Inaccuracy in Rotary axis

Hello..everyone.

I am using LCNC 4 axis router or a mill I can say.
Purpose: Decorative engraving on silver items like dish, glass etc.

Now the rotary A axis is not accurate enough. The positional repetative inaccuracy found almost 1 degree which for a 2 inch dia glass is almost 0.5mm. Being engraved using very narrow diamond dresser, inaccuracy is very prominent and it is visually not acceptable.

Machine runs with RT-preeempt kernel, two parallel port, software step generation.

Now, there was trip issue with joint.3 i.e A axis. Ferror=1 and Min_Ferror=0.5 which allows feed of 650 and rapid upto 1000. Thre is a 6 inch self centering chuck mounted on bearings and stepper motor directly fitted into, ratio: 1:1. Motor fitted=

Other axes i.e X-Y-Z is not having any issue and accuracy is acceptable there.

Now, what is wrong? is it rotational inertia? For information, there is mild torsional vibration during homing in A axis and motion is not very smooth either. Presntly microstepping is at 64. Step rate does not permit higher microstepping.

how to get accurate result? Will a closed-loop stepper solve? Or I need a bigger motor?

How is the 4th axis driven? have you checked for backlash or binding?

Also it looks to me like you do not have enough reduction ratio
in the 4th axis. Your steps per degree scale is only ~35 and this is
close to 1/2 of a full step at 64 usteps/step which is the limit of accuracy
of the step motor (1.8 degree per step but 0.9 degree accuracy)
This 0.9 degree error spec on the step motor translates to ~0.8
degrees uncertainty on the 4th axis.

Well..my A axis photo is attached.
The stand (mushroom white color) has two ball bearing. the motor is directly coupled. the keyway is precision and no backlash.
In addition I reduced A axis velocity and acceleration and increased step/degree.

Now there is two issue:

No.1= Program trip:-
Despite reduced velocity and acceleration, program still trips with Min_Ferror=0.25. INI file attached.
But same program completes with Min_Ferror=0.4 with no other changes in parameters.

No.2=A axis inaccuracy:-
As already in earlier post, there is about 1 degree error in repeatation of the same line on job.


Question 1: how Min_Ferror affects accuracy??

Question 2: will a reduction ratio of say 5 or 10 solve the error issue?? If yes, a timing bely system will work or not??

Direct drive 1:1 motor coupling will never work for a rotary axis on a milling machine. The motor will not have enough holding power or accuracy. Most are geared well below 20:1 or 100:1 or more in servo applications.

Well, I am planning for a servo gearbox of ration 20 or 25.

Will that increase the torque enough??

Dear friends,

In continuation to earlier posts, I added a 1:10 ratio servo gearbox and a 12kg-m holding torque closed loop stepper to my rotary.

But result is not all ok.

Please refer the attached photos. There is clear deviation in engraving path. some straight lines are generated in place of following the curve.

The axis configuration is also attached. G-code file attached.

In X (having open loop stepper): 1mm=640 steps and with 75 dia job that I am doing now, in A axis (having closed loop stepper), 1mm=1015 steps.

Latency is below 8k.

What is wrong I am doing? thanks in advance for any help.