Re: [Emc-users] Backlash comp monitoring
On 01/23/2020 11:07 PM, David Berndt wrote: Hello All, I'm having a bit of an issue with some geometry not coming out as expected. I suspect it's related to backlash and cutting forces moving the table a bit but I don't have linear scales to monitor/correct for that sort of thing so I'm left wondering a bit. I was hoping that being able to see what backlash comp is doing on screen as my toolpaths run might enlighten me some. Is there any pin or ?? that can be monitored to give me more insight? What motion control system are you using? Mesa, Pico, software stepping? In any case, there is some position output that can be fed into Halscope to read the real-time position and graph it. The problem with backlash compensation is that you are trying to cover up a defect (usually slack in the leadscrew) with software that can't actually read the slack. So, any force from the cutting tool can put the table on either side of the slack, and the computer can't sense it. Also, when changing directions, it requires the motor to instantaneously jump from one side of the slack to the other. So, the best fix is to remove the slack. Depending on the mechanism, Acme nuts can often be sliced in half and compressed or otherwise adjusted to reduce backlash. One problem with Acme screws is that there can be more wear in the middle of travel, making it hard to adjust out the slack throughout the range of travel. Dual ballnuts can be set up with a preload. Jon ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Backlash comp monitoring
On Fri, 24 Jan 2020 at 07:40, David Berndt wrote: > >Is there any pin or ?? that can be monitored to give me > more insight? 2.8 has: joint.N.backlash−corr OUT FLOAT (DEBUG) - Backlash or screw compensation raw value joint.N.backlash−filt OUT FLOAT (DEBUG) - Backlash or screw compensation filtered value (respecting motion limits) joint.N.backlash−vel OUT FLOAT (DEBUG) - Backlash or screw compensation velocity 2.7 has: axis.N.backlash−corr OUT FLOAT - Backlash or screw compensation raw value axis.N.backlash−filt OUT FLOAT - Backlash or screw compensation filtered value (respecting motion limits) axis.N.backlash−vel OUT FLOAT - Backlash or screw compensation velocity -- atp "A motorcycle is a bicycle with a pandemonium attachment and is designed for the especial use of mechanical geniuses, daredevils and lunatics." — George Fitch, Atlanta Constitution Newspaper, 1912 ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Backlash comp monitoring
there is a small description in the source code: 'compute_screw_comp()' is responsible for calculating backlash and lead screw error compensation. (Leadscrew error compensation is a more sophisticated version that includes backlash comp.) It uses the velocity in emcmotStatus->joint_vel_cmd to determine which way each joint is moving, and the position in emcmotStatus->joint_pos_cmd to determine where the joint is at. That information is used to create the compensation value that is added to the joint_pos_cmd to create motor_pos_cmd, and is subtracted from motor_pos_fb to get joint_pos_fb. (This function does not add or subtract the compensation value, it only computes it.) The basic compensation value is in backlash_corr, however has makes step changes when the direction reverses. backlash_filt is a ramped version, and that is the one that is later added/subtracted from the position. 24.01.2020 09:38, Marcus Bowman пишет: It would also be very useful to understand the algorithm being used to apply backlash compensation. Is there a link to an explanation, somewhere? ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Backlash comp monitoring
On 24 Jan 2020, at 05:07, David Berndt wrote: > > I'm having a bit of an issue with some geometry not coming out as expected, I > was hoping that being able to see what backlash comp is doing on screen. It would also be very useful to understand the algorithm being used to apply backlash compensation. Is there a link to an explanation, somewhere? Marcus ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users