[paulus.v]

Hello all,

Imagine a circle drawn 3-dimensional,so that the circle plane does not coincide with one of the XYZ planes. If I want to follow this circle with a 3-axis mill, is it any possibility to define the g-code using arcs? Or the only way to program a circular movement in all 3 planes is by reducing it to straight lines?

Thanks

[islaww]

You should be able to code two half circles with Z movement if you can define the highest and lowest points. One half circle with Z down, the other with Z up.

That said, there should be no issues with a segmented circle, assuming the machine is well setup. Reality is that even when cutting with G2, G3 moves the machine is given straight line segments. That is the only way GCode machines move. There are no arc moves. Only straight lines. The only difference is that with G2-3 moves, the controller writes the segments and not the CAM program.

[victorofga]

I think it pending by the control...

virtually you can define a workplane and on that workplane you can use the circular interpolation, what actually tied to a defined plane

using the ""standard""" G codes, youre right only the 3 predefined plane available..

that's why there are the G17-18-19 codes..

as gary already pointed, it is your hardware..

[sinha_nsit]

The simple answer is NO.
Either use a CAM software or write a macro yourself. Yes, straight-line segments will have to be used.

[paulus.v]

Thanks for the answers.

Now I realise I have explained my problem wrong. In the above example if you look to the circle perpendicular to XY there is an elipse.

But what I try to obtain is an interpolation of two circular arcs. I attach images for a better understanding. There is half a circle in a plane defined by the arc end points perpendicular to XY plane (first image). And this circle is "bent" to form another circular arc in XY plane (second image).

I want to know if it is possible to define it using a single line of code.


Attachment 317516

Attachment 317514

[sinha_nsit]

Yes, if that line calls a macro. It is also possible to define a new G code involving this macro.

[partur]

What if you rotate the axes?
Attachment 318100
Attachment 318102

[Kiwi]

Write you code for one plane eg G17 and Rotate using G68 (Fanuc)
The hardest part is determining the second plane angle when setting the controller, if your arc's plane is rotated on two planes.

[paulus.v]

please look at my post #5 where I explain better what I need. There is no circular arc in one plane, but the arc is also bent.

Actually I need a simultaneous arc movement in two planes.

[Superman]

You haven't stated what machine & control you what to do this...

Imagine a circle drawn 3-dimensional,so that the circle plane does not coincide with one of the XYZ planes. If I want to follow this circle with a 3-axis mill, is it any possibility to define the g-code using arcs?

The only way to get circular moves, is to place the material in such a way, that the arc is seen by the tool as "planar" ( meaning all points on the arc are in-line on 1 plane ) ....this is what "PARTUR" is meaning

Or the only way to program a circular movement in all 3 planes is by reducing it to straight lines?

- this is the simplest way,,,, long code but simpler
- best done using a CAM system, as it would auto-decide on each segment length when applying or maintaining a tolerance to the wireframe geometry

[sinha_nsit]

please look at my post #5 where I explain better what I need. There is no circular arc in one plane, but the arc is also bent.

Actually I need a simultaneous arc movement in two planes.

Is it like a helix?

[paulus.v]

Is it like a helix?

It is more than a helix. I need synchronised movement in 3 axes as well but for helix the third axis (Z) has a linear movement. I need it to be circular. It is like having, for example, G17 and G18 in a single line of code, with the same start/end points but two different radii.

I was thinking of a simple program to generate 3D patterns g-code. But as I need to output straight lines g-code, the programming gets way too complicated.

http://wallers.pk/wp-content/uploads...anel-164-p.jpg

Thank you all for the replies!

[sinha_nsit]

CAM is the only convenient method.

[Kiwi]

This code will machine an arc on XZ Plane rotated 45 deg and tilted 10 deg. Similar to pic on Post #5,.

Using Fagor Control. (The same should be doable with Fanuc using G68)

G18
G49 A7.1071 B-7.1607 C44.1225
G01 X-30 Y0 Z0
G03 X30 Z0 I30 K0
G17

[norbert.barnard]

Now I realise I have explained my problem wrong. In the above example if you look to the circle perpendicular to XY there is an elipse.

But what I try to obtain is an interpolation of two circular arcs. I attach images for a better understanding. There is half a circle in a plane defined by the arc end points perpendicular to XY plane (first image). And this circle is "bent" to form another circular arc in XY plane (second image).

I want to know if it is possible to define it using a single line of code.


Attachment 317516

Attachment 317514[/QUOTE]

Hallo Paulus,

i see the following:
pic 1 show an arc, looking from bottom right corner.
pic 2 show this arc bended around another arc (i.e. outside of a cylinder)

to machine this as a toolpath:
A.on a xyz machine,
1. in cad, this is then defined as a spline (cnc capable 1 line instruction)
2. in cadcam, spline broken down to straight lines

B. on a machine with a rotary axis,
3. one of the arc's could to be made with the movement of a rotary axis, the actual application would decide if this is possible.

C. Yes, a purpose-build macro could also be used, calculations would result in straight-line segments,
would the design-effort be worth it, only the user can decide.

enjoy.

Norbert

[RCaffin]

> I want to know if it is possible to define it using a single line of code.
No.

You can do anything you want if you can specify the arc in mathematics. Then you go into CV mode and just program it with 1000 calls to a G1 X Y Z instruction, incrementing the XZ, Y & Z values with each step.

Cheers
Roger

[paulus.v]

As I understand now, cnc controllers are happy to work with straight lines, probably the arcs are also executed as straight lines or points of intersection.

Writing a macro seems a good solution but not for me, as I am not that much into programming and mathematics.

I have already started to program using CAD/CAM software but most CAM programs do not like 3D sketches, without solids...

[CNCFr]

The simple answer is NO.

... but wrong.

With each controller which allows the definiton of coordinate systems rotated in space it's easy to produce cirlces arbitrarily located and rotated in space.
I know at least one control (Siemens 840D) wich allows the definiton of circles in space by providing three points in space without the need to define a special rotated cordinate system for this purpos.

By the the way: Cirlcels in space can not be created by using sections of helixes. Helixes are 3-dimensional curves while circles are plane curves (2-dimensional curves). Or mathematically speaking: The torsion of circles is zero while helixes have torsions unequal zero:

https://en.wikipedia.org/wiki/Torsion_of_a_curve

[StarAppJohnUK]

hi all,

I may have misunderstood what your looking for but I have a simulated helical mill macro, that does a 3 axis move via point to point programming.

NB: it will only work on a swiss cnc lathe but can easily be modified to suit your machine. let me know if its of any use

[RCaffin]

at least one control (Siemens 840D) wich allows the definiton of circles in space by providing three points in space without the need to define a special rotated cordinate system for this purpos.
I stand at least partly corrected. Fair enough. That Siemens controller can handle some fairly sophisticated mathematics internally.
Now, I wonder how many other controllers can do that? Or put another way, is that code portable? I am curious.

Cheers
Roger