Hi Michael,

>Sorry, no there isn't any automated way to take a radial arrangement of curves and move them
> along a straight line path.

Could scripts be helpful in this? I have not written scripts for Moi yet but I'm willing to give it a go if that may ease the process.

>Could you possibly give some more details on your equipment and your overall plan of how you
> want to produce your finished design?

My machine has 3 axis to move the tool in space: X (left/right), Y (front/back), Z (up/down) and a fourth one, A axis ( 0-360deg) that rotates the stock.

In my setup,the stock axis is oriented parallel to the y axis, the X axis locked with the tool right above the stock axis. And the x axis info in sent to the rotating axis. The machine is set so the one rotation correspond to the linear path (object 4).


>the distances between different peaks and valleys in your shapes is different in the "straight"
>one versus the 3D one, that produces different shaping when you loft a surface through the
> straight one for example.

As the distance between the 3d ogject surface and the center is identical to the distance between the unwrapped surface and the center of the stock it seems to me the two should be identical, no? Of course more Intersection Lines would produce a closer lofted surface.
But you are probably right. I will need to cut a part to put my head around this one :)

Gaston


Gaston

Hi Gaston,

> Could scripts be helpful in this? I have not written scripts
> for Moi yet but I'm willing to give it a go if that may ease
> the process.

Well, in theory yes, but I'm afraid that it would be quite a difficult task to produce such a script.

It's basically like you would be trying to create a toolpath generator. It's just overall the job that I would have expected a CAM program to be doing for you.


> In my setup,the stock axis is oriented parallel to the y axis,
> the X axis locked with the tool right above the stock axis. And
> the x axis info in sent to the rotating axis. The machine is set
> so the one rotation correspond to the linear path (object 4).

Ok, I think I understand what you are talking about a lot more now, thanks! :)


> As the distance between the 3d ogject surface and the center is
> identical to the distance between the unwrapped surface and the
> center of the stock it seems to me the two should be identical, no?

It's not so much about the distance to the center of the stock - it's about a difference in the distance between portions of each profile curve.

For instance when there is a depression, each projected section will be close to one another in 3D, but when you arrange these same sections in a linear progression, the spacing will be increased there.

How things are spaced is going to have some effect on the lofted result. So that's why you're going to have some difference if you try to use those linear arranged pieces to build a new loft surface through them. You'll probably tend to get different kinds of lumps and bumps in the shape because of that...

But possibly it will give you some kind of close enough result, I don't really know.


Ideally I guess you would never build a new loft surface at all and just take a whole bunch of radial profile curves from the 3D shape and use those as your toolpaths - but I always thought that is more or less exactly the kind of task that a CAM program is supposed to do for you?

- Michael

Here is a screenshot of a cam package that does 4th axis wrapping. On the right, a 2 inch radius cylinder. The light blue is the cylinder unwrapped. Its Y value is not "4 inches". it is 6.2832.



Here it is in top view:




I can draw a circle and generate code and cut the circle:



but the software takes my "Y" values and converts them with this:

conversion axis*(360/(2*pi))/radius

If I wanted to manually draw the same circle to feed to my 4th axis, it would look like this:



You would have to run a conversion on all your line geometry that would generate a "Y" move to create the proper shape.

Michael,

>Ideally I guess you would never build a new loft surface at all and just take a whole bunch of radial
>profile curves from the 3D shape and use those as your toolpaths - but I always thought that is
>more or less exactly the kind of task that a CAM program is supposed to do for you?

You are right, as far as I know, the CAM programs that support 4th axis would analyse the 3d object and the tool path but those CAM softwares that support 4th axis are way over my hobbyst budget.
My CAM program, as explained by BurrMan previously, is a "four side machining". Each side is machined in a raster fashion. The tool moves at constant speed across and within the boundary of the part and after each pass the other axis is incremented. During the whole process, the tool follows the peaks and valleys of the object.
One option is to machine the top part with normal xyz operation, rotate the part 90 degrees, machine the side then repeat for the bottom and the other side.

Colin told us that the Flow command in Rhino4 does exactly what I want to do (see the tutorial attachement to Colin's message 2654.8)
As Moi3d does not have such a command, my idea is to draw the object in 3d, then modify it so the CAM would "think" it is machining one side of the part while in fact the object would be rotated and the entire machining would take place in one operation.

Gaston
My idea was inspired by Julian Beever's sidewalk painting.
http://users.skynet.be/J.Beever/dungeon.html

Hi Gaston, what about this one: http://www.imsrv.com/deskcnc/

It sounds like it handles 4-axis rotary type stuff...

It looks like the software is available for $250:
http://www.cadcamcadcam.com/software-deskcnconly.aspx

There is a demo version: http://www.imsrv.com/deskcnc/dcncdown.htm

It may be worth checking out.

- Michael