From: james on
I am attempting to model a helical cut on a shaft.

That is, a cut like the spiral that you see on a barbershop pole. But
rather than just a painted red stripe, I want a groove that's the width
of the stripe, and cuts into the cylinder. An example would be a 1/2"
diameter cylinder, with a 1/16" endmill helical cut going around it,
just like the barbershop stripe.

I am coming to the conclusion that it is not an easy thing to do at
all.

Has anyone accurately modeled a helical cut on a shaft?

I now believe it's not as simple as sweeping a rectangular cross
section. This seems right, until I try to model the entry and exit
points of the endmill. These entry and exit points clearly show
there's a problem.

I've tried making the sweeping cross section be tangent to the path of
the helix, and all seems well until modeling the "start" and "end"
positions of the endmill. If they walk off the ends of the barbershop
pole, the error isn't noticable. But if you try to start and end while
still on the barbershop pole, the error is obvious.

Pretty interesting stuff, and frustrating!!!

I read some old posts on a mechanical desktop or something or other
forum, and people were claiming it's not quite possible to do. I'm
beginning to agree with them.

From: james on
Here are some pictures of the endmill exit point I was refering to:

http://www.precisionmechanicaldesign.com/images/helicalcutexit1.jpg

http://www.precisionmechanicaldesign.com/images/helicalcutexit2.jpg

From: Jeff Howard on
> Has anyone accurately modeled a helical cut on a shaft?
>
> I now believe it's not as simple as sweeping a rectangular
> cross section. ....

I think you are correct. Planar sections won't work and the reason is seen if
you consider the "pitch angle" (cutter tangent) for a given pitch helix at
different radii. At the work piece axis (zero rad) the angle is zero, relative
the axis, and moves toward 90 with increasing radius.

The best deal I've found is to model a "ribbon" surface (by what ever means is
available to you) that would represent the trace of the cutter axis. If, for
instance, you use the work piece axis as a sweep trajectory, the helix as the
section X vector control, and section (profile) plane defined as normal to work
piece axis; you can than sweep a line connecting the axis and helix to produce
the desired surface. (You'd actually want to sweep a portion of the line or
trim the surface at groove diameter.) Some "loft" or "blend" functions appear
to work as well, I believe. Depends on how they map the input curves (axis and
helix). Wrapping a curve on cylinder and then surface normal "pulling" it to
groove dia can also give you a good set of curves for a loft type function or
maybe sweep. Doing a symmetric, surface normal thicken of that surface should
produce an accurate groove. Groove walls should be, I can't remember if it
looked like a good rep of groove bottom as cut by an end mill or not. Ball end
would make things just a little more complicated, but I think combining a swept
circle cut and the thicken might work.

(Ya know, I've never actually heard this from a machinist but think what they
will really be interested in is a curve that describes the intersection of the
cutter axis with a cylindrical surface of some (any arbitrary) diameter. I'm
guessing they'd be happier having that curve than trying to reverse engineer it
from model surfaces. More important if the curve isn't something as simple as a
helix (?).)

From: John Layne on
james wrote:
> I am attempting to model a helical cut on a shaft.
>
> That is, a cut like the spiral that you see on a barbershop pole. But
> rather than just a painted red stripe, I want a groove that's the width
> of the stripe, and cuts into the cylinder. An example would be a 1/2"
> diameter cylinder, with a 1/16" endmill helical cut going around it,
> just like the barbershop stripe.
>
> I am coming to the conclusion that it is not an easy thing to do at
> all.
>
> Has anyone accurately modeled a helical cut on a shaft?
>
> I now believe it's not as simple as sweeping a rectangular cross
> section. This seems right, until I try to model the entry and exit
> points of the endmill. These entry and exit points clearly show
> there's a problem.
>
> I've tried making the sweeping cross section be tangent to the path of
> the helix, and all seems well until modeling the "start" and "end"
> positions of the endmill. If they walk off the ends of the barbershop
> pole, the error isn't noticable. But if you try to start and end while
> still on the barbershop pole, the error is obvious.
>
> Pretty interesting stuff, and frustrating!!!
>
> I read some old posts on a mechanical desktop or something or other
> forum, and people were claiming it's not quite possible to do. I'm
> beginning to agree with them.
>


Dr J.D Mather Has some very good tutorials that may be of use checkout 5b.

http://home.pct.edu/~jmather/content/DSG322/solidworks_surface_tutorials.htm

John Layne
www.solidengineering.co.nz
From: Dominic V on
James,

Try performing two extruded cuts simulating the plunging of a slot
drill (you can't plunge end mills in the real world) into the start and
end of the helix. This may be difficult to set up, but you should be
able to work it out. Then perfom a sweep along the helix with a
rectangle normal to the helix like you were before. These "plunge
cuts" are good machining practice, and should provide the runout you
need at the end of the helix to overcome this problem. It is something
that you don't notice in the real world and most machinist would not
know about it, as it cannot be distinguished from normal tool chatter
in smaller sizes.

Dominic V.