Barrel Cam Modeling

[TomU]

I'm looking for ideas on how to properly model the slot in a barrel cam. I have attempted multipe techniques and I'm starting to spin my wheels. I have an idea that what I want to do can probably be done with the variable section sweep, but that feature is definitely not my area of expertise.

I've worked through the "wave washer" suggested technique on PTC's site (http://www.ptc.com/appserver/wcms/standards/freefull_cskdb.jsp?&im_dbkey=88651&icg_dbkey=900). The problem with this exampl is that the part is not a constant thickness (whereas my cam slot needs to be). To get a constant thickness, I changed the sketch profile to a single line, which created a single wavy surface, and then thickened it. Problem is, the sides are no longer vertical. To get around that, I lengthend the VSS section to include some extra on each end, then extruded inner and outer cylinder surfaces, merged everything together, and solidified. This succeeded in creating a constant thickness washer with straight sides.(See attached PDF.)

Problem is, thistechnique does not transfer well to the cam. The centersurface won't offset nearly as far as it would need to. Creating inner and outer curves and building boundary blends didn't work either.

Tom

This thread is inactive and closed by the PTC Community Management Team. If you would like to provide a reply and re-open this thread, please notify the moderator and reference the thread. You may also use "Start a topic" button to ask a new question. Please be sure to include what version of the PTC product you are using so another community member knowledgeable about your version may be able to assist.

[TomU]

Here is a PDF that shows the wave washer modeling.

Tom U.

[TomU]

Digging on the web, I see that solid works has the ability to sweep one solid through another. Anyone know if there is something similar in Pro/e?

The Swept Solid function allows an actual solid shape (say a milling cutter) to be swept and cut from another solid. A Cut-Sweep only cuts a sketched profile which gives a slightly different result.

files.solidworks.com/supportfiles/Whats_new/2008/English/whatsnew.pdf

[cid:image001.png@01CD0D01.232491E0]

[KrisR]

Could you do the washer as flat, do one section of the wave as a cut, then pattern the wave cut? I would think this would work.....

[TomU]

The purpose of the wave washer wasn't to create the washer, but rather explore techniques that could translate to the cam. Here is a picture of what the cam looks like right now (though not correct yet).

[cid:image001.jpg@01CD0D7F.340D0A90]

[blenk]

We have had similar issues trying to model this feature at our company too. I've tried to create it using VSS, using curves and surfaces, even using Pro/NC. I've even tried drawing a flat pattern in sheetmetal and forming it into a circle, but this really isn't the design intent I want to use. The best luck I have had is using the VSS and keeping the cut normal to the center axis of the cylinder but no matter how I've tried it the walls of the cut will twist as soon as they start to cut at an angle across the cylinder. Would be great if Pro/Creo/Wildfire could simply rotate a cutter normal to an axis following a tool path. Maybe it can but it is beyond me, I'd be very interested into what others have done for this.
[cid:image002.png@01CD0D80.D39A1D40]


Brandon Lenk
Senior Designer
TEAM Industries-Bagley
(218) 694-4128
-<">mailto:->

[JonathanHodgson]

We model our barrel cams by creating a flat pattern, then using a
Toroidal Bend.



I'm fairly sure the geometry isn't 100% accurate; but it looks about
right, and since the milling guys program it from the flat pattern
anyway, it makes it easier to include that information in the drawing.





Jonathan

[TomU]

I've been getting quite a bit of feedback on this issue. Unfortunately, every single solution yields cam slot surfaces that do not properly contact the roller. The ONLY thing I've found to properly orient the slot surfaces to the roller is to offset them from a center surface (created through the desired trajectory). I would like to be able to sweep something normal to the outside of a cylinder surface AND normal to the trajectory at the same time. The sketch section needs to "tip" as it moves along the trajectory so the cam surfaces stay flat to the line of contact the roller makes with the slot. Here are a few pictures that show what I'm talking about.

This first one shows how a roller will gouge the slot surface. The roller should only be in contact on the red line. This is what is created using a VSS, Toroid bend, graph with trajpar, etc.

[Description: cid:image009.jpg@01CD0D85.38920EC0]


This picture shows slots that were created by offsetting one through the center of the path. These are correct! I just can't use this approach for the real cam profile (surfaces become self-intersecting at the offset values needed.)

[Description: cid:image010.jpg@01CD0D85.38920EC0]


If there is a way to alter the VSS to allow the blue rectangle to tip as it sweeps around so it's parallel with the red line, we'd be perfect!

[Description: cid:image001.jpg@01CD0D8B.09C2A9A0]

[TomU]

Someone suggested using a helical sweep. This DOES tip the sketch plane to match the pitch value. Does anyone know if there is a way to use a helical sweep but tell it to follow a graph/trajectory/curve?

[cid:image004.png@01CD0D9F.8BD2EDB0]

[bfrandsen]

I have seen Creo 1.0 presentations using a graph to control the pitch of the helical sweep. Not sure if this is available in earlier releases.

Another idea could be to start creating a VSS surface with the current flat trajectory making an edge being you cam edge. Then use that edge as trajectory for the VSS that makes the solid cut holding the sketch normal to the its trajectory.

Best Regards,
Bjarne Frandsen

[TomU]

Barrel Cam Update


* Using a variable section sweep seems to be the best solution so far.

* The primary trajectory needs to be the cam profile wrapped or projected around the cylinder.

* There needs to be a second trajectory for use as the X-trajectory (under Horizontal/Vertical control)

[cid:image002.jpg@01CD0E68.1CEDDA30] [cid:image004.png@01CD0E60.C36703D0]

I am still struggling with how to correctly model the surfaces in the areas when the offset combined with the rotation will generate a sharp corner. The picture shown above is with the slot 45mm thick. It needs to be 40mm thick. Here is what it looks like when it's the correct thickness.
[cid:image005.png@01CD0E64.0025FB20] [cid:image008.jpg@01CD0E68.1CEDDA30]

I am presently attempting to build the top and bottom slot surface independently to see if I can overcome this issue. Thanks for all the suggestions so far.

Tom U.

[TomU]

Several of you have asked for more information. Here is a copy of the actual model I'm using. It has curves at the edges of all surface and the VSS shown at only 40mm thick. All the features in the model are suppressed to reduce file size. You will need to resume them after opening. The file was created in WF5 M120.

Tom U.

[bfrandsen]

Hi all
I think I found a nice solution for this Barrel Cam Modeling making it in 3 features. I am not an expert in mechanism design, so I might have overlooked something, but anyway here is my suggestion:

Use VSS to create a surface perpendicular to the cylinder. The height from the base can be driven be the CAM graph (10.00 is the start value). Dimension the surface to the depth need for the CAM (5.00) and let the other end be free of the cylinder (1.00).
[cid:image002.png@01CD0E9A.373A9130]

This give a Quilt surface like this:
[cid:image003.png@01CD0E9A.373A9130]

Thicken the Quilt and define it as a Cut:
[cid:image004.png@01CD0E9A.EC9088A0][cid:image005.png@01CD0E9B.437BA9B0]

This gives a Dihedral Angle at the bottom of the groove close to 90 degrees.
[cid:image006.png@01CD0E9B.B08C0AE0]

And a constant groove width:
[cid:image007.png@01CD0E9C.912CC5D0]


/Bjarne

[TimMcLellan]

Here is also a copy of the sample I sent Tom yesterday. This example also
leverages "approximate" copies for the origin & x-vector of the VSS to
reduce potential self intersecting geometry.



I had also include a "Full Round" in the model to help illustrate the cam's
performance. For those of you who are playing with this leveraging a round
on the internal edges will help you visualize if the results are as needed.
You will find that creating a simple edge round (with correct results)
should yield parallel internal tangent edges of the round throughout the cam
surface. It should also allow you to create a full round as illustrated in
this example. Switch the full round OFF and set the value to about 4 and you
should see what I am talking about on the inside surface of the cam



(WF5 & features suppressed to reduce file size).



Hope this helps others,

Tim McLellan
Mobius Innovation and Development, Inc.

[ptc-2860737]

Hi all

While I've been pondering this problem I've been trying to set up a
mechanism to be able to see the motion of the pin following the groove,
but Pro/E will not allow a cam follower connection. Does anyone know how
to go about setting it up so a pin will follow the groove for a motion
study?


Doug Barton
Mechanical Designer
Parker Hannifin Canada
Electronic Controls Division
1305 Clarence Avenue
Winnipeg, MB R3T 1T4 Canada
direct 204 453 3339 x309
fax 204 452 7156
-
www.parker.com/ecd

[KrisR]

Here is the cam-follower connection - it is in Applications->Mechanism.

[cid:image003.jpg@01CD0E74.AB30F500]

[TimMcLellan]

Just keep in mind that the motion envelopes generated in Mechanism are
facetted (i.e. surface, STL...)



Tim McLellan
Mobius Innovation and Development, Inc.

[ptc-2860737]

Thanks, however I already know how to set up a cam follower connection.
What I said was Pro/E would not allow it (in this case). I believe this
is because the cam surface is curved in two directions which makes it
invalid. Although I also tried just the profile curve instead of the
surface and that didn't work either.

From Help


The question was, is there a method other than a cam follower connection
to get the pin to follow the cam profile.


Doug Barton
Mechanical Designer
Parker Hannifin Canada
Electronic Controls Division
1305 Clarence Avenue
Winnipeg, MB R3T 1T4 Canada
direct 204 453 3339 x309
fax 204 452 7156
-
www.parker.com/ecd

[blenk]

Bjarne, I applied your idea to the track I was trying to make and it appears that it works great. I used a second trajectory instead of a CAM graph. The sides of the track stay at 90 degrees even where the track cuts across the cylinder at an angle. Thank you very much for sharing this with us!

VSS created with curve on end of cylinder and pre-determined track curve:
[cid:image008.png@01CD0E77.56626B00]

[ptc-2860737]

Yes. Maybe not enough resolution to see anything worthwhile.

Here's an image of an assembly of parts. The pins are just assembled
point on curve.




Doug Barton
Mechanical Designer
Parker Hannifin Canada
Electronic Controls Division
1305 Clarence Avenue
Winnipeg, MB R3T 1T4 Canada
direct 204 453 3339 x309
fax 204 452 7156
-
www.parker.com/ecd

[TimMcLellan]

Tom,



Here is your part.



Changes include:

· Converted profile to GRAPH feature. Scaled dimensions to simplify
the VSS for 360° rather than using the arc length as you had defined. This
simplifies the CAM path and allows for easy modification. Also, modified
splines and converted complete sketcher/graph to spline (one entity in
sketcher graph).

· VSS paths use new GRAPH and sweep as I had illustrated yesterday
(origin & X-vector to control orientation). Note: both VSS use same graph
to drive the path over 360°

· VSS is extended outside of outer surface and inside of inner
surface then merged.



Hope this helps everyone.

Tim McLellan
Mobius Innovation and Development, Inc.

[TimMcLellan]

Sorry, everyone. After I sent the last file I noticed a small error with
the cutting tool in the assembly.



Please download the latest example which includes Tom's part updated and an
assembly where you can simulate the tool cutting (mechanism).



Have a great weekend,



Tim McLellan
Mobius Innovation and Development, Inc.

[TomU]

For those who are following this thread, here is a cam that was cut like it would be in real life. There are revolved cuts (804 of them to be exact) normal to the outer cylinder surface, tracking the desired trajectory around all the way around. Doing this clearly shows the off-axis gouging that occurs from the cutter's tip as well as what happens when the slot is so wide that the offset becomes sharp in areas. If only there were only some way to get Proe/Creo to properly create clean surfaces representing this roller surface (not the gouging necessarily, but at least the partially sharp corner area).

[cid:image001.png@01CD10EF.4AB88FD0]

[cid:image002.png@01CD10EF.C4458150]

[mjenkins]

I created an angled offset axis and revolved the profile around it.

But I think this may not be what you are calling a barrel cam.

In Reply to Tom Uminn:

I'm looking for ideas on how to properly model the slot in a barrel cam. I have attempted multipe techniques and I'm starting to spin my wheels. I have an idea that what I want to do can probably be done with the variable section sweep, but that feature is definitely not my area of expertise.

I've worked through the "wave washer" suggested technique on PTC's site (http://www.ptc.com/appserver/wcms/standards/freefull_cskdb.jsp?&im_dbkey=88651&icg_dbkey=900). The problem with this exampl is that the part is not a constant thickness (whereas my cam slot needs to be). To get a constant thickness, I changed the sketch profile to a single line, which created a single wavy surface, and then thickened it. Problem is, the sides are no longer vertical. To get around that, I lengthend the VSS section to include some extra on each end, then extruded inner and outer cylinder surfaces, merged everything together, and solidified. This succeeded in creating a constant thickness washer with straight sides.(See attached PDF.)

Problem is, thistechnique does not transfer well to the cam. The centersurface won't offset nearly as far as it would need to. Creating inner and outer curves and building boundary blends didn't work either.

Tom

[TomU]

This works because the follower is a ball. Once the follower becomes a cylinder, things get much more interesting.