Hello Guest it is March 28, 2024, 08:26:40 AM

Author Topic: A very weird prob... pls assist...  (Read 6496 times)

0 Members and 1 Guest are viewing this topic.

A very weird prob... pls assist...
« on: June 04, 2010, 07:49:55 PM »
Hello Guys:

Here is a very weird problem i m facing with my mill: (4 axis - G540 -ballscrews - SmoothStepper)

After tramming the mill to a acceptable degree,
And measuring with the dial and physically verifying the distance on the DRO is the actual distance traveled,
I made a small cube (10x10x6mm)  in my CAD prog to find the 0,0,0 of the machine. ( I do roughing and finishing on top/bottom)
After cutting  - measuring - adjusting the offset i got a pretty good accuracy. (about 0.05mm tolerance).

Here is the weird part:
when i cut my test part i always get the same result, but when i cut other parts they about 1mm smaller in the Z axis. (ie: my part should be 6.1mm in Z and it comes out 5mm)
The problem occurs no matter:
 - If I use SS or not,
 - Two different computers,
 - I run simulations to exclude any CAM issues.
 - While using the LTP driver and press the verify button   I don't have lost steps in the report.

In the bottom line:
 When i cut some rectangular parts I cannot recreate the error, When i do pocketing or anything else I end with physical geometry which is missing about 1mm in Z.

I would highly appreciate any ideas on this very weird problem.

Offline Hood

*
  •  25,835 25,835
  • Carnoustie, Scotland
    • View Profile
Re: A very weird prob... pls assist...
« Reply #1 on: June 04, 2010, 08:04:10 PM »
What does your code look like, is there any height offsets or work offsets being called?
Hood
Re: A very weird prob... pls assist...
« Reply #2 on: June 04, 2010, 10:19:00 PM »
No i don't call any height or work offsets in the G-code.

I simply set all axis to zero at the point where all axis meet, raise Z, add my stock and run the code

Thanks Hood.

Offline Graham Waterworth

*
  • *
  •  2,668 2,668
  • Yorkshire Dales, England
    • View Profile
Re: A very weird prob... pls assist...
« Reply #3 on: June 05, 2010, 05:31:45 AM »
Is the value always the same?

If you run again without resetting datums using fresh stock is the error doubled?

Graham
Without engineers the world stops
Re: A very weird prob... pls assist...
« Reply #4 on: June 05, 2010, 10:26:08 AM »
Hello Graham,

yes the error is duplicated, this points to a possible CAM issue but i run the simulations and I dont see a prob there.
Still i should try a different CAM to exclude this.

Thank you

Offline Jeff_Birt

*
  •  1,107 1,107
    • View Profile
    • Soigeneris
Re: A very weird prob... pls assist...
« Reply #5 on: June 05, 2010, 12:56:52 PM »
Hi Yianni,

My first thought when I got your note was tool offsets being applied as Graham and Hood have eluded too. I have been contemplating possible tests to verify machine operation. One easy way to test the machine operation apart from any generated GCode is to use the MDI line. With the Z-axis zeroed with the tool touching the top of the stock, moving the Z to -2mm and then cutting a slot should show a 2mm deep slot. Another options is to use one of the built in Wizards in mach to do some simple operation.

In your CAD/CAM program you could make two concentric pockets the outer 2mm deep and the inner 4mm deep. If the outer pocket is shallow and the inner is 2mm deeper than the outer you know your CAM is putting in a tool offset.
Happy machining , Jeff Birt
 
Re: A very weird prob... pls assist...
« Reply #6 on: June 05, 2010, 01:51:59 PM »
Hello Jeff,

I did something similar with MDI and could not find any errors other than tiny misalignment due to physical tramming.

Next step is to swap Z with A axis (cables and pins) and see whats going to happen I will know more on Monday once I try it.

Thank you.

Offline Graham Waterworth

*
  • *
  •  2,668 2,668
  • Yorkshire Dales, England
    • View Profile
Re: A very weird prob... pls assist...
« Reply #7 on: June 05, 2010, 01:59:27 PM »
You say the error is duplicated but is the error twice what it was with a single run?

Did the tool feed into the top of the second part at depth plus the previous error?

What I want to know is if the error is incrementing or not.

Graham
« Last Edit: June 05, 2010, 02:00:59 PM by Graham Waterworth »
Without engineers the world stops
Re: A very weird prob... pls assist...
« Reply #8 on: June 05, 2010, 02:23:21 PM »
Hello Graham,

Allow me to apologize as English is not my native language.

When i say the error is duplicated I mean i put in new stock, run same code and end up with same results like the first cut.

I did  not try to run the code twice on the same stock.

I m not sure I understand what you mean by "if the error is incrementing or not"

Thank you

Offline Graham Waterworth

*
  • *
  •  2,668 2,668
  • Yorkshire Dales, England
    • View Profile
Re: A very weird prob... pls assist...
« Reply #9 on: June 05, 2010, 04:27:50 PM »
on the first run lets say that the error is 0.5mm too deep, on the second run with no resetting of the datum is the error in depth 0.5mm or 1.0mm

Graham
Without engineers the world stops