Re: [CAD_CAM_EDM_DRO] Re:RE: Propeller blades, was: 4-AXIS FOAM CUTTER

> > Date: Sat, 9 Jun 2001 15:50:25 -0600
> > From: "Tim Goldstein" <timg@...>
> >Subject: RE: Propeller blades, was: 4-AXIS FOAM CUTTER
>
> > I want to machine what I would describe
> >as the front face of a propeller (one blade only). Out of your arsenal of
> >programs what software would you suggest to do this. What the item actually
> >will be is a prop blade form for a model air plane propeller that is then
> >molded on the face of the from a balsa sheet. So the form must have both
> >camber (an arch from the leading edge to the training edge) and it twists
> >from one end to the other just like the face of a propeller.
>
> Vector
>
> If you copy paste one offset cross section with simultaneous multiple displaced copies,
> rotation, and scaling, you will get a series of cross sections that fits your
> definition pretty well. If you then reverse every second and connect, you will
> have a tool path. (offset the first cross section by one radius, before the
> Copy-Paste). That's probably the most simplistic approach(and may work
> very well for your purpose). You could also define several cross sections
> (possibly in an aeronautical design program), drawn precisely, connect them with a
> spline or linearly, and then loft a skin surface over the wireframe form.

Software such as this may be a very good solution. But as I've lately
had to fix a lot of problems with such programs, I find that there are
often limitations and end up getting tempted to write something from
scratch just to make this shape.

Tim proposed a fairly simple shape, having only camber and twist,
without any rake to the blade or taper in its chord. What I would do
is either fit a curve to the airfoil, or express it is a finely
detailed table of points. Then I'd decide on the point about which it
twists. I'd write an equation to twist it as a function of position
along the length (optionally, one could also taper the chord and sweep
it at this point). This gives a collection of points defining the surface.

The collection of points then has to be converted into a toolpath for
a ball mill. The problem is easy, because being convex we known the
mill will only be tangent to the surface at one point at a time.
Decide if you are going to travel along the chord stepping over in
span, or vice versus. At each point on the surface, calculate tangent
vectors in X and Y by taking the derivative (slope) to the next point.
Cross multiply to get the normal to the surface. Offset up the normal
half the ball diameter and you have a point on your toolpath. (It
should be noted that it is the cutting point which makes a raster-like
path in this approach - the mill will traverse on slightly curved
paths in all axis, so this is not good for a machine with backlash)

One of the things I like about this approach is that you can easily
change one of the parameters - say twist, or the basic shape, and view
the result. Depending on how a cad package is written, it may be hard
to do this if you've lofted a skin over a bunch of ribs entered either
seperately or derived from an original shape.

As for which approach is better, it all depends:

If you have software (or money for it) that does what you need
satisfactorily and you know how to use it, do so. This approach will
also let you use the software's built in procedures for roughing.

If you can describe what you want precisely, remember a little vector
geometry, and are comfortable programming in any language, you can
have fine control over the results that way. But you'll have to come
up with a solution for roughing on your own, such as running the
program multiple times with a simple Z offset (and perhaps
automatically cancelling out the part of the toolpath when you know
you would not be cutting anything).

Also, it may make sense to fixture the part on an angle (maybe even on
its side) so that you never need to cut with the tip of the ball
mill.

Chris

--
Christopher C. Stratton, stratton@...
Instrument Maker, Horn Player & Engineer
22 Adrian Street, Somerville, MA 02143
http://www.mdc.net/~stratton
NEW PHONE NUMBER: (617) 628-1062 home, 253-2606 MIT