Community Tip - Stay updated on what is happening on the PTC Community by subscribing to PTC Community Announcements. X
I have a link soomewhere for this, I'll try and dig it out. I never tried it though.
thank you.. i'll be waiting...
It kind of depends on which pattern you subscribe to. Wiki has a good guide image: http://en.wikipedia.org/wiki/Golf_ball
Have you considered creating an appropriate texture
Wouldn't a "fill" pattern work?
A texture is an image that can be wrapped onto the surface of the sphere. With a little effort, you can even have the rendering engine mimic the dimple "look". Unfortunely it will not do much to help create the actual dimple feature.
I'll see how close I can come to this. Sounds like an interesting challenge.
That model was a lot of effort by someone. I wonder how those pattern tables were generated.
Very nice work, whoever did it (the History interrogation still doesn't work in Creo 2.0 M040).
It does return a username EBOWERMAN.
So I started here and it certainly is not complete.
I used the Wiki example on the left as a guide. This creates a 252 hole golf ball.
I am only using this as a skeleton starting point. The dimple sizes are not optimized nor is the size.
Reverse engineering both images shows that it is based on a nested set of 12 pentagons and 20 triangles. With Pro|E (all versions) you can assemble these shapes and they end up locking themselves into the proper form.
From the assembly I drove a sketch back to the pentagon to determine the height of the 5 triangles on the pentagon so I could determine the diameter I ended up with. Again, none of the holes are optimized as yet.
The idea is that you want to be able to define this shape within a part file and not an assembly. Holes, or points, still have to be mapped on a sphere. The dimple sizes are not consistent... you see the full scope of the challenge. And this is not even the definitive standard for golf balls although I do like this one for the challenge at hand.
The next level of subdivisions should determine exactly where the dimples should fall. In that way, points could be added to determine the appropriate location on the sphere.
. i really appreciate the effort you have done here... thanks
The pattern of the dimples in the picture that you supplied actually looks pretty simple when looking at it through a spherical coordinate lens. My first approach would be to use another program (Python, MATLAB, Excel, etc.) to calculate the locations of all the points of the dimples and save the points in a separate file. I would then have pro/e load those points with the Offset Coordinate System Datum Point Tool (I think... I have never tried this). I would then make one dimple feature, and use the point pattern tool to fill the rest of the dimples.
I think this would work to make a ball in the image that you attached. You would quickly run into logistical issues if trying to replicate the ball the Frank attached. You could probably still use the method, but you would likely have to break it into smaller pieces.
When I find some time, I might have to give this a shot. I will let you guys know how it goes.
Hi All,
here is my model of golf ball: GOLF - Gentleman Only, Ladies Forbidden?
Vladimir
http://communities.ptc.com/message/165911#165911
Scroll down and you`ll see golf ball with explanation. I`ll see if i can find my old model from this post tommorow.
Mani, you have unintentionally started a surprisingly challenging discussion and quite a few off-shoots.
And since you started it, I should follow up. I spent most of the week trying to "understand" the problem rather than just making due. Between the challenge itself and making Creo do what it should really did provide a lot of insight and even frustration.
This is what I ended up with: 260 dimples on a 1.68" sphere. This model is attached as a STEP file.
I ended up using 3 levels to manage the point plotting onto the face for patterning. This technique could be used for any pattern, but I choose the Wiki version with a few extra holes at the centers of the resulting hex pattern.
The 1st level was as discussed before, a modified dodec/icosahedron (32 faces).
I was finally able to generate this shape with a comprehensive sketch that controls planes from which the "fill" patterns were created.
Next, I needed to stellate one of the pentagons. Through the sketch, I was able to determine the of the center on the bounding sphere. Then 2 key sketches were created to determine the centers. These odd sketches will determine the "centers" of the dimples at the points of the pie shaped pieces.
A surface sphere of the final size was created. These sketches were used to create Offsets. This provides Vertexes to use for patterning later on. The offsets were patterned on a hemisphere.
Now, 95% of the work is done. I have vertex locations on the sphere surface. Next a solid sphere is created and a 1st dimple is revolved from the solid face.
Now, I -thought- I could make datum points on the surface using the "darts" I created. But patterning to those points does not maintain a surface normal. All the axes of the pattern became unidirectional without options. Even if I tried to orient the points, no joy. The answer was to map the points onto a sketch in the pattern command. This was easy enough. Now you have the options of mapping to surface normals. remember to choose the Origin as well.
Now I am able to modify the 2 simple sketches above and optimize the distribution of the dimples by tweaking the sketch constraints. You will notice I used one size dimple. Even greater optimization could be done if you separated the dimple creation in 2 steps.
From here, this patterns were copied, pasted w/ 180 degree rotation, grouped, and rotated about an axis 5 times. And last, the fillets were added.
To clean up the file, I hid all the reference geometry and also the 260 axes that were created from the dimple cuts and saved the layer status.
And the end result is attached
Very impressive work. Just creating the modified dodec/icosahedron requires a bit of thought and foresight, not to mention the path you took to complete the other two steps. Awesome model!
A ball with some dimples... How hard can that be? Oh... Wait...
Thanks Kyle!
As an aside, there is one way to quickly create vertexes for a spherical surface. You can export a half-sphere surface to VRML (.wrl) and open (import) it. This creates a facet feature (whatever that is). If the export went well, it creates a nicely segmented sphere... although on a square pattern. This does provide many vertices for patterning or for generating a more complex patterning sketch.
The other observation that this exercise presents is the spherical spiral that could be used for patterning with 3D datum curves, even if these curves were generated from equations.
I have spend years working with VRML files. With some advanced editing techniques, they can be very useful in capturing large datasets and even manual manipulation of the data. Point files should be easy to create from VRML files. These are essentially point clouds with polygon definitions.
coord DEF FaceC Coordinate {
point [
0 -0.21335975 0,
0 -0.21286241 -0.014560042,
0 -0.2113727 -0.029052266,
0 -0.20889722 -0.043409362,
0 -0.20717408 -0.05047996,
-0.012087606 -0.20729499 -0.049041558,
-0.023472902 -0.20729499 -0.044723812,
-0.033493964 -0.20729499 -0.037806884,
...
...and polygon connections using the point numbers as reference for connectivity:
normalPerVertex TRUE
coordIndex [
5, 25, 24, -1,5, 24, 23, -1,5, 23, 22, -1,
5, 22, 21, -1,4, 5, 21, -1,26, 25, 5, -1,
26, 46, 45, -1,26, 45, 44, -1,26, 44, 43, -1,
25, 26, 43, -1,47, 46, 26, -1,47, 67, 66, -1,
47, 66, 65, -1,47, 65, 64, -1,47, 64, 63, -1,
46, 47, 63, -1,68, 67, 47, -1,68, 88, 87, -1,
68, 87, 86, -1,68, 86, 85, -1,67, 68, 85, -1,
89, 88, 68, -1,89, 109, 108, -1,89, 108, 107, -1,
89, 107, 106, -1,89, 106, 105, -1,88, 89, 105, -1,
110, 109, 89, -1,110, 130, 129, -1,110, 129, 128, -1,
110, 128, 127, -1,109, 110, 127, -1,131, 151, 150, -1,
...
...and they even contain face normal definisions!
normal DEF FaceN Normal {
vector [
0 -1 0,
0 -0.997 -0.068,
0 -0.99 -0.136,
0 -0.979 -0.203,
0 -0.971 -0.236,
-0.056 -0.971 -0.229,
-0.11 -0.971 -0.209,
-0.156 -0.971 -0.177,
-0.194 -0.971 -0.134,
-0.221 -0.971 -0.083,
-0.235 -0.971 -0.028,
...
This could simplify the original request if an acceptable pattern can be visualized within this model.
I have attached the hemisphere_prt.wrl file for your convenience.
Spherically pattern - new pattern for Spherically shape
What do you think? It is good idea for new feature in Creo? - Vote here
Ideas should be a bit more descriptive, so I added a comment. Maybe you could add it to the idea.
Trick is to make half sphere not whole sphere and then project pattern on to it. Fill Pattern type must be circular. First dent u create in center of halfsphere surface. when you complete halfsphere you just mirror it.
Now comes problem with distance between pattern items on 1 halfsphere and other. You must use a bit of math in relations to achive nearly perfect ball.
Here is the model: