Hello Guest it is June 08, 2023, 01:30:56 PM

Author Topic: Tangental Cutter problem!  (Read 4410 times)

0 Members and 1 Guest are viewing this topic.

Tangental Cutter problem!
« on: September 16, 2009, 06:38:05 AM »
My Fabric sutter is having some issues.

It is a 2.5D pnuematic Z axis with tangental A axis.

The problem is that after a cut the trajectory planner starts to rotate the knife for the next move before the Z Axis has lifted. Iam using a Z Gcode  of -200 to +200 with a feed rate of F10000. This is to provide time for the Z to go down prior to a cut and the same for the lift

Watching the gcode running once the cut is complete and the code is calling for +200 and before it reaches 0 where the Z should lift, the A axis is rotating to a postion for the next cut.

I have attached a short video of the problem http://www.youtube.com/watch?v=6stJyMRDmkc
Re: Tangental Cutter problem!
« Reply #1 on: September 16, 2009, 07:02:26 AM »
How does mach3 know the position of the Z-axis? does it have an end switch telling mach3 the position?
Re: Tangental Cutter problem!
« Reply #2 on: September 16, 2009, 06:06:34 PM »
I have just been through my old messages to Brian,

It seems this problem occurs on G00 moves. If I change the Z to G01 moves the Z retracts before the A spins.

This was a bug from last years rev and it seems as it was forgotten in the latest rev.

I have let Brian know so he can do the fix in the next rev.


Re: Tangental Cutter problem!
« Reply #3 on: September 17, 2009, 03:26:42 AM »
Because I have investigated A-axis behavior a lot of time, I'll try to share my knowledge.
1. Rotate is done every time with rapid feed ratio. This is not good.
2. Mach 3 has no way to know what you want to do. So, you should up the knife before fast moves (G0).
3. G1 move uses tangential control, so knife direction will be decided before linear interpolation.
4. + and - rotates in different directions, but +0 and -0 in absolute coordinates always are +. This is very bad.

So, IMHO you should up the knife, move with G0 to desired position (will not rotate) and do linear interpolation X,Y with very small Z down in the same line. Because rotate will done with rapid feed, position will be exact before cutting. Also, this give smooth start of cutting. Of course, you should compensate in NC file all this.
Probably, this is the possible solution with the Mach 3 current behavior. If your knife is center based to Z-axis, forget for A0 position when the knife is into the material. You will produce holes in some situations.

This is my opinion, and probably I'm on a wrong way at all...