You’d think a nose plug for swimming would be easy.

Short version: Nope.

You wouldn’t think that designing a simple nose plug for keeping water out of your nose while you’re swimming would be hard at all.

You would, sadly, be terribly wrong.

At least in my experience, even with a physical example of the thing sitting in front of me able to be directly manipulated, calipers in hand, it provided more than a slight challenge to reproduce.

Now — I make no claims of being an expert 3D modeler. I really came into the field as a whole once I discovered parametric modelers were available to any Joe on the street and, by some whim of twisted mutation, I find it far easier to get a grasp on describing what I want parametrically to a computer rather than most of the surface modeling or volume modeling systems that I had been exposed to before. I am not God’s gift to modelers. If you’ve read any of my previous bits on 3D modeling, your very aware that I have no illusions there.

Most of the things that I’ve designed over the last couple of years have been new creations. While they may have had parts which were intended to operate and interoperate with already extant physical items (my car cup phone cradle, for example), they have been inherently relatively unlimited in their inherent designs.

Last week, my girlfriend (legitimately an artist, unlike myself) pluckily returned from the YMCA, slipped off her nose plug, and plopped it on the desk in front of me.

“I wish that we could just make these things at will. They’re always getting lost, getting stolen, falling into the pool… You should see if you can design one of these. Like this.”

I looked a bit dubiously at the thing. “You know, this has a lot of really organic curves and it might be easier to do if I started with making a photogrammetric 3D model of your nose…”

“Nope, just build the thing you have in front of you.”

And that’s when it all went sideways.

Because she couldn’t have an easy one to model, oh no.

No, no, she had to have one of the Speedo Liquid Comfort Nose Clips (whose name almost compels me to follow it with a chirpy “by Mennen!”).

I wish that they actually had a 3D model of that thing on their page, not the least reason being that it would’ve saved me a good week and a half of banging my head against the wall trying to get it right, but also so that you could see that it isn’t a simple one piece thing that sweeps down and around but effectively is composed of two compound curves on separate planes which emerge from the side of the circular clamp end at a tangent. It is, not to put too fine a point on it, annoyingly complex — and that’s before you actually notice that the interior edge that is intended to lay against the nose is flat to that expected surface.

And it’s even more complex than that if you want to start noticing that there is a complex distortion in that plane which means that both arms are curved into directions.

For the sake of sanity, I will advise that you don’t notice that last — because that’s what I did.

Let’s take a look at my third or fourth stab at this thing. (Yes, third or fourth. I did mention it balked me at every turn, yes?)

This was an attempt which was born out of a desire to make more use of the Patch environment in Fusion 360. If you haven’t played with it, the Patch environment is where you create and manipulate surfaces rather than volumes. Still using parametric methodologies, but affecting surfaces in the geometry thereof. That sounds like a great fit for this problem and ultimately ended up being a key part of my eventual solution — but this was not there yet.

Notice that, most glaringly, the lines which define the sections that go over the arch of the nose are parallel to the floor rather than parallel to the expected surface that would be beneath. That was a side effect of the fact that I started with a top down view of the piece, drew the expected arcs, took them into the Patch environment, and then extruded them orthogonal to the plane on which a drew them. Unfortunately for me, that plane was along the y-axis, so the extrusion ended up being along the x-axis — flat to the floor. Once I had the surface extruded, I used Thicken in order to turn them into volumes and then looked at what I had wrought with disgust and horror.

The general shape is right, and I captured the idea of the split arches fairly well, but for the purposes intended, that is to sit on top of the nose and keep it closed, this would be terribly uncomfortable. Trying to use more traditional Boolean intersections in order to get the side profile right was an absolute mistake, and while it would have been more successful to have thought about creating a plane offset and angled to represent the cross-section of the nose and then to have used that as a way to carve out the underside, ultimately what we’re looking at is something that would’ve been way more complicated than it needed to be.

Now let’s look at something a little better.

That is a lot better, and not just because it’s a better render. You can see that the inside of the sweeps are actually orthogonal to the line, which is what we want. The bars themselves smoothly connect at a tangent and blend into the body of the nostril pinch itself, and overall it just looks a lot more comfortable to wear. There are a few rough spots when it comes to merging some of the rounded edges, particularly when the volumes merge messily at the crossover, even after a Boolean join, but this is a piece that you could actually wear.

For a change, the the step-by-step build replay doesn’t require a whole lot of discussion. It speaks for itself.

One of the things it doesn’t speak for itself is why I did the design in Fusion 360. After all, Onshape recently added the ability to use curve projections to do what Fusion 360 does with 3D sketches, essentially allowing you to make a pair of plane sketches and automatically figuring out what the intersecting curve would be if they were both projected off of their planes into one another. In theory, this is exactly what I needed to get these curves absolutely right.

In practice… Let’s just say that it didn’t turn out quite that way.

Curves as a first class object are interesting. That is, OS is not treating this extended object in space as a sketch nor as an object, nor as a surface, but in and of itself as a standalone entity which has its own entry in the navigation bar to the left. When dealing with things like projected curves, that’s really cool because you’re going to end up using Loft to connect up these 3D profiles in space in order to describe volumes.

Unfortunately for me, I really just don’t have a grasp of how the intended workflow is supposed to integrate with projected surfaces. OS has some procedural/process-related issues with managing surfaces into volumes which often make it quite unintuitive for myself to really grapple with that kind of construction. At least at this point it doesn’t feel natural. Maybe that’s just me — or maybe the methods for really digging in and using that stuff haven’t been solidified by the community yet.

Instead I worked it out in F360 and feel rather proud of the results. That’s not to say that I won’t continue the struggle to really understand how first-class curves work in OS, but sometimes you just have to accept that one tool is just not working for you.

That’s an important point: as creators, we have multiple tools. Even tools which are broadly similar and do the same thing often have different ways of achieving those goals — after all, if they didn’t why would we need multiple tools that didn’t differ? Don’t be afraid to put one down and pick up another even if you have to spend time learning and understanding multiple tools. I guarantee you, the effort you put in will be paid off by more fully understanding and exploring your craft, whatever that craft maybe.

In the meantime, I have a prototype to tinker with.


If you’ve enjoyed what you’ve read here today and want to support my continued efforts to bring engineering, art, and the occasional philosophical divergence to the masses, please feel free to send me a tip. Or thousands of dollars, I’m really not that picky. It’s through the efforts of you and others of like mind that content like this gets created.

Thank you for your time.