Tilted Work Plane Issues

[5thnewb]

New to 5th axis work in general. We have a machine (Matsuura MX-330) that was bought specifically for one job to be ran lights out, with prep work being done during the day. We want to be able to run eight of these parts unattended. This machine has an A and a C axis. Job was programmed using Mastercam and Camplete. We are trying to use the tilted work plane function (G68.2) to mill the majority of the part at A0, then rotate to A-90, and mill different features at C0, 90, 180, and 270. We are using a probe macro to pick up the X,Y, and Z on every part. 

At A0. we are milling the first side and outside profile mostly complete, including our Z zero for all programming and our primary Datum. X0Y0 is the Center of the part. When we were just running one part proving out the program and process, I thought I noticed a little variation here and there, but mostly chalked it up to not having the most rigid setup, and working out speeds and feeds. When we got to the point of our first test run of what would happen overnight we realized we had a problem. When we rotate to A-90, and mill features that relate to the face of the part(Y0 at A-90, Z0 at A0) our dimensions are getting smaller through the eight part run. For example, we mill a feature in the X axis parallel to the face of the part. the dimension from this feature to the face of the part shrinks approximately .0004 per part. We have a few other ways to verify this happening, but we dont know why. It seems like we are incrementally losing some of the positioning from the G68.2 Function. Here is a sample of the beginning of a tool at A-90. 

T14M06

M08

G49G00G53Z0.0

G131P2(HIGH SPEED CONTOURING ON)

M132(A+C UNLOCK)

G00G90G54A-90.C90.S16500M3

G54.4P1

G68.2X0.Y0.Z0.I90.J-90.K0.

G53.1X-2.9Y1.05

M131(LOCK)

G43Z4.H#517

Does this look ok? Any ideas as to what might be going on? forgive me if this is not clear enough. I can add any more necessary details if needed. Thanks!

 

[JParis]

If it's good in Camplete, it'll be good on the machine

[5thnewb]

Yeah, thats what I hear. And this may be a machine problem. Wondering if a parameter is possibly off. I also may not have explained the issue well enough. At A0. the part is getting thinner, as if the Z axis is growing from heat. If we were not probing the next part, this would make sense. When it swings to A-90 our established face is off the amount of growth we have in the spindle, which would in turn throw our Y off after rotation. But we are probing for a new Z each part, so any difference should be picked up and adjusted each time. The way it is now we lose about .0004 each part, like clockwork, so that the dimension on the 8th part is about .003 different from the first. we have ran this cycle through backwards, meaning pallet 8 to 1. We have tried a half hour dwell between cycles, and each time we can compare the first part ran of the eight, to the first of the next cycle, or the third to the third and so on, and they are almost identical. 

[Leon82]

Usually when we probe z for every part the tracks relatively well. Have a probe the part right before you cut the feature. You can't probe it and then a half hour later after running blast coolant expected to be thermally stable when you cut that feature.

 

Are you using a lot of through spindle coolant? That will heat the enclosure 10 to 20° above room temperature at prolonged periods of time.

 

[cncappsjames]

1 hour ago, 5thnewb said:

...

G53.1X-2.9Y1.05

...

Does this look ok?

 

This is a problem. Not sure if it thee problem, G53.1 needs to be on a line all by itself.

 

[cncappsjames]

Just now, Leon82 said:

Are you using a lot of through spindle coolant? That will heat the enclosure 10 to 20° above room temperature at prolonged periods of time.

Right now we're in the heat of summer here in Southern California. No pun intended. Non climate controlled shops in my area are starting off around 75 first thing in the morning and running 100+ before noon and maintaining that well into the late evening depending on building sun exposure.

Even with no high pressure coolant, I was in a Matsuura MAM72-35V yesterday. It was easily 20 degrees hotter in the machine AND humidity was of course through the roof. That's definitely gonna have a drastic effect on things.

Aluminum's thermal expansion coefficient is roughly .0001" per inch, per degree.  Do the math. Steels aren't nearly as sensitve, but a 35 deg swing will have an effect. Probing can solve a variety of issues if implemented properly.

[5thnewb]

17 minutes ago, cncappsjames said:

This is a problem. Not sure if it thee problem, G53.1 needs to be on a line all by itself.

 

Oops, I fat fingured there. The G53.1 is indeed on its own line.

[cncappsjames]

Your code format is correct.

Have you checked your #19700 parameters with the sphere and either Matsuura eZ-5, a 3rd Party Axis Check program, or Renishaw Axi-Set?

[crazy^millman]

18 minutes ago, cncappsjames said:

Your code format is correct.

Have you checked your #19700 parameters with the sphere and either Matsuura eZ-5, a 3rd Party Axis Check program, or Renishaw Axi-Set?

I believe you found that issue at one shop years ago where that was off and fixed their similar problems right up.