Radial Sweep? 5

[CPosner]

This is what I am trying to accomplish:

I have been having trouble generating a surface like shown in the attached sketch. I have a curve on a cylindrical surface and a radial line that must always stay parallel to the surface of the cylinder while sweeping. Is it possible to achieve this? If so, How would you achieve this? Possibly attach a part displaying this?

Much Thanks!

 

[DBezaire]

Re-attach sketch please

 

[CPosner]

Sketch

 

[jackk]

I'm not sure if I understand exactly what you're doing, but here's goes....

1. project the two endpoints of your curve onto the centerline

2. draw a short line between the 2 projected points

3. make the multi-section surface between the curve and the short line

do I win the prize?

 

[Azrael]

I'll have a crack...

-Use sweep, line type
-pick curve on cylinder wall as guide1
-pick center line as guide2
-pick center line as spline
-hit ok

 

[CPosner]

Jackk, your method worked for me, but I had to add two guides(connecting each end point of the curves to each other) It removed a dip in the surface between the two curves.

Azrael, that didn't work for me. I just errored out and I couldn't get it to work.

All in all, I am still not really sure if the surface I am creating is even manufacturable. i think the key is to make something manufacturable.

BTW, does Catia have a feature like SW2008 with the solid sweep??

Thanks all!
Chad

 

[Kapitan]

As is often the case there's more than one way to make this surface. I've tried some of them and here's the results. Apart from one, they looked identical but when analized there are small differences in area. If the first one is considered as having unit area then the others are:

(1) Blend - Curve, Curve Area = 1.000
(2) MSS - Section, Section Area = 0.996
(3) MSS - Section, Section, Guide, Guide Area = 0.970
(4) FreeStyle - Auto and Approximated Area = 0.999
(5) Fill - Four Boundaries Area = 1.007

When I first read this post my initial thought was that it was going to end up a ruled surface, which can be envisaged as an infinite number of equal length lines radiating from the axis out to the curve on the cylindrical wall. Being a ruled surface it is also the simplest mathematically, but when I thought more about it I realised that was nonsense - because of the cubic type spline laid on the cylinder to emulate the pencil sketch. As the spline has an inflection point the ruling lines would climb the axis and plunge before going up again - and this would result in an sharp ridge near the axis. Using those lines in construction would give an anticlastic surface of no use.

At the back of my mind had been something in one my old books: the strickling of sand done by mouldmakers when constructing the helical surface of a ship's screw propeller. A board, that could slide axially and swing on a vertical pole (DE), followed a curved template (ABC), sweeping the sand as it moved.

From Practical Geometry and Graphics (D A Low, 1929)

The Blend was the first one I tried, followed by an unguided MSS one which had bowed edges - so that it was more at ease with itself; but at the same time the bows would probably make it not practical. The guided MSS one was almost identical to the Blend - but required the construction of two linear guides. These are implicit in a Blend between two elements.

The FreeStyle Blend (both Auto and Approximated) was as expected; almost the same as the GSD Blend - there's nothing in it really. The Fill version was done to illustrate yet another way of doing it, but I'd be careful with Fill, it can lead to messy solutions. In the end it's obviously up to the user to decide which is the best - only the user knows exactly what is required.

Sweep would not accept the inflection case, as it meant a reversal of axial direction. After adjusting the height of the two internal points of the original cubic (on the cylinder) Sweep gave a result - which had a pinched area halfway up the axis and a shelf-like area out towards the cylinder wall. The pictures below compare the Sweep and a Blend made from the same given geometry. Once again the Blend gave the better surface - this time by far.

Surface made using the Sweep tool

Surface made using the Blend tool

Compare the two cases when the Cutting Planes are in this orientation...

Sweep Blend

The first solution will not always be the best - but in this case I think it is, because the Blend gave the better result. The guided MSS was a close runner-up, but needed some construction work.

These are all done in FreeStyle with identical ACA Highlight settings. Click them to enlarge...

(1) (2) (3)


(4) (5)

Area is not the only difference in all these surfaces: the smoothness can be seen using the old-fashioned waterlines method. I've used the Cutting Planes tool in FS but it can be done manually in GSD.

This picture is of the Blend (1) but also applies to (3) and (4)

This shows the bowed sides of the unguided Multi-Sections Surface

So far the Blend tool has been used at the most elementary level: from one curve to another. In this case there are no Supports to work with, but they could be constructed (green). This is a rather simple surface that we're after, I tried using two supports but there was no appreciable gain.

Photobucket has an Resizing tool...

 

[jackk]

Once again, Kapitan, a post worthy of archiving. Great lesson on suface modeling and analysis. You win the prize and a star!

 

[CPosner]

Kapitan,

Excellent Post! Very informative! I was aware of the fact the in what I was trying to achieve that there cannot be inflection points in the curve.



In this case, I am curious about a Drum/Barrel/Cylindrical cam modeling where in terms of manufacturing, the surface needs to modeled like how a cutting tool would cut that radial slot/surface.

In order to achieve inflection in the curve, I think the modeling approach needs to be different. What I have also tried is developing a curve onto the cylindrical surface and then offsetting that developed curve to the depth required and then I can make a multi-section sweep...I still feel that the surface that is modeled is still not how it will look after manufactured, but maybe close enough. I'm going to get some screenshots...


on a tangent...Kapitan, that device for the screw propeller casting sounds interesting...I'd like to see pics of that.

 

[Azrael]

Kapitan, the effort you put in is worth a star