[ddragon]

Hello
Just about finished my 4th axis. Does anyone have an inexpensive recommendation for software that can creat g code for a 4th axis project?

Thanks

[awerby]

Take a look at DeskProto. It has several 4th axis strategies including along X, around A, and n-sided indexing. You can try it for free for 30 days.

Andrew Werby
www.computersculpture.com

[ddragon]

Thanks. I'll do that.

[lancut]

Can't go wrong with CNC Wrapper.

CNCWrapper - Home Page

[ddragon]

Hi
Yes. I have cncwrapper. But I also need something that will do more than wrap.

[MarcusWolschon]

Be aware that DeskProto only does Pseudo 4 Axis milling using A, Y and Z But not real 4 or 5 axis machining since X doesn't move.
It also fails to handle path distances correctly as you get near the center and the same linear distance means a higher rotational movement.
For speeds the same is the job of the machine control program and many (Like MACH3 don't do that, leading to vastly lower speeds near the center and higher speeds on large diameters.)

[MarcusWolschon]

I started to write a 4 and 5 axis CAM.
A prototype can be found as "simplemultiaxiscam".
I'm waiting for the CAM API in FreeCAD to implement a final version.

[awerby]

Have you actually tried DeskProto? It certainly does have a 4th axis strategy where X moves. I'm not sure what you mean by "Pseudo", but it's pretty effective in making parts using the 4th axis, without the limitations of a "wrapper" type of program. I've never noticed that the path distances were incorrect towards the center; it seems that if that were the case, the parts it made would be distorted, but they haven't been.

If you want to write a better program, by all means go for it, but it doesn't seem like your efforts have culminated in anything that the person who was asking for help in this thread would find useful quite yet. How about posting some examples of 4-axis parts you've made, or videos of your machine doing "true" 4 and 5 axis machining, rather than just bad-mouthing your competition?

Andrew Werby
www.computersculpture.com

[MarcusWolschon]

I'm using Deskproto regularly.
What strategy moves X, Y, Z and A?
The part would not be distorted at all but a toolpath distance of e. G. D/5 becomes more like D/64 taking forever and melting stock material that must be machined at a certain minimum speed.

Gesendet von meinem INO_ONE_PLUS mit Tapatalk

[awerby]

You said "DeskProto only does Pseudo 4 Axis milling using A, Y and Z". It can also do 4-axis milling using A, X and Z as well as N-sided indexing. If you're now insisting that it's "Pseudo" unless all axes move at once (for some reason) then fine; write your own program that does that.

It sounds like you haven't got Mach3 set up correctly, if your toolpaths are taking that long near the center. You need to check "Use distance for feedrate" and put a value (.0001" or .001mm) in the Rotation axis values.

If the program you wrote works so well, why are you using DeskProto anyway?

Andrew Werby
www.computersculpture.com

[MarcusWolschon]

A, X and Z is not 4 axis. It's 3 axis while one of them happens to be a rotational axis.
Try doing a parallel strategy on A with a non-meander movement.
At 360 degrees it will raise the tool, travel to A=0,Y+1 and put the tool down.
It doesn't handle the fact that 360 and 0 degrees are the same thing.


As I said... I did but it's a working prototype.
It even reads the existing Deskproto tool files.
Too slow until I get the hitpoint code and data model from FreeCAD. They haven't finished their API to allow tool path generators in yet. So I have to wait until I can integrate may working code.
I'm using Deskproto for all the regular 3 axis work.

The MACH3 setting you mention is not the solution but the problem.
You have to enter a radius by hand while in reality the radius changes while you mill away material.
It should ask you where in space your axis is.

[awerby]

A, X and Z is not 4 axis. It's 3 axis while one of them happens to be a rotational axis.

[It's a way to make parts using a rotational "A" axis that can't be done with a "wrapper" type of program, one of several strategies this powerful but inexpensive program offers to do this, which other programs in the same price range lack. If you want to insist pedantically that it's "not 4 axis" I suppose that's your right, but please don't spread misinformation about a product that works quite well for many people who otherwise couldn't carve their models using their rotary 4th axes.]

Try doing a parallel strategy on A with a non-meander movement.
At 360 degrees it will raise the tool, travel to A=0,Y+1 and put the tool down.
It doesn't handle the fact that 360 and 0 degrees are the same thing.

[It can, if you set it up correctly. In the Library of Machines, under the Advanced Settings, you need to click the box that says "A axis values can exceed 360". If you do that and it still stops at 360 and rewinds, you've probably got your Mach3 settings wrong. Here's a quote from the Mach3 manual: "Rotational: Rot 360 rollover, if checked, will measure a rotary axis modulo 360 (0 to 360 then restart at 0). Otherwise, it will keep counting up (for example, two revolutions would be 720)".]


As I said... I did but it's a working prototype.
It even reads the existing Deskproto tool files.
Too slow until I get the hitpoint code and data model from FreeCAD. They haven't finished their API to allow tool path generators in yet. So I have to wait until I can integrate may working code.
I'm using Deskproto for all the regular 3 axis work.

[Doesn't this "working prototype" of yours even do 3-axis machining yet? Why don't you write your own API? The people who wrote DeskProto didn't wait for someone else to do their work for them.]

The MACH3 setting you mention is not the solution but the problem.
You have to enter a radius by hand while in reality the radius changes while you mill away material.
It should ask you where in space your axis is.

[As I understand it, it gets that information from the Z-axis DRO, once you've put a minimum value in the rotational axis box. If it's not working for you, you might ask about it in the Mach3 support forum.]

Andrew Werby
www.computersculpture.com

[MarcusWolschon]

A linear encoder (not really the DRO display http://en.m.wikipedia.org/wiki/Digital_read_out)
Cannot supply the radius.
It can only supply the absolute position on one axis.
For the radius all axis need to be known and the user must have somehow entered the position of all rotational axis and how they are connected.

Gesendet von meinem INO_ONE_PLUS mit Tapatalk

[arizonavideo]

I downloaded the jar file and will give it a look soon.

Thanks for working on it Marcus.

[awerby]

A linear encoder (not really the DRO display Digital read out - Wikipedia, the free encyclopedia)
Cannot supply the radius.

[Mach3 doesn't support linear encoders, as far as I know.]

It can only supply the absolute position on one axis.
For the radius all axis need to be known and the user must have somehow entered the position of all rotational axis and how they are connected.

Gesendet von meinem INO_ONE_PLUS mit Tapatalk

[You're over-complicating things here. If we're talking about a 3-axis mill with a rotary A axis, used the way DeskProto supports, then the Z axis is zeroed at the A axis center. The Z axis DRO will report its position at any point; that's equal to the radius of a circle centered in the middle of the part, which can be used to modulate the feedrate.]

Andrew Werby
www.computersculpture.com

[joe aiello]

Marcus wrote:
"I'm waiting for the CAM API in FreeCAD to implement a final version. "

As I am a newby in cnc stuff, hoping someone else would ask - why is the Freecad package so important as oppose to Heekscad, Cambam, Gcoodtools or similar not appropriate in your application?

....joe

[MarcusWolschon]

It's just what I chose.
Commercial code of cause is no option as I need access to their model. I found many free options to not support 4/5 axis machines, so it's no use implementing strategies for such machines there as the very data model would be optimized around assumptions like the cutter always coming straight down vertically.

Gesendet von meinem INO_ONE_PLUS mit Tapatalk

[MarcusWolschon]

Awerby: that makes a great many assumptions. Like having exactly 1 rotary axis and that axis being immobile.
(it may rotate the tool, it may be a 5 axis setup, you don't even know if it's along X or Y without having asked the user and 0 may very well be, where the end stops are working with absolute coordinates to simplify things )

Gesendet von meinem INO_ONE_PLUS mit Tapatalk