Help, Hurco 5 axis simultaneous - rippled surface.

[sgt_rock]

We have a Hurco VMX42HSRTi which is a head+table (B+C) machine. We are trying to cut our first true 5 axis part (Its an aerospace wing). The C axis is 24" diameter. The part is also about 24". If we machine the surface using just 3 axis or 4 axis (B tilted) the part looks great. When we try to machine the part using all the axis combined then it shows "ripples" in the part towards both ends of the outer areas of the surface (the furthest away from center of the rotary). The center areas look fine. We need to use the full 5 axis paths to completely machine the edges of the wing so we can't just use the tilted B axis by itself.

We also noticed that after cutting the part with all 5 axis and then re-cutting it using just 4 axis that there was about a .002 difference in the outer ends of the surface where it didn't clean up like it's not following the same contours.

The distributer came out to look at the machine but didn't find anything out.
We've tried changing the SFQ (Surface Finish Quality) parameter at the machine and it did not improve.
The part is programmed using tool axis vectors (IJK) and .0002" intol/outtol (same for all types of tool paths).

Has anyone done any true 5 axis simultaneous parts on their Hurco's? I'm wondering if it's the machine or the Hurco software in it.

Here is the surface using just 4 axis.


Here is the same surface using all 5 axis.


Here is a stock photo of the machine layout.

 

[Mtndew]

Is the part vibrating at all? Hard to tell from your photos how secure it's tied down.

 

[sfriedberg]

Hard to tell from your photos how secure it's tied down.

To me, it looks like the blank is clamped to the fixture in 10 places. So the periphery should have good support?

 

[empower]

are you unclamping B/C in your gcode in this toolpath? if its not, could be the brakes are turning on/off rapidly between each move which could cause this.

also try loosening up your toolpath tolerance to like .002" and see if the quality improves. if it does, the control is having a hard time keeping up with the code.

 

[MaxPrairie]

are you unclamping B/C in your gcode in this toolpath? if its not, could be the brakes are turning on/off rapidly between each move which could cause this.

also try loosening up your toolpath tolerance to like .002" and see if the quality improves. if it does, the control is having a hard time keeping up with the code.

I was also leaning towards a setting in the cam that would help improve the simultaneous code. What cam is this programmed on?

 

[BudgieW]

What does the code you are running look like?

 

[5 axis Fidia guy]

Is your rotary axis direct drive? or is there a gear of some sort? If it's not a direct drive you will never get rid of the surface blemishes no matter how your program and tolerances look. I've gone thru this in my Fidia days and glad they are over.

 

[boosted]

Is your rotary axis direct drive? or is there a gear of some sort? If it's not a direct drive you will never get rid of the surface blemishes no matter how your program and tolerances look. I've gone thru this in my Fidia days and glad they are over.

No argument that direct drive is better, but there are plenty of mechanically driven rotaries that leave flawless finishes. Some designs are far superior to others.

You are probably going down the right path though. B axis in the head is always difficult to control. Even on premium machines quality of simultaneous paths on the tilting head tends to suffer. That Hurco probably has a mediocre gear assembly in the head as well as the rotary, and probably not the greatest setup for lookahead and shape compensation.

Hard to say for sure this is a mechanical issue though. So many variables that nobody can make a confident assessment just going off the grainy photos.

 

[sgt_rock]

It's only when we start swiveling the B that it looks bad.

 

[sgt_rock]

It's the same part, same fixture, same tool for both versions of the toolpath. It's only if we swivel the B continuously that it has the bad finish.

 

[sgt_rock]

Here is a snippet:

T39 M06
(1/4 CARB BN, 3 FL)
S12000 M03
M128 (Tool Center Point Management ON)
G08.2 X14.8783 Y-2.9646 Z1.5 I-.336874 J-.172969 K.925525 (ASR TCP)
G43.4 (Toolpath Linearization)
M35 (B AXIS UNCLAMP)
M13 (C AXIS UNCLAMP)
M08
G90 X14.8783 Y-2.9646 Z1.5 I-.336874 J-.172969 K.925525
X15.6589 Y-2.5637 Z-.6448 I-.336874 J-.172969 K.925525
G01 X15.6631 Y-2.5558 Z-.6599 I-.336874 J-.172969 K.925525 F144.
X15.6645 Y-2.5416 Z-.6705 I-.336874 J-.172969 K.925525
X15.6642 Y-2.5248 Z-.6768 I-.336874 J-.172969 K.925525
X15.663 Y-2.5073 Z-.6798 I-.336874 J-.172969 K.925525
X15.6613 Y-2.4895 Z-.6803 I-.336874 J-.172969 K.925525
X15.6594 Y-2.4718 Z-.6789 I-.336874 J-.172969 K.925525
X15.6573 Y-2.4543 Z-.676 I-.336874 J-.172969 K.925525
X15.6551 Y-2.437 Z-.6719 I-.336874 J-.172969 K.925525
X15.6529 Y-2.42 Z-.6667 I-.336874 J-.172969 K.925525
X15.6528 Y-2.4197 Z-.6664 I-.336875 J-.172968 K.925526
X15.6527 Y-2.419 Z-.6658 I-.336875 J-.172966 K.925526
X15.6525 Y-2.4176 Z-.6645 I-.336875 J-.172961 K.925527
X15.6524 Y-2.4162 Z-.6633 I-.336875 J-.172956 K.925527
X15.6523 Y-2.4159 Z-.663 I-.336876 J-.172955 K.925528
Y-2.4158 Z-.6629 I-.336876 J-.172955 K.925528
Z-.6626 I-.336876 J-.172954 K.925528
X15.6522 Z-.6624 I-.336876 J-.172953 K.925528
X15.6521 Y-2.4157 Z-.6622 I-.336876 J-.172953 K.925528

 

[sgt_rock]

Problem solved

Well here is what the problem/fix was. The reseller came out again but this time they "tuned" the axis. If you can imagine what an EKG graph looks like, they moved the machine changing axis direction back and forth and noted the peaks and valleys on an internal graph, then modified some parameters to "smooth" out the peaks and valleys on the accel/decel for the B axis. My comment is that seems like something you should do when you first install the machine on a customer's floor. (We've had the machine for just over a year and have had some "motion issues" with it). The part smoothed out and looks good all over now.