Need Help!- Part gouging

[Rumm64]

Hi
I'm new to the forum but have been lurking for some time.
Found lots of useful information here.

Now we've run into some troubles at our Fadal machines and
thought I'd post to get ideas to what may be happening.

First off please take a look at the part I've pictured, and you'll see
the part gouging.

A little info on what we're running:
Fadal VMC6030 Machine
Using Unigraphics NX4
MDFA post pocessor

The fisrt picture is a Z LEVEL finish pass with .017 step downs

The second picture is a CONTOUR AREA pass ( 1 continous pass)
You'll notice the tool path looks to be bouncing in and out of tolerance.

All tolerances are set to Part intol 0 -- Part outtol .001
and same for Bountry Tol. I have cranked down the tolerances with little to no affect.

I've actually run the PTP file thru Vericut and these under cuts do NOT show up.
All I can come up with is how the Fadal is interpret the data.

And heres the real kicker -- we run the same toolpath at your Mazak machine and there is NO gouge marks in the part.
I'm just looking for ideas as to what maybe happening here.
Any suggestions????

Thanks

[scadvice]

...is the Fadal and the Mazak. Just wondering. Steve

[carbidecraters]

I would check Z for backlash. I wonder if you are getting some kind of noise from an outside source effecting your servos. I dont know if its possible or not but I would also check the motor coupler.

[FUN4ME]

We had a similar problem.
While cutting on the X axis we were getting ripples on the finish.
We put an indicator on the Z axis and set it on the front of the part
Without manually moving the machine it would oscillate about .001 inch
You can check all of your axis this way.
We had the resolver or revolver (what ever you call it) replaced on the Y axis and the problem went away.
Hope this helps

[Delw]

Those look pretty deep like around .005-.010 with a ball endmill.

I am leaning towards a program problem, vericut is nice but it wont show small cuts that well. do you get that same pattern in all the parts that run that program and in the same area?

your post processor could be set up wrong for the fadal.

I was leaning towards the machine being sloppy or bad resolver however the parts have the marks on all axis's x and z and it looks also like on y in the first pic.

you can check this easy enough by running a 20 inch x move on an indicator and a 10 on the y. run it like 20 inch a min and then once you did it put hand pressure on the machine.

I do lots of surfacing and I am realy thinking its the program.

one other thing if your not running and g9 and your taking heavy cuts the ramping up and down may be a problem with tool pressure. I just assume your running a g9

[makingchips]

Originally Posted by Delw Those look pretty deep like around .005-.010 with a ball endmill. I am leaning towards a program problem, vericut is nice but it wont show small cuts that well. do you get that same pattern in all the parts that run that program and in the same area? your post processor could be set up wrong for the fadal. I was leaning towards the machine being sloppy or bad resolver however the parts have the marks on all axis's x and z and it looks also like on y in the first pic. you can check this easy enough by running a 20 inch x move on an indicator and a 10 on the y. run it like 20 inch a min and then once you did it put hand pressure on the machine. I do lots of surfacing and I am realy thinking its the program. one other thing if your not running and g9 and your taking heavy cuts the ramping up and down may be a problem with tool pressure. I just assume your running a g9

I have to agree. It looks like you were milling with the center tip of a ball endmill. If this is the case the endmill is deflecting due to plowing the material.

MC

[Rumm64]

OK thanks for the replies

To answer a couple ?'s

Mazak is a 1992 model
Fadal vmc6030 is a 2001
Fadal vmc8030 is a 1996
Fadal Mag vmc 6535 HTX is 2008

OK alittle more info for you all
The first picture in original post (wall)
this part was run in Fadal Mag vmc 6535 HTX
the second picture was run in Fadal vmc6030
So what is happening is not isolated to 1 machine
but to the Fadal machines. What we've been doing
is rgh and semi in Fadals and finishing in Mazak.
As far as plowing the part -- I don't believe this is the
problem -- rgh part to .100 stock --semi to within .030 stock --finish
And when we run dulicate parts the gouge is in the exact same spot
on the duplicate part. (Back to how Fadals are reading the data)
I know you all probably think I'm nuts,but again keep the ideas coming
I'm all ears.
As far as programing problem -- Why not get the same gouge when ran in
the Mazak -- unless it's the post processor??
This problem started about the time we went to Unigraphics NX2

We'll check the motor coupler and backlash but ??????

"I just assume your running a g9" -- no

We havn't tried to run the duplicate parts in 2 seperate machines
to see if gouge is duplicated or not -- may give this a try

Again any ideas are welcome

Thanks

[carbidecraters]

I would also cut and paste and draw the program in the machine control and see if there are some strange Y moves or some strange Z moves.

[FUN4ME]

I see you are not runing a G9, our machines default to G9.
try adding a G8 to the program and see if that helps.
the G8 can be on its own line I usually place it after the H call out line.

[Delw]

Originally Posted by FUN4ME I see you are not runing a G9, our machines default to G9. try adding a G8 to the program and see if that helps. the G8 can be on its own line I usually place it after the H call out line.

Yup I screwed up I should have said g8 not g9.




N1 O1111 ( 675in2vice4cav 01/29/09 )
N2 M6 T1 (0.375 DIA. END MILL)
N3 G0 G17 G40 G80 G90 E11
N4 M3 M8 S8500
N5 X-9.1875 Y-4.5
N6 G8
N7 Z0.1 H1
blah blah blah


it happens only on 2 sperate fadals so its definatly not a machine problem. its a code problem thats pretty much a givin. since you threw in a switch in programs thats would I would be looking at.

carbidecraters I dont think he could paste the code here and if he did most people wouldnt be able to understand it due to surfacing it would be hard to read every line and find the problem cause I am betting the program is probally 5+ megs long. it looks like his step over is about .020-.050 using a 1/4-3/16 ball endmill.


I'm not familiar with unigraphics nor its post processors, or the meshing controls in it, however I assume your using polylines and converting them to g code in the post proccesor, I would look for that section in the post proccessor.

[HuFlungDung]

If it were a difference in post processing, what could it be? Is machine comp being thrown into the mix on a random basis? For some who program using wear comp instead of radius comp, an amount of .010 might be expected.

Perhaps some small arcs are being misinterpreted (R value mixup), although those parts look like mostly interpolated straights, one would have to look at the code to see.

I could also run the code through OneCNC, since it has its own simulation module

[Rumm64]

We'll be running test parts over the weekend shift.
And will be trying diffent methods as suggested on here.
Will know more on Monday -- Thanks to all for your imput.

Again all suggestions will be considered --keep them coming.

I'll keep an eye on here from home
(Incase you may have additional ?'s or suggestions)

Have a good weekend everyone