Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Smooth motion from machining project

I am working with a Fanuc 200ic LR Mate and R30ia Mate Controller. It is set up for the vision system, but unfortunately the camera was damaged when the system was removed from production line. Translating g code to something the robot can understand is my focus right now. I am having trouble with the robot moving very slowly and appearing to almost stop on each point created by the software. Sometimes it even jerks between points. I am using fusion 360 for my g code source. I have tried using the "smoothing" feature that decreases the amount of g code created. I have also increased step values in fusion along with lowering tolerances in the hopes of decreasing the "resolution".  Playing around with the min and max step values in robodk don't seem to make a difference either (or i'm not doing it right). I'm guessing my problem is in the robots motion planner. Either a system variable needs adjustment or it simply can't handle the amount of points i'm sending it. I noticed at one point the created "ls" file was using "fine" points and then later it used "cnt1". I cant remember the changes I made when that happened. Is there a way to increase the "cnt#" amount? How bad will that affect accuracy? I'm guessing it would involve editing the post processor script? Or is there any other information that will help my situation? Thanks for any help and have a great day!
You obtain a CNT1 when you use a rounding instruction with a value of 1%. More information here:

This will affect accuracy, so you need to find a good compromise between accuracy and speed. Increasing the rounding value will make the speed remain more constant while accuracy will be worse. For Fanuc, the CNT represents a percentage. Most other robot controllers need a value in mm.

This instruction is added automatically when you activate the rounding in the program events from your robot machining project.

Another tip, you can filter points here to have less density of points along your curve:
Tools-Options-Program-Minimum step size (mm)
@Albert How does this relate to using  If we generate a .ls or run on robot from the .py script what would be the best way to edit the cnt value to avoid all points being set as fine?  

Would it be best to add a set speed command in the .py or to set the smoothing?  Do you have any examples of that anywhere?  Trying to synch 2 robots using the and its working wonderful to keep in sync, but they move quite slow. I may have missed the set rounding options in any of the speed calculations in the sync scripts if they are there.  

My other thought was similar to pwolfe, since we are generating the program off targets from  Would that lead to a post process edit as well?
Thanks for the info Albert! The rounding feature did the trick. I must have looked over that a dozen times or more.

I definitely need to find a way to expand the memory of the robot. There is only about 18MB available, and a small little test like the wooden bowl example takes up half of that. Lol. I remodeled a similar bowl in fusion. Using a parallel machining pattern with 5mm step over, it still creates programs fairly large.

Thanks again for the help, much appreciated!
We are aware of this memory limitations with some controllers. We have a post processor will help with this issue. The post processor is called:
Fanuc RJ3 DripFeed
It works on Fanuc RJ3 controllers and later (including R30iA, R30iB, etc).
This post processor automatically splits and feeds Fanuc LS or TP files to the robot over FTP. 

Added to the typical LS/TP programs that are generated for a Fanuc robot you'll find a PY script that you can execute to automatically load programs over FTP. This file manages the transfer (load/unload of programs) as your main program gets executed. If you select Ctrl+F6 (Send Program to Robot option), the program feeding is triggered automatically. More instructions are provided in the console window when you run it.

More information about how to select a post processor here:


Users browsing this thread:
1 Guest(s)