[SVM]

I am making a bracket that will clamp onto a 1" OD pipe. I model it in SolidWorks, make the 3D G-code with FreeMill (just to try it) and cut with Mach3 with a 1/2" ballnose.
The resulting diameter is smaller than the 1" I need because of the 0.01 increments I have in the tool path. In other words, since the ball leaves a little material between the passes, this is enough to cause the ID to be too small to fit around the pipe.

Is there a way to predict this and model the diameter of the cut slightly larger to compensate for this effect?

Steen

[ger21]

Using Freemill to cut a hole is definately not the best way to do it. A simple 2D drawing and ACE converter www.dakeng.com or SheetCAM www.sheetcam.com will be more accurate, and a LOT faster to machine. Unless of course it's a complex 3D part.

[SVM]

Gerry,

It is a not really a hole, but more like a half shell. This is one half of a clamp that goes around a piece of 1" pipe so it is on fact a 3D piece. The outside of it is real 3D and very contoured, but that is not a problem.

Thanks for the reply!

Steen

[SVM]

I have attached a .JPG picture of the part for clarification ... I hope I did it right ...

Steen

[Kiwi]

Does your tool path follow the arc as this will give you a better finish?
If the path runs lengthwise you will have large cusps on the sides...larger near the top.
Suggest you clamp the two halves together with shim between faces and clean with reamer.

[ger21]

You can cheat a little in MeshCAM, and tell MeshCAM that the tool is a little smaller than it actually is. This will cut it a little bigger.

[SVM]

Kiwi,

My toolpath is lengthwise right now, but I could cut it across instead. I am sure you are right that it would cause less of a problem.

The reamer trick would defenitely work too and would produce a nice finist. But I was trying to get by without having to buy (and wait for) any additional tools.

Since this clamp is supposed to grab onto the pipe, it is actually nice to have the rough surface.
I am just thinking it should be possible to come up with a way to calculate the actual ID based on the tool diameter, the ID of the cut and the stepover.

Thanks for the reply!

Steen

[HuFlungDung]

I'd suggest that you post your gcode file on here for a quick run through a backplot. Depending on whether your code is offset for tool center or for tool tip, there could be a considerable variation in the resultant path.

Is this true 1" OD tube, or is it 3/4" pipe, which might be 1.05" OD?

[Geof]

Scroll down the page you get at this link and you will find a formula for cusp height on a flat surface. The whole article goes into a lot of detail about surface finish, stepover, etc, etc.

http://www.haascnc.com/solutions_3D.asp#3D

[Kiwi]

Steen
I've had a play and come up with a simple program.
Is this what you were wanting?

[SVM]

Thanks a bunch guys, all great responces!

Hu: The pipe is a true 1.0" OD pipe so thats not the problem.
The code is generated with FreeMill ... not exactly the most sofisticated post on earth, but it was ... Free! It does not have any features for setting tip or center.
I have BobCad 21 but have not yet dug into the details of that. FreeMill was just an easy way to get a part I needed done.

I am looking at MetaCut and will download it to see what it does.


Geof: Thanks for the article link. That was a very informative read!


Kiwi: Nice tool! Yeah that pretty much does it, good illustration too!
Now ... if you could make it work backwards from the "Diameter Inside Largest Cusps" and produce the "Profile Diameter" as an output, then it would be exactly what I was talking about!
However, dont spend your time on that, I can experiment with the Profile Diameter and recalculate until it is close enough to the final ID I need. That will work fine.
Would you mind sharing the formula you use to calculating the Cusp hight on an Arc face?

Thanks a bunch for taking out time to make the program!


To all: As you can probably tell, I still have a lot to learn about CNC. I have an Enco mill which I converted myself, also build the electronics etc. I am using it for prototype work and am enjoying learning the process as we go.
I really appreciate you taking out time to answer my question.
Hopefully there will come a time when I can start paying back!

Rgds
Steen

[Kiwi]

Steen
I've added some more features to the Cusp Calculator.
As the cusp height varies around the arc the coord the program generates are not a true arc.
Have a look, this may help you generate a better curve.

[SVM]

Kiwi,

Man you are good!

You know .. thinking about the distortion of the resulting arc if I enlarge the profile diameter, and the fact that it is mainly the 1st cusp on each side that causes the problem, it might be better to create a toolpath which has a finer stepover near the rim and thereby knocking out the 2 offending cusps.
I'll have to code that by combining 2 files. I'll try that today and see how well that fits.

Thanks for the effort!!

Btw, I'll be passing through NZ in October and again in December.

Rgds
Steen

[Kiwi]

Steen
Making smaller step-overs at the top will make a truer shape.
This is how I calculated only the cusp at the top of the curve.
I used a different formula for the cusps that were not trimmed with the parallel part of the ball nose cutter.

The dimensions maked are all known.
L1 = Sqr(5^2 - 4^2)
L2 = Sqr(L1^2 + (45+4)^2)
L3 (Cusp Height) = (100/2) - L2

Cheers.

[SVM]

Kiwi,

Ah ... pretty simple when you think about it. Thanks for the illustration and the explanation!

Thanks
Steen

[Kiwi]

Steen
This is how I calculated the other cusp heights....a bit more involved.

The dimensions maked are all known.
L1 = (100/2) - 15
A1 = ASin(L1) * 45
L2 = (100/2) - 17
A2 = ASin(L2) * 45
A3 = (A1 + A2) / 2
A4 = A3 - A1
A5 = (45 * Sin(A4)) / 5
A6 = 180 - A4 - A5
L3 = (5 * Sin(A6)) / Sin(A4)
L4 (Cusp Height) = (100/2) - L3

[SVM]

Kiwi,

Certainly a bit more involved, but still fairly straight forward geometry. I see what you are doing and kick myself for not thinking through the problem before asking :-)

Thanks again!

Steen