SUCCESS!! A toolpath cut normal to a cylinder (cylindrical cam)

Jeff, you are right. For generating the CNC code the centerline curve
on the cylinder is probably very much preferrable to the actual
geometry. But the geometry problem makes for an interesting modeling
and teaching problem, especially for designers who will get at most one
3hr machine tool class that will never be more than the most basic of
introductions. And there is some value of the model in proving the
motion of in an assembly, providing the software can do that (and you
are well aware of that problem).

 

.... the geometry problem makes for an interesting modeling

I'm totally in agreement. My primary interest in this sort of problem is
to simply comprehend the shapes, their interactions and modeling softwares'
capabilities as the understanding is applicable to other problems. I also
wanted to suggest that getting a good part might not necessarily validate a
particular modeling approach.

 

Maybe the "Prof" can analyze just how close this technique gets to

I think (and that's about as good as it gets with me) what you can do to
test the geometry is:

..Create a point on the helical curve.
..Create an axis thru the point normal to the cylinder axis.
..Create a plane thru point normal to (cutter) axis.
..Extrude a circle of groove dimension thru the groove and look at the
intersection.

I believe what you'll see (I'd have to do it to make sure; that memory
isn't playing tricks on me) is that there will be intersection curves that
approximate a long helix (?) about the cutter (which would correspond to
the graduated normal directions mentioned earlier?).

 

What I want to know is how you and/or Dr. Mather are determining if the
method is accecpible or not?

Spork, how do YOU determine if a perticular method is viable or not? I'm
also trying to keep in mind that your question may have already been
answered (or not) and that you have moved on to finish your job!
I know that in the past I have gotten an answer to my post in reply 2 or 3,
yet the thread went on for a week.

However I do think this is a facinating discussion.

Muggs

 

Hi, Muggs.

To tell you the truth I can only guess at a "proof". Methods I've used to
satisfy myself include running pins (constrained to a plane and normal to
the cylinder axis) of graduating diameters thru the groove in an assemby
while checking for interference and putting an array of pins along the
curve and checking for interference (there's a pic of this on Sean's site
in the Pro/E section).

 

LOL Devon
you just know J.D. is a well healed teacher with a report card like that.

 

....running pins (constrained to a plane and normal to

Should be noted that the pin's surface was not actually constrained to the
groove surface but a coincident sphere slightly larger than pin diameter
(and fixed to the pin) was constrained tangent to the groove surface. The
in-plane and axis perpendicular constraints in conjuction with this should
be pretty realistic.

 

LOL Devon

Go ahead. Make me spey coffee on the keyboard. 8~)

 

hmmmm
but a sinousoidal-motion cam path developed about a cylindrical body is not
a helix. Further, for the general case of a cam motion, it may have dwells
in it. This path can still be defined by a piece-wise series of motion
functions D = f(x, r, theta). Therefore...therefore...I dunno either.

You got me thinking though. Consider the rectangular sketch that is used to
mimic the cutter as it sweeps along a path. Accurate geometry would be
generated **as it sweeps** only if the normal of the sketch plane is
continuously tangent to the path tangent. And only to the accuracy of the
spline that is generated by node points, calculated from the motion
function. Does anyone know how the sweep function works internally? Can it
be assumed to be continuous (within a tolerance)?

If so, then why wouldn't this method work? Maybe it is just too darn much
work to undertake for a one-off solution.

I'm_Curly_Today

 

... but a sinousoidal-motion cam path developed

I could be wrong, but I believe this will apply for all but annular and
axial cuts and it's probably most exaggerated when annular and axial
components are equal. Have to ponder on that. Well, no; it's beyond my
pondering capacity. What I'd have to do is model it and see what it looks
like. 8~)

 

"Also it is not a sweep of one solid along a path of another resulting in
the intersection removed."
This is not possible is SolidWorks.

What I sent you was a "method", not a finished product. Since you don't have
SW 2005, you were not able to see the methods I used for this sample.

I use this method all the time. I have a Patent Pending for a unique plastic
mold, using this method.

The cut "solid" path is generated from an "x,y,z" coordinate table. In the
example I sent you, the "cut solid" is normal to the surface of the cylinder
and the profile is tangent to the curved path. This method allows you to
generate any path that is both normal and tangent.

YOU, take another look at this method.

Also, next time, be more courteous and if you have a statement directly to
me, just send me an e-mail.

Best Regards,
Devon T. Sowell

www.3-ddesignsolutions.com

 

LOL ...way to go Devon!

.....hey teacher, leave us kids alone... )

 

This is the method I like to use.

Show me an example. I was not aware that it was possible to sweep one
solid along a path on another solid with the resulting intersection
removed. Please try this problem before replying and provide example
files. You can post files at the SolidWorks forum at mcadforums.

Devon,
I apoligize if you didn't understand this post before you provided me
with an example that did not address the problem at hand. I admit I
was baiting you from the start as I didn't think it was possible to
sweep one solid cutting another in SolidWorks as you have confirmed. I
would be interested in seeing your solution to the exam same problem
for Tutorial 4b on my website that was the original subject of this
thread.
J.D.

 

And I probably should use the word "baiting". I am in the habit of
challenging my students when they are providing an answer that I "don't
think" is correct. I am basically saying, "are you sure about this,
provide me a proof before I start teaching everbody your method." I
learn a great deal from having 20 students tackle these problems every
year from 20 different angles. I tend to use the most interesting
(read tricky) problems that test whether we know what we think we know
(that includes myself). Making the challenge available to a wider
audience via the web simply increased the diversity of techniques and
increases the chances that a "best method" can be identified.

 

I also must admit the same . . . and my apologies to both Jeff Howard
and to Muggs for not being able to adequately evaluate what they have
sent me. So far -- I'll try over the next few days to find the time.
I've been out of town at a "dog and pony show" for my current main
client for a couple of days now and I'm just home.

To determine whether I've got accurate geometry I've used methods
tantamount to sighting down a gunbarrel, and that is certainly NOT good
enough -- especially knowing that what you see on screen is not always
precisely what you get in geometry. But I've gotten done what needed to
be done via rapid prototyping, and that's as much as I need for the
moment. I'm glad the thread has gone on as long as it has, and it is
indeed an interesting discussion. I have run the problem of actually
creating such a part via machining by a mold maker in Massachusetts, who
also indicated that all he really needs is a guide curve (as Jeff also
pointed out). But given the type of curve I needed, he would have to
translate that into linear motion of the end mill coupled with
rotational motion of the cam while the cutting operation is proceeding.
Without a mathematical equation to drive the guide curve that
translation is NOT all that easy.

At any rate, I have run into more than a couple of scenarios in the real
world of MACHINE design (not swoopy stuff as in product design) in which
I had to jump through some hoops with SolidWorks in order to create
geometry that I recognized would be quite easy to create with CAM
software. And that just shouldn't be. I can give another example if
anyone is actually interested.

Best to all,
'Sporky'

 

I can give another example if anyone is actually interested.

Yeah, I'm interested. My body has gone to hell, but I still like to exercise
my mind.
But do us all a favor and start a new thread, please.

Muggs

 

Thanks, Corey. (And BTW, although you're a newcomer relative to some,
your contributions to the newsgroup and to my own personal knowledge is
very valuable, IMHO, and greatly appreciated.) I have indeed
accomplished what I need to accomplish for the moment, but this does not
lessen my frustration with SolidWorks in regards to the issue of having
to go through so much to create something (accurately) which is simple
to do (accurately) with any respectable CAM software.

The application here (see
http://www.h2omarkdesign.com/img/cam-and-follower.pdf) is in the vein of
a spring-loaded firing pin (punctures a CO2 cartridge) which has the
follower pin pressed in midway through its center. The follower pin
rests on a "saddle" portion of the cut slots (one slot on each side of
the cylindrical cam) while the firing pin is cocked. To "fire" the
assembly the firing pin gets slightly rotated, which pushes the follower
pin off the saddle and allows it to be "shot" down the main portion of
the cut slot by the compression spring at the back end of the firing
pin.

'Sporky'

 

Geez, this is the 2nd thread I've started on the subject, and you want
me to start ANOTHER one? But yeah, it has become a bit tedious, hasn't
it?

 

Thank you for the kind words. I will try some things if I get a chance
today, see if we can't simplify it.

Corey

 

hmmmm
hmmmm, back at ya, pardner. 8~)
I think we are talking about a variable pitch helix, possibly with a few
revolved and extruded segments thrown in?

Well, seeing what it looks like proves its un-ponder-able without more
visualization horsepower than I can muster.

If it's of any interest....

http://img99.exs.cx/my.php?loc=img99&image=ar5z167se.jpg
http://img99.exs.cx/my.php?loc=img99&image=br5z87qw.jpg
http://img99.exs.cx/my.php?loc=img99&image=cr5z45oc.jpg
http://img99.exs.cx/my.php?loc=img99&image=dr5z21zh.jpg
http://img99.exs.cx/my.php?loc=img99&image=er5z2nrmlcutteraxis4tz.jpg

..... and I'm not deluding myself (possible) I think these illustrate why a
2D section won't (precisely) work. It's basically a line swept along what
would be the cylinder axis with the x-vector controlled by a helical curve.
The surface should represent the path of the cutter axis. The second
surface is swept along a line perp to cyl axis on that surface, x-vector
normal to the surface. It's edge should indicate the cutter's tangent
direction for any radial thrown off it.

(It does appear that the twist is greatest toward the outer portions of the
cylinder when pitch is ~.7 * circumference, which might mean something to
the mathematically inclined.)

.... If it's of any interest. Have a good one.