I have a Kuka KR100 with KRC2 controller and I have a HSD 7.5kW spindle attached to the end of it. I can cut light foams (Airex and Rohacell) very smoothly but when I cut high density epoxy tooling block I get a "pattern" on the finished surface. I am using MasterCam with Robotmaster software. Can anybody advise me on the best robot settings for smooth surface machining? I have attached an image of the surface pattern I don't want as it was machined on a small block of tooling block. Are there any settings that smooth the spindle movement?
Kuka CNC Milling
-
PrecisionPete -
July 1, 2013 at 4:58 PM -
Thread is marked as Resolved.
-
-
I'm getting similar pattern as well with my KR200-KRC2. I don't think it is a matter of software. We are using Solidworks, MasterCam, Rhino, Powermill.....I am starting to believe that we have reached the limits of a non-High Accuracy robot.
-
I'm not a milling expert, but from what I'm told, patterns in the milled surface are usually due to a mismatch between cutter RPM and feed rate for a given material. Also, since the robot is simply not as rigid as a CNC tool, it may be necessary to use much lower feed rates than a CNC machine would use for the same material. If the "recoil" from the cutter engaging the workpiece is too high, the robot might flex and create a surface pattern as the tool "chatters" against the work surface.
-
What's your "resolution" in Mastercam?
I've seen it a few times before (bending, milling etc. etc. etc.) - if your res. is set to low (0.01 or lower) it'll create a sh*tload of points within "nothing" - been testing it, with "higher" res. - which "smoothens" it a bit - but of course - at a "lower" tolerance...I just made a large tubular laser installation - on a simple cut, the difference between res. 0.01 and 0.1 (ish) were something like 100 lines of points...
cut'n'paste of similar problem at customer - some geeky bending software...
simple process - strange solution...Code
Display MoreDECL E6POS P2AA1[16] P2AA1[1]={x 1375.85999,y 2647.98999,z 1550.40002,a -116.599998,b 2.31999993,c 179.539993,s 2,t 35,e1 1891.31006,e2 0.0,e3 0.0,e4 0.0,e5 0.0,e6 0.0} P2AA1[2]={x 1379.30005,y 2478.75,z 963.090027,a -116.599998,b 2.30999994,c 179.539993,s 2,t 35,e1 1891.31006,e2 0.0,e3 0.0,e4 0.0,e5 0.0,e6 0.0} P2AA1[3]={x 1817.79932,y 2467.70703,z 1786.98242,a -116.599998,b 2.33999991,c 179.539993,s 2,t 43,e1 445.480011,e2 0.0,e3 0.0,e4 0.0,e5 0.0,e6 0.0} P2AA1[4]={x 1817.79932,y 2467.70703,z 1786.98242,a -116.599998,b 2.33999991,c 179.539993,s 2,t 43,e1 445.480011,e2 0.0,e3 0.0,e4 0.0,e5 0.0,e6 0.0} P2AA1[5]={x 1817.79932,y 2467.70703,z 1786.98242,a -116.599998,b 2.33999991,c 179.539993,s 2,t 43,e1 445.480011,e2 0.0,e3 0.0,e4 0.0,e5 0.0,e6 0.0} P2AA1[6]={x 1817.79932,y 2467.70703,z 1786.98242,a -116.599998,b 2.33999991,c 179.539993,s 2,t 43,e1 445.480011,e2 0.0,e3 0.0,e4 0.0,e5 0.0,e6 0.0} P2AA1[7]={x 1817.79932,y 2467.70703,z 1786.98242,a -116.599998,b 2.33999991,c 179.539993,s 2,t 43,e1 445.480011,e2 0.0,e3 0.0,e4 0.0,e5 0.0,e6 0.0} P2AA1[8]={x 1817.79932,y 2467.70703,z 1786.98242,a -116.599998,b 2.33999991,c 179.539993,s 2,t 43,e1 445.480011,e2 0.0,e3 0.0,e4 0.0,e5 0.0,e6 0.0} P2AA1[9]={x 1817.79932,y 2467.70703,z 1786.98242,a -116.599998,b 2.33999991,c 179.539993,s 2,t 43,e1 445.480011,e2 0.0,e3 0.0,e4 0.0,e5 0.0,e6 0.0} P2AA1[10]={x 1817.79932,y 2467.70703,z 1786.98242,a -116.599998,b 2.33999991,c 179.539993,s 2,t 43,e1 445.480011,e2 0.0,e3 0.0,e4 0.0,e5 0.0,e6 0.0} P2AA1[11]={x 1817.79932,y 2467.70703,z 1786.98242,a -116.599998,b 2.33999991,c 179.539993,s 2,t 43,e1 445.480011,e2 0.0,e3 0.0,e4 0.0,e5 0.0,e6 0.0} P2AA1[12]={x 1817.79932,y 2467.70703,z 1786.98242,a -116.599998,b 2.33999991,c 179.539993,s 2,t 43,e1 445.480011,e2 0.0,e3 0.0,e4 0.0,e5 0.0,e6 0.0} P2AA1[13]={x 1817.79932,y 2467.70703,z 1786.98242,a -116.599998,b 2.33999991,c 179.539993,s 2,t 43,e1 445.480011,e2 0.0,e3 0.0,e4 0.0,e5 0.0,e6 0.0} P2AA1[14]={x 1817.79932,y 2467.70703,z 1786.98242,a -116.599998,b 2.33999991,c 179.539993,s 2,t 43,e1 445.480011,e2 0.0,e3 0.0,e4 0.0,e5 0.0,e6 0.0} P2AA1[15]={x 1817.79932,y 2467.70703,z 1786.98242,a -116.599998,b 2.33999991,c 179.539993,s 2,t 43,e1 445.480011,e2 0.0,e3 0.0,e4 0.0,e5 0.0,e6 0.0} P2AA1[16]={x 1817.79932,y 2467.70703,z 1786.98242,a -116.599998,b 2.33999991,c 179.539993,s 2,t 43,e1 445.480011,e2 0.0,e3 0.0,e4 0.0,e5 0.0,e6 0.0}
btw - that .DAT is 5073 lines...
-
I would think that it is connected with the feedrate (what SkyeFire suggested) - why don't you try on lower feed and the same RPM? Or you did? Can you make a video?
It also depends what is the length of the tool? May be you can try with a shorter tool?
I have KR150 -2 with 12KW HSD and it is flawless - even with a hand made postprocessor - the end result is what i am aiming usually -
Also, with a program that is made up of several thousand very small motions, it is possible that the robot path planner simply cannot maintain a smooth continuous-speed motion profile through all these points. I would suggest running an oscilloscope trace of the robot's motion through this program and look at how consistent its speed is.
Also, if your KSS version is high enough to support it (5.6 for certain, 5.5 maybe), you could try importing these motions into a SPLIN motion block. Although, I would recommend deleting most of those extra points in favor of just ones that have a significant motion delta between them.
-
thats actually pretty smooth for a robot(and I have a KR100HA). Skyefire is right, you may be able to slow down, as it does seem to be some sort of vibration; could also be your approximation settings. If it were simply axis 1 backlash(which is a huge source of positioning error), it would probably be more of a step.
what are your feed and speed, cutter size etc? send me the part, I will make it perfect on my haas VF2SS
-
Problem resolved. Thank you to everyone who posted help. It was a feed related issue. When I was able to control the feed rate I was able to cut a really smooth finish to the surface of my epoxy tooling block test piece. I slowed the feed rate down to 20mm/s and the surface finish is perfect. There was a bug in my software that did not let me modify the feed rate as expected (robot default $VEL.CP=3 was dominate everytime)
My Kuka KR100P with KRC2 controller now appears to be a very accurate, smooth and controllable 3-axis CNC machine. I have a few more tests to do with various other materials (wood, ply-wood, TB650 and PU foam) and with 5-axis surface tool paths but I'm happy with the results so far.
Cutting at slower speeds I have no chatter, vibrations or recoil. -
well $VEL.CP=3 is really too much... i am cutting MDF with $VEL.CP=0.06 - 0.07 and sometimes i need to slow it down
keep it cutting!
-
Now I can control the feed rate I am getting some good, smooth results.
I have used a $VEL.CP of between 0.1-0.02 to great effect on wood and epoxy tooling block (e.g. TB650)
I still have a few more variables to play with and test and I'm constantly looking for new materials to cut. But my robot/spindle combination already meets my expectations/needs. -
glad to hear you got it sorted out. I have to check this forum more often.
another couple things you can do is change the approximation value.
The Vel.CP maxes out at 3 and usually is the result of entering in in/min into Mastercam when the Kuka is in m/s (but then there's the config switch to convert the values)
I've actually been able to get some pretty amazing finishes in aluminum with robots! The 'high speed' toolpaths in mastercam are incredible for this because they drastically reduce the spindle-load. they make very fast low-load 'peel' cuts. The downside is they produce tons of code but its worth it
where are you located Pete?
-
Hey Tyler, can you post some video or at least images? I am thinking to test my robot in aluminum but i am a bit afraid to do so not sure about the right settings...