Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5

Sporadic path deviations between Fusion360 and reality

#1
Sporadic path deviations between Fusion360 and reality
 
after a long time of experimenting and building, my milling machine with a KR360 (KRC2 edition 05) and a self-made turntable is now ready and works so far well. In most cases, the milling tool follows the specified path without any problems and with sufficient accuracy. Unfortunately, sometimes there are movements that do not exist in the simulation that destroy the workpiece.
But first from the beginning:
   
This is how the CAM calculation looks in Fusion360. The tool path is correct and not too much material is removed during the roughing process.
   
Here again a little more detail. The chin region is cleanly roughed up.
   
This is how the path looks in RoboDK. Even though it's hard to see, the path looks good so far. I can't see any movements that would explain the faulty milling to me.
   
To be on the safe side, I also loaded the model of Fusion360's simulation into RoboDK, no path penetrates the model.
   
In reality, the result looks like this. At the point marked in red, the milling head moves much too deeply into the material. Such errors happen again and again on the test parts. Fortunately, this error has already occurred in line 5500 and I have observed it with my own eyes. The milling tool repeatedly enters this position cleanly. A sporadic electrical fault can thus be ruled out. It looks like material has been removed from the blue arrow in the wrong place one layer before. First I suspected that the $ADVANCE =3 is too large and therefore the accuracy is too low. Unfortunately, even a $ADVANCE =1 brought no change, the tool moves exactly the same in the red marked zone
Now I don't know how to solve this (hopefully small detail). Otherwise everything runs pretty well!

This is how the code looks:

Code:
&ACCESS RVP
&REL 4
&PARAM TEMPLATE = C:\KRC\Roboter\Template\vorgabe
&PARAM EDITMASK = *
DEF Pusteblume ( )


;FOLD INI
;FOLD BASISTECH INI
GLOBAL INTERRUPT DECL 3 WHEN $STOPMESS==TRUE DO IR_STOPM ( )
INTERRUPT ON 3
BAS (#INITMOV,0 )
;ENDFOLD (BASISTECH INI)
;ENDFOLD (INI)

;FOLD STARTPOSITION - BASE IS 17, TOOL IS 1, SPEED IS 45%, POSITION IS A1 55,A2 -50,A3 50,A4 5,A5 0,A6 -5,E1 0,E2 0,E3 0,E4 0
$BWDSTART = FALSE
PDAT_ACT = {VEL 45,ACC 100,APO_DIST 50}
FDAT_ACT = {TOOL_NO 1,BASE_NO 17,IPO_FRAME #BASE}
BAS (#PTP_PARAMS,80)
PTP  {A1 55,A2 -50,A3 50,A4 5,A5 0,A6 -5,E1 0,E2 0,E3 0,E4 0}
;ENDFOLD

;FOLD LIN SPEED IS 0.5 m/sec, INTERPOLATION SETTINGS IN FOLD
$VEL.CP=0.5
$ADVANCE=3
;ENDFOLD



PTP {A1 -23.33140,A2 -49.55890,A3 104.71900,A4 152.47600,A5 58.26180,A6 -74.75400,E1 0.44815,E2 0.00000} C_PTP
LIN {X 34.437,Y 205.000,Z -185.360,A 180.000,B -0.103,C -90.000,E1 0.44815,E2 0.00000} C_DIS
$VEL.CP = 1.00000
LIN {X 34.437,Y 175.000,Z -185.360,A -180.000,B -0.103,C -90.000,E1 0.44815,E2 0.00000} C_DIS
...
...
LIN {X -99.712,Y 305.000,Z -316.287,A -180.000,B -0.103,C -90.000,E1 0.44815,E2 0.00000} C_DIS

END
I would be very happy if someone with more experience with RoboDK could give me some tips.

Chris
#2
Hi Chris,

First of all, nice project.

To be able to help, I would need to have access to your RoboDK Station. Could you provide that, please?

Jeremy
Find useful information about RoboDK and its features by visiting our Online Documentation and by watching tutorials on our Youtube Channel


#3
Hi Jeremy, 

Thank you for your answer and willingness to help me with my problem. 

In the middle of the week I will have the possibility to access the robot again. How can I give you access to my RoboDK station? Should I upload the RoboDK file somewhere? 

Greetings 
Chris
#4
You can attach the .rdk file to your reply on the forum, select the file, click upload, then attach it to your post
#5
(04-26-2021, 08:53 AM)colinb83 Wrote: You can attach the .rdk file to your reply on the forum, select the file, click upload, then attach it to your post

Okay, that's easy. I thought the file was much larger.

Here is my station. The turntable and fixtures are not quite complete in the model, but that shouldn't be a problem.

Chris


Attached Files
.rdk   Forum.rdk (Size: 8.63 MB / Downloads: 337)
#6
Can you provide the .apt file you imported?


"Lindemann_low_poly_decimate_unterkonstruktion v8.apt"

There's a few spots where the path seems odd:

   

The issue could come from the APT, or the way RDK imported the APT. 
My goal would be to manually split the APT file into smaller chunks to better see the path in RoboDK. 

Jeremy
Find useful information about RoboDK and its features by visiting our Online Documentation and by watching tutorials on our Youtube Channel


#7
That's a really interesting point. Since I don't know which APT based post processor is running in the background of the Fusion360-RoboDK add-in, I manually tested several post processors.
Here is the comparison between the import via RoboDK Add-In and the manual import of the APT file with the postprocessor APT ISO 4343. The path looks much smoother without the marked edges.

   
 
   
 
Next I tested the postprocessor APT CAMplete, here the path is visually almost identical to the automatically imported one. Including all the edges that may have caused the problems.
 
   
 
In Fusion 360, the toolpath looks good, without any deviations. Now I have a hard time telling if Fusion's postprocessor is causing the errors, or if RoboDK is misinterpreting the APT code. For me it would also be no problem to import the path always manually into RoboDK, if it then runs properly.


Chris


Attached Files
.apt   APT ISO 4343.apt (Size: 1.89 MB / Downloads: 297)
.apt   APT CAMplete.apt (Size: 1.84 MB / Downloads: 294)
#8
Hi Chris,

We switched from the Fusion CAMplete post to the ISO4343 post a few months ago as we noticed these issues which are linked to arc movements. I'm sorry to hear the CAMplete post destroyed some of your pieces.

I recommend you to update RoboDK which should update the Fusion plugin:
https://robodk.com/download

Alternatively, you can simply update the Fusion plugin.

Let us know if you still see issues.

Albert
  




Users browsing this thread:
1 Guest(s)