf works the way it does (it’s mysterious and is probably better learned without an end shape as the target (your framework is probably 10x the number of curves that can practically be expected to work as inputs)).
What Jim says is true, you’re way better off using the standard surfacing toolchain. Check out the training guides and the official McNeel Youtube channel and also thirtysixvert’s primary surfacing video tutorials.
From the curves shown (maybe even fewer) I think you’ll be able to achieve everything you’re after with sweeps and lofts, maybe some blends.
In absence of good teachers, self-studying industrial designers should learn NURBS surface modelling with Rhino’s Level 1 and Level 2 training, learn about all-important patch layouting approaches, and learn to understand basic concepts about NURBS curves and NURBS surfaces.
The network surface is known as notwork surface for a good reason ; )
One of the reasons why the NetworkSrf command fails is due to curves that do not properly intersect or touch each other. Another common issue is a polyline that hasn’t been joined beforehand, meaning you’re not actually selecting the curves correctly during the process—along with other possible causes.
Hi @Leonard_Elias_Böker
As it has already been pointed out, there is no way something as complex as this can be created in a single operation - it’s just not how NURBS work. Also your curves leave quite a bit to be desired in terms of precision and there’s also a few places where the curves are quite inconclusive.
Attached is a quick version that might at least get you started (just the main body). It is by NO means perfect, since the input curves are not quite there, but hopefully it’s a starting point and will give you an example of how a thing like this can be made. What I would do is to start over and make all the planar surfaces and true fillets first and then construct the transitions and blends from that.
First of all, I am always amazed by how helpful everybody is in this forum. A special thanks to @JimCarruthers, @Normand, @3Drawing_net, @Lagom, @Zaxxon, and @brvdln for reacting, commenting on, and even modeling my issue.
I managed to solve it myself by splitting the model into basically two segments (or halves) and then lofting the remaining open edges. After that, I used several Sweep2 commands to blend the openings, followed by some fillets.
While I know this isn’t a “proper” industrial design approach—and would need much more refining to be considered clean geometry or a professional workflow—my goal for now was simply to model a workholding tool with a more appealing appearance.
It is printing at the moment! This tool will be used in two weeks on a UR10 robot to felt geometric patterns into fabric to create foldable structures. If you are interested in the outcome, I’ll post an update here with some pictures.
I plan to look further into SubD modeling and cleaner workflows with NURBS curves to ensure a better output next time.
