¿ªÔÆÌåÓý

ctrl + shift + ? for shortcuts
© 2025 Groups.io

Formula for axis extension


 

Hi,


I have a job to do some 3D machining in foam on my router, it's two-sided machining and unfortunately bigger than my machine.


I can't face having to mess around indexing the part in x so I thought I would add a simple moving bed to my moving gantry machine, when the gantry moves one way the bed moves the other. I've done this using an off the shelf ballscrew actuator and the 4-axis stepper driver that was doing nothing.


I set up the x-axis as a dummy axis so the DRO works but it does not output pulses and then in the formulas section did


A = x/2

B = x/2


Where A and B are set up to drive the gantry and bed respectively. This works with basic moves from the MDI:




However, when I did a test cut circles came out really funny. I'm assuming this is because the lead of the bed is twice that of the gantry and although they both have the correct pulses per mm set, in the middle of an interpolated move their differing resolutions mess things up.?


Can anyone confirm or give me some more insight into how the formulas do their thing? I'm thinking I can cheat the pulse/mm and modify the portion of x each axis moves to correct for this but I want to be sure before I make my brain hurt :)


Cheers,


Graham


 


Well I solved my own problem. After some experimenting I realised something I probably already knew from my previous experiments with parallel kinematics and that is that the kinematics are only calculated for end points so are more or less ignored inside G02 and similar commands, either you split everything into tiny lines or you find another way.

What I did realise is that if my moving gantry and moving bed had the same pulses/mm then I could just run one as the slave of the other and then divide the pulses/mm of the master by two and not need the formulas.

BUT mine are not equal lead, one needs 80 pulses per mm and the other 50. But I thought, why not still do the same thing but calculate a new pulse/mm that balances the two. I did and it works like so:

Distance moved = (1/80)x + (1/50)x ? ?where x is the number of pulses applied.
Distance moved = (1/80+1/50)x

For 1mm movement

1 = (1/80+1/50)x

x = 1/?(1/80+1/50) ? ?pulses/mm

So I stuck this number (37ish) into the box for the x-axis and set my a-axis (connected to the moving bed) as a slave.

All works fine and circles are now round :)

A video of it working is currently uploading.

Cheers,

Graham
?


 

¿ªÔÆÌåÓý

I think thats an excellent approach, imo.
Kudos.


On 02/03/2015 23:39, grezmos@... [mach1mach2cnc] wrote:

All works fine and circles are now round :)

A video of it working is currently uploading.

Cheers,

Graham

-- 
-hanermo (cnc designs)


 

Thanks! ?Here is is doing some actual cutting:



Cheers,

Graham

?




---In mach1mach2cnc@..., <gcode.fi@...> wrote :

I think thats an excellent approach, imo.
Kudos.

?