Jerky Motion with Small Circular Movements

testyourdesign

testyourdesign
I noticed that my mill is not running as smoothly as it used to. I get this jerky motion when when ever a circular movement is requested. Its very noticeable with adaptive clearing toolpaths from Fusion360 but it also shows up other processes as well and its leaving tool marks in the finished parts as well. I tried changing the settings in fusion360 but nothing changes. You can see and hear it in this video I posted on Instagram. Has anyone else noticed jerky motion with small circular movements with V3.42 and backlash compensation enabled?

Cheers, Stephen Brown
 

testyourdesign

testyourdesign
Well after a bunch of research into Fusion360 CAM Smoothing Functions and Tolerances I determined that the program is working perfectly fine. I just tested my system with backlash set to zero and all the jerky motion went away. Looks like V3.42 backlash compensation is not happy with the start of circular or arc motions. I am not sure if its isolated to imperial units or if its the same with metric systems as well but I definitely have an issue. Its staying set at zero for now.

Cheers, Stephen Brown
 

masso-support

MASSO Support
Staff member
with backlash compensation, there will be some delay in motion when switching direction because to counter the backlash the axis needs to move to the opposite direction first and then move the required movement as requated by gcode. The small delay that you see in the video is when the axis changes its direction to remove axis backlash.
 

testyourdesign

testyourdesign
@masso-support

I don't have a problem with the program slowing down to make the adjustment but the abrupt movement to compensate for the backlash is causing tool marks on the finished parts. I slowed the motor acceleration down to 1 inch/sec^2 and it improved slightly but it still gouged the parts as the change in speed causes the endmill to dig into the metal.

Is it using a G1 command to make the move for compensation and the same feed rate or is it using a G0 at the rapid feed rate? Its needs some sort of acceleration control to make the cut less interrupted.

Cheers, Stephen Brown
 

Attachments

  • 20190903_210033.jpg
    20190903_210033.jpg
    527.1 KB · Views: 14
  • 20190903_224024.jpg
    20190903_224024.jpg
    887.3 KB · Views: 9
  • 20190903_224110.jpg
    20190903_224110.jpg
    389.5 KB · Views: 14
  • IMG_20190903_114524_334.jpg
    IMG_20190903_114524_334.jpg
    549.1 KB · Views: 12

testyourdesign

testyourdesign
Its different on each axis with X = .0147, Y=.0049, Z=.0115 inches. I tested them individually with .010 one at a time and found the same issue however, it seamed to occur on the opposed direction. When I added .010 on X the issue happened when it moved in Y direction.

Cheers, Stephen Brown
 

evermech

evermech
Hi @testyourdesign

are you climb cutting or cutting against the feed direction? Might make a difference which way the cutter is pushing or pulling on the backlash in your system. Maybe try a couple simple circles with the feed going the opposite direction just as an expirement. You might need to feed in one direction for inside geometry and the other for outside profiles. It also could be a benefit to uses shallow depth of cuts, combined with higher spindle speeds and feed rates to lower the cutting forces minimizing the effects of tool deflection and rigidity issues.

Guy
 

testyourdesign

testyourdesign
@evermech

Thanks for your response. Its definitely more noticeable on roughing cuts but it also happens on finish passes as well. I am not being aggressive with my cuts and it never did this with the older version of the software before backlash compensation was implemented. The gouging issue goes away when I set the backlash compensation to zero so I know its not my program or the feeds-n-speeds I used.

I realize that it's a complicated feature for them to implement but it needs a little more work to make it right for production. The feature basically works well but it needs refinement so that the transition between the various segments happens a little smoother.

Cheers, Stephen Brown
 

evermech

evermech
@testyourdesign

Could be interesting to try the different feed directions just to see what the heck it does

just thinking. Otherwise sounds like you have everything in control. Good luck

Guy
 

testyourdesign

testyourdesign
@evermech

I'm just using climb milling like most CNC machines. I'll run some conventional milling test cuts tomorrow to see if there is any benefit but the benefit is outweighed by the cost if I have to spend a bunch of time to split the program into various cutting passes to compensate for the backlash. Based on what I saw today, I'm going to look into improving the actual backlash on my X axis and Z axis then try to run without software compensation for the time being.

Cheers, Stephen Brown
 

evermech

evermech
@testyourdesign

i just finished reading up on backlash and backlash compensation. It appears that climb cutting makes everything worse, but of course that's how you want to cut most everything. Not an easy fix for sure.

Guy
 

machinedude

machinedude
i think working the backlash out on the mechanical end is a good plan. even .005 on the best axis is getting up there but getting the other two at least that good should help improve things some. i always liked having a double nut that could be pre-loaded in opposing directions but with those small mills it's tough since the travel is limited to begin with. not sure what kind of lead screw is driving each axis either?

one other thing that might help is slowing the feed rate on the lead ins and outs and maybe extend you finish passes beyond the starting points. that's something you can set in fusion 360 so it should be easy to try some different settings. maybe even a spring pass on your finishing pass would probable help some.
 

testyourdesign

testyourdesign
Thanks for all the feedback but I think some of you might have missed the point of this post. I can easily spend more money to tweak the mechanical functions of my machine or live with what I have without major concern but other users may not have that luxury. I am more concerned about the way backlash compensation has been implemented on the Masso Controller as my simple hobby machine is not critical to the community of Masso Users.

It is my belief that the numbers I posted should not have an effect on the performance of the software or controller. If the values I used are to large then Masso should have established the limit and notified us. I am not familiar with the way Masso implements this feature within the software nor do I really want to know but given my feedback they might be able to determine a better way to implement this feature or adjust the way they implement the function in order to improve the product for the community as a whole.

Considering that they chose to perform the backlash calculation at the end of each move means that the software will always have to pause before implementing the next move. From my machining experience I know that a pause in motion causes tool marks in the parts. A machinist is taught to place those pauses in motion at strategic locations in order to minimize the effect of the tool marks on the finished parts. As a software and controller manufacturing expert dealing with a wide range of trades, Masso may not be familiar enough with detailed Machining techniques to realize that they implemented a function which will be rejected because it does not meet a basic parameter that is demanded of the operator. In my experience tool marks are cause for rejection by quality control in most cases so they may have an uphill battle on their hands.

As always my goal is to provide the feedback they requested while gaining a better understanding of the science of CNC machining and helping others avoid the pitfalls I've experienced.

Cheers, Stephen Brown
 

evermech

evermech
As always my goal is to provide the feedback they requested while gaining a better understanding of the science of CNC machining and helping others avoid the pitfalls I've experienced.

Cheers, Stephen Brown

This is why I think it could be a good idea to try both an inside and outside circular cut using a conventional machining technique, to test the ability of the backlash compensation. Then @masso-support will know for sure what they are up against

Guy
 

testyourdesign

testyourdesign
If you look at the video I posted you will see that I did machine both inside and outside featured using conventional machining approach. I am going to start with a fresh look at my motor tuning to make sure its still dialled in properly and then take the ball screws apart, clean them, and reassemble them to make sure everything is secure and functioning properly.

Cheers, Stephen Brown
 

evermech

evermech
Sorry Stephen I didn't see the videos otherwise I would not have pestered you about it.

im thinking the issue is in capable hands.

Regards Guy
 

testyourdesign

testyourdesign
Okay just finished double checking everything and made some adjustments to both my motor settings and backlash settings. For some reason the numbers were way off. Based on the new measurements the actual backlash is much less than my previous settings. I now have them set to X= .005, Y=.0052, Z=.0099. I know the Z axis looks loose but I am using a timing belt on that motor so the added backlash is understandable. I posted a couple videos on my YouTube channel that show how I did it.

Now to run some more parts to see if they run a little smoother this time. :)

Cheers, Stephen Brown
 

testyourdesign

testyourdesign
I like the way all the Masso settings are so easy to understand. It all works very well. Here are links to the videos.




Cheers, Stephen Brown
 

testyourdesign

testyourdesign
@masso-support

Ran some test cuts after double checking the backlash values and noticed an improvement in the finish. The previous gouges might be a symptom of having to high a value for the backlash settings. With X=.005 and Y=.005 the Mark's are much smaller.

Cheers, Stephen Brown
 
Top