Stamped sheet metal part

[Ken]

What's a better way to digitally model this?

I made a Boolean assembly of extrudes and sweeps. Problem is that the material between the two basic forms is not consistent. Nor the desired result. Might not even be achievable. Or way too expensive.

It's supposed to be a sheet metal part that's stamped or cold forged onto underlying objects that are basically a cylinder on its side and a cylinder on its end. Alternately, with some tweaking to the digital model, it can be injection molded plastic.

Hoping for some way of "draping" a consistent-thickness loft onto it while maintaining perfect horizontal alignment.

[Benson Shaw]

Experiment with the shell tool. Start with surfaces.

-B

Edited December 28, 2013 by Benson Shaw

[Kevin McAllister]

Or try the free beta of Rhino for Mac. I've been playing with it for objects that need to be very specific and are not easily modelled in VW. The new import/export to Rhino is actually very good and you can bring this type of object back in as a generic solid.

KM

[Ken]

Shell tool yes, but first need an easy way to define the NURBS surface at the transition.

I vaguely recall some way of using Digital Terrain Model (DTM), which is now called Site Model? There used to be a way of vertically projecting downward to 3D points thereby producing a "ground cover" which hopefully is just a NURBS surface. I did not explore it back then when I heard about it. Anybody know?

I've downloaded Rhino 5. It's interesting -- a whole new set of tools to learn.

[Ken]

Here's where I'm stuck. I have three NURBS surfaces. They all touch. I figure to combine them somehow into one, then manually move a 3D vertex on each side for added strength (at the connect between the two basic forms).

Then use the Shell tool.

But how to combine NURBS surfaces into one NURBS surface?

Object is turned upside down for clarity.

[mike m oz]

Have a look at the 3D Power Pack Tips:

http://www.vectorworks.net/3Dpowerpack/tips.php

Also explore the other 3D Power Pack pages including the sample models for download.

http://www.vectorworks.net/3Dpowerpack/samples.php

[Ken]

I've also tried a shelled loft as connector piece. But Add Solid doesn't work to smoothly combine with the other pieces. It's 3 CSG solids.

The bottom is also still messy.

[Benson Shaw]

Hey, Ken - Looks like you added a transition component. I think there are several approaches, but here is one way to adjust the existing objects.

Create some new objects to use for trimming with Solid Subtract.

Create a solid version of the pan (or maybe just the inside of the pan), ie a conical solid, similar to a solid rubber sink drain plug. Subtract this from the tongue and the transition to remove the projecting pieces.

If needed create a solid version of the tongue, ie prox a half cylinder, to trim bits from the pan and the transition.

If there are still bits of the components projecting, make a solid volume of the transition shape and subtract that from the other pieces.

Vectorworks should Add Solids if all the projecting parts are removed and the 3 components are completely engaged.

-B

[mike m oz]

Ken, can you please post a drawing showing the sizes and shape of the objects and their spatial relationship. One of us might then be able to post a file showing you how to achieve what you want.

[Kevin McAllister]

VW does have a tool to create the transition component - Create Fillet Surface. Its one of the most clunky commands though because it provides almost no feedback and mostly fails. I've attached an example. Note that I made the dish part much taller to run the command and trimmed it after. The surface normals (the red arrows) need to be aligned.

The harder part is combining all these surfaces and getting it to shell.

Kevin

Edited December 30, 2013 by Kevin McAllister

[mike m oz]

Ken, I've had a go at modelling a similar object. The attached file contains 1 layer for each step (10 in total).

There is probably a quicker and easier way to do this and hopefully other users will post their method.

[Ken]

YES!

Fillet Surface is indeed the tool.

Good step-by-step, Mike. Good suggestions, Kevin et. al.

[Ken]

Mike,

How did you apply "Fillet to join surfaces" in your Step 4? It's not the Fillet Surfaces tool?

I had to do the steps in a different order. Fillet Surfaces require NURBS surfaces, so I had to convert the solids first. Then chopping off the the conical shell from above left me with no top piece. So I had to add back a sweep object.

But problem is that the upper outer edge of the conical shell won't take Fillet Edge, before or after the Solid Add.

The transitions are wonderful though. Just some minor surface blemishes. OpenGL set to show edges for emphasis. Not sure if my requirement of the top surfaces having to be horizontally aligned makes this form impossible as a singular resulting object.

[Kevin McAllister]

Ha, that's where I got stuck with Mike's example tutorial too. Then I realized he added the two solids together first. If you do that, then the fillet is created with the 3D fillet tool (not fillet surfaces) in edge mode (in the tool's settings dialog) on the resulting solid addition.

Kevin

[Ken]

The problem with Fillet Edge on a solid (or on a Boolean addition) is the seemingly hit-or-miss transition at the lip for this particular form. Here's the closest I got.

Then, at the end, it still doesn't take Fillet Edges all around. The most important edge to fillet is the top of the conical platform, to get a big fillet. All other edges just need to get a micro eased edge. Attached is the VWX file (disregard scale).

[mike m oz]

How did you apply "Fillet to join surfaces" in your Step 4? It's not the Fillet Surfaces tool?

In the comments I forgot to include Use Add Solid to create a single object

A corrected file is attached.

Edited December 31, 2013 by mike m oz

[Kevin McAllister]

The problem with Fillet Edge on a solid (or on a Boolean addition) is the seemingly hit-or-miss transition at the lip for this particular form. Here's the closest I got.

Then, at the end, it still doesn't take Fillet Edges all around. The most important edge to fillet is the top of the conical platform, to get a big fillet. All other edges just need to get a micro eased edge.

Ken, the only way I found to get the smoother transition you're looking for was to make the "hat" part taller to allow for a bigger fillet radius along the join between the two pieces. See my example image of fillet surfaces above for what I mean. I think its about how the fillet command works and may not be solvable.

Kevin

[Ken]

Probably not the "hat" part taller. I've just now tried a much larger radius for the fillet.

Seems the issue is more like the bottom positioning of the cone solid ? specifically how it aligns with the edge of the other object. Or not align.

I think you're right. This 3D geometry may not be solvable in Vectorworks.

[mike m oz]

Ken, I've had another go at modelling this using your objects with a simplified process. This time I did the fillets in 2 parts using a variable radius fillet (2" to 0.01").

I can't figure out how to get a greater flare and a smoother transition though.

[Ken]

Interesting, Mike.

I'm assuming you did a variable radius fillet using a secret formula? I've tried it again too. Sometimes I could not get the upper flat circular rim to take a fillet. Sometimes I could. Here's your Step 4 which takes the fillet okay. It looks decent.

But while that upper fillet is important, the transition "flare" is more important.

I'll have to go back and take my chances with Fillet Surfaces.

Here's a silly attempt at a hybrid or "half this, half that" ? also known as a Frankenstein (stitching two dead pieces together).

[Benson Shaw]

Flogging the dead horse, but totally fun modeling challenge.

Without dims or other parameters, I generalized that goal is transition which is tangent to both tongue and pan in both top and side view.

Thought I had it, too. But the shell of my transition area ends on a different angle than the shell of the cut pan. So the relative thickness and edge positions do not match.

This should be way easier in Vectorworks.

-B

Edited January 1, 2014 by Benson Shaw

[Ken]

It's a good exercise, Benson.

Wouldn't it be funny if the goal was just to sketch a nice looking frying pan with a nice round hole? :crazy:

Actually, the cost of an injection mold even at prototype stage can exceed twice the price of full retail Vectorworks. Or get cut in half with good clean modeling. I can't give up yet.

[Benson Shaw]

I was thinking shoe horn, but skillet it good. I will think about cutting a hole in my cast iron in every future use.

Sometimes it's faster and easier to just sculpt the pattern with clay, wax, wood, plaster, whatever, then mold.

For stamped metal die work, the model accuracy requires eased or radius "joints" throughout (fillets/tangency), and some stretch/thin factor from the press process. Yikes!

That suggests another Vectorworks model path - model the die, then drape or extract the surfaces and trim to final shape.

Or use modeling software that is focused on this kind of object.

-B

[Benson Shaw]

Tried it with the Drape Surface.

The edges crinkle some, but the process works.

I also tried it with a big fillet on the die, btw cone and cylinder. That makes the tongue flare, so may not be desirable.

Final transition can be eased by filleting the 2d outline.

OK, Happy New Year Everyone!

-B

[mjm]

Another exciting round for all of us watching from home!

Thanks guys!