Replies to This Discussion

Permalink Reply by Lowell Walmsley on December 18, 2012 at 2:02pm

That doesn't sound good.

Is it possible you can give me some curves so I can see what's going on in that case?

It sounds likely that its because of some other difference that you get different answers, but the only way to know is to look at the example.

Thanks,

Lowell

• ▶ Reply

Permalink Reply by Lowell Walmsley on December 20, 2012 at 4:56pm

I think I see what's going on.

The bottom is pretty flat so the rulings can go pretty much anywhere.

With the flat area, the locations are probably being influenced by numerical noise in the calculations.  There may be a difference between v4 and v5 as far as choosing between very low twist locations.  I'll look at that.

This is a good example to see what's going on.

Run DevSrf and pick the centerline then the inside chine.

Set Ruling density to 1 and Twist tolerance to 0.7

With Least twist selected, all of the rulings are green, meaning that they have less than 0.7 deg twist.

Check Show ranges and you'll see that the fans between the blue lines for the rulings on the flat part are very wide.  Those are the limits of where the twist is under 0.7 degrees.

Change Twist to 0.6 and the ranges get smaller at the curved part but don't change visibly at the flat part.

Change to 0.5 and one of the rulings in the middle of the curved part turns red, meaning it's twisted more than 0.5 deg.

Change from Least twist to Shortest and see that the rulings in the flat part go straight across between the rails and stay green, which means that they are able to achieve 0.5 deg twist at the shortest distance between the rails. The rulings in the curved part all move as far as they can within the range fan to try to get to the shortest distance between rails.  They can't get to that and still maintain 0.5 deg twist so they turn red.

So for that surface, with Shortest selected, you can get all of the rulings to under 0.7 deg twist.  If you choose Shortest, you get the complete bottom surface. When you make that surface, you get the folded over part that got us here in the first place.  I'll work on that.

Lowell

• ▶ Reply

Permalink Reply by Lowell Walmsley on December 21, 2012 at 3:11pm

Here's a description of what Twist means in the DevSrf dialog

Attachments:

• DevSrf Twist.3dm, 64 KB

Permalink Reply by Lowell Walmsley on December 14, 2012 at 9:54am

Dennis,

When you nudge control points until an acceptable surface appears, how do you evaluate when the surface is acceptable?

Lowell

• ▶ Reply

Permalink Reply by Dennis Harjamaa on December 14, 2012 at 11:50am

Hi Lowell,

I try to use the smallest possible number of control points in the two edge curves.

If a surface actually appears it is good. Most of the time after picking the two edge curves using Loft I get a horrible faceted mess. Very small nudges to the control points will make a nice looking surface appear. The nudge setting is only 0.1mm or less. For me getting the History command in V4 was gold! In V3 this process took ages.

I evaluate the results by comparing the surface areas, edge and girth lengths of the 3D and developed panels. These tend to match perfectly.

Dennis

• ▶ Reply

Permalink Reply by John Taylor on December 14, 2012 at 4:18pm

Hi Dennis

I like the idea of using history with loft I'll give that a try.

I'll use both loft and DevSrf depending on the amount of curvature and what needs to be done with the surface.

The issue I've have with loft is the poly surfaces created. Often hull surfaces we work with are offset, trimmed or split to fit the plates together with the desired thickness.

John

• ▶ Reply

Permalink Reply by John Taylor on December 20, 2012 at 11:54am

Hi Lowell

This is the original set of 3D Curves in V5 for the skiff, before the sheer line was lifted 50mm. 

The back (stern) of the surfaces would be trimmed off afterwards

John

Attachments:

• 3D Curves for Skiff.3dm, 42 KB

• ▶ Reply

• ‹ Previous
• 1
• 2
• Next ›
• Page