Having problem with blending arcs together

[Nandi]

Hi Jochen!

  I'm developing a 3-axis plasma cutter software. It's uses g-codes for calculating the path of motions and using blend function for continous moves. Everything works fine including continously blending short straight line moves, but sometimes blending arcs resulting high following error. Mainly when i try to blend little arc segment with a big one. This problem doesn't occur when i set blend factor to 0. I tried many different trajectory and arc parameters with the "Sequence of Blended Lines and Arcs.vi" example, but it resulted the same effect. With some trajectory parameters i can blend the short arc segment with the big one without following errors, but i still don't know what causes this problem. This situation is same on two different cnc machine. One of them is using 1 servo and 2 stepper axis in closed loop with different encoder resolutions. The other one is using 3 stepper axis with different step/rev in open loop. The effect is the same. Blending little arc with a big one (approximately r1:r2=1:8), resulting following error and the motors jumps a little.

- PCI-7340, umi-7774
- LabView 8.5, Motion 7.6

Waiting for hints! Thank you!

[Nandi]

Sorry for my bad english! Ask me if you need further informations for solving this case.

[Jochen]

Nandi,

could you please provide a screenshot of a position vs. time plot and a following error vs. time plot? This would help me to better understand the problem and hopefully to locate the occurrence of the following error within the move sequence. It would be great if you could provide these plots for all axes seperately (but taken from the same move).
Additionally it would be great if you could provide the exact move parameters for the arcs (radius, velocity, travel angle, start angle).

How do the arcs blend into each other? Is the result a sharp angle in the contour or do they transition smoothly?

Thanks and kind regards,

Jochen Klier
National Instruments

[Nandi]

Jochen,

thank you for dealing with my case. I attached the vi, which have the plot you asked and demonstrates the occurence of the position error. The trajectory parameters and some notes are in the vi. I hope there is a solution for the problem.

Yours sincerely,
Nandi

[Jochen]

Nandi,

sorry for the late response, but I have had to gather some equipment first to reproduce this behavior. I have setup a very similar system with a servo and a closed loop stepper and I have used your code for a test.

The bad news is, that I couldn't reproduce the problem. First I have worked with two servos but I couldn't see a significant following error with those, but in fact I wouldn't have expected a sustaining following error with a servo at all, as servos can compensate following errors during the moves.
In a next step I have used a servo and a closed loop stepper but I have seen the same result. No significant following error during the move. I have also switched the X and Y axes, but there was no change. I have also played around with the other move parameters but again, no significant following error.

With this said, I don't think that we are facing a general issue here that is related to 73xx motion control boards.
To me this looks more like an issue that is related to problems with load inertia and high acceleration values. In fact if you reduce the radius of an arc but keep the vector speed, the forces become higher and in your case they seem to become too high for the stepper motors. A stepper motor can't compensate step loss during the move. That's why you see this constant following error.

I hope this analysis hits the bull's eye.

Kind regards,

Jochen

[Nandi]

If someone can help me i would really appreciate it. I'm running out of time and i can't find out the reason of this malfunction.

[Nandi]

Jochen,

  Thank you for your investigation. Maybe i must write some code for filter out this possibility and reduce the trajectory parameters for the arc with little radius. I have to do some test for the corresponding acceleration and velocity values.

Thank you for your help, and i'm trying do my best.

[Jochen]

Hi Nandi,

maybe I have to revise my statement in my last post. In fact I was able to reproduce the problem by adding additional load to the stepper axis and increasing the velocity, but this doesn't seem to be the point here.

Now I have done another test:
I have decreased the velocity to extremely low values and voila - now I can see the axes' position jumping. In my case this is not enough for a sustaining following error, but I can see the quick movement with both, stepper axes and servo axes.

I will pass this information to our motion control R&D team for further investigations. I'll be back to you as soon as possible.

Regards,

Jochen

[Jochen]

Nandi,

if it turns out, that there is really an issue in our boards' firmware, it might be a good idea to use either the workaround that you have suggested or to switch over to contouring (please refer to the shipping examples).
Especially with short moves, contouring could be a better option for your application anyway.

Still we will do further investigations and I will let you know about the  results.

Thanks,

Jochen

[Nandi]

Jochen,


   thanks for your help. I will wait for the result. I have to reject the contouring mode, because it means i have to rethink the whole software i wrote (etc.: vel. override, motion execution architecture..) And in the other hand we doesn't noticed any problems with blended shortmoves after reading the blend complete status at maximum rate. Anyway thanks for all the collaborators for the investigation who affected in this case. I have to work on another project until the result is done.