Dynamic Component problem

I am trying to create a 2D symbol for a bath as a DC and I have nearly cracked it. But not quite.

The geometry seems simple enough but for a DC you have to break everything down as you only want certain things to be scalable. In this case, the outline of the bath “bowl” is separated into radiused corners, a radiused end, and an outline. It is the outline that is causing the problem as it does not scale correctly when you move away from the base size. Note that it has some hidden lines in it.

Whilst the defect does not matter much for this model, I am interested in understanding the principle so that I can get it right in more complicated cases. It may be that I have not constructed the model in the optimum way or that my relational data could be improved.

I know there are some much cleverer people than me out there who will probably be able to advise. Here is my model:

Bath.skp (213.4 KB)

It’s the way the edges of the Outline component are getting resized that is causing the gaps. The arcs for the corners are fine. Or do you want the arcs to scale, too?

I don’t want the arcs to scale. In fact, the only things I want to scale in this model is the rectangle that defines the outside edge of the bath and the group called Outline that defines most of the inner edge. This group is also a rectangle except that I have hidden some lines so they don’t show. If it was a simple rectangle, it might have worked but then it wouldn’t have looked right.

What’s the fix, then?

Maybe I should make the space between the arcs a simple rectangle with three hidden lines and have a separate rectangle with hidden lines for the rest of the bowl?

Scaling that Outline rectangle means you’re scaling the length of the line segments within so the hidden ones are getting scaled along with the rest.

How about making the outline just four line segments without the hidden corners?

Solved it with the idea in my last post. See here:

Bath.skp (93.5 KB)

But maybe your idea could be simpler still? Have you tried?

No. I’ve been working. As long as you’ve got it solved, though…

Just tried your way with three lines (two joining the arcs to the semicircle and the other joining the two arcs). That doesn’t work as the gaps at the ends of the notional rectangle get scaled and therefore leave gaps if you scale up.

Maybe the solution I found is the only way.