about scara joint 3 issue ...

the image .... of belt drive C axis ..... the other machine is equal but gear driven C axis ....

Over the "cilindrical carter" there are a motor directly connected to a planetary gearbox ... direct connected on joint0 .... under the second "carter" there is another engine and gearbox connected by belts or gears ( depending on model ) to the joint1 .... joint3 / C axis is seen and is without housing ... has a gearbox that connects to the pipe ( end effector ) that runs ....

With image is more clear I think ...

Regards
Giorgio

andypugh wrote:

Is that the C-axis motor on the bottom of the second joint?

Yes...

andypugh wrote:

Is it possible you simply have the sense of rotation of the C axis motor wrong? (Or, at least, not the direction Scarakins expects)

I think scarakins expects that if joint0 go to ++ and joint1 go to ++ ... joint3(C axis) go to ++/++/+joint3 .... what you means for??

the problem:



pheraps at every velocity (proportionally .. less at slow motion ... more at hight motion .... F5000 = slow F90000 = hight)...
regards
Giorgio

UPDATE 1: I replaced # DEFINE PM_PI with #DEFINE PI = 3.xxxxxxxxxx of scarakins ...... the situation has improved slightly but little by little ...

andypugh wrote:

If green is what you want and red is what you get

This is right ......

andypugh wrote:

C-axis scaling? Have you checked that 1 revolution is exactly one revolution?

This is just check ... it's all ok.

tommylight wrote:

Not my forte, but, if the black are the robot joints, green is what you need it to do and red is what you get, that is a sign of proportions and lengths not being set properly, also the geometry of the arms and position in regards to one another, and moving angles.

you're right , it's the first thing I thought of me as well ... but the parts are manufactured in our workshops and the error is +/- 0.02mm ... while I see an error of 4-12 ° depending on the speed .... only for test 3week ago I try to insert an erron on leght of arm .... on a 10mm error on arm off 500mm, it is barely visible to the naked eye on the axis " C " while you see a lot regarding the joint and joint 0 1 ....

At these point I think the problem is max step acceleration problem related .... correct me if I'm wrong , but if I have a closed loop , any error ( acceleration and speed ) the error is corrected from the closed loop in LCNC / JA .... but if I have an open loop ... if the acceleration of ' axis " C " was too different from acceleration of joint1 + joint0 ???? What happen??

I do not know the answer to this ... but I figured LCNC / JA could have problems ... I try to play with these parameters now ...

UPDATE2: joint0 and joint1 is set STEPGEN_MAXACC = 9000 ... in the old test I set C axes STEPGEN_MAXACC = 9000 .... now I set C axes STEPGEN_MAXACC =300 STEPGEN_MAXVEL =200 MAX_ACCELERATION = 220 MAX_VELOCITY = 180 ..... same error on the same situation .... Now I'm sure lLCNC / JA adjusts the acceleration and deceleration even in the step / dir without any type of error according to the kinematic choice .... so it is not the problem ....

Regards
Giorgio