Scanning Outline

Hello,
Can anyone help me understand why after I preform a scan my scanning outline has a couple of spots where the line deviates to the inside or the outside and is not consistent all the way around the part? I'm attaching a picture below with the suspect area highlighted.

Thanks for you help!

Attached Files

DAN_M over 7 years ago

1) You could be "stock on" in those areas. The part might actually be bad.

2) You could have some FOD on your part or stylus tip causing a false reading.

3) This could be related to your alignment. Might be getting false readings if all 6DOF aren't adequately constrained. Can you post your code? I'd love to take a look.

I suggest you completely clean & deburr your part and then clean your tips. Then re-qualify your tips and then re-run the program.

If you can (I can't see your part so i don't know if this would work or not), have an inspection overlay made and measure the form on a comparitor to try and verify this.
Cancel
Vote Up 0 Vote Down

Sign in to reply

Cancel
bigtallanddopey over 7 years ago

What value do you have the find noms tolerance at? Normally when I have seen things like this it is a find noms issue, either too small and not taking the deviation into account, or too large and the software is looking at another surface. IT should be half the probe diameter plus whatever deviations from nominal you would expect on that part.

Another thing to check could be the CAD vectors, make sure the vectors of the surface are all correct.
Cancel
Vote Up 0 Vote Down

Sign in to reply

Cancel
louisd over 7 years ago

It looks like all three spots in question are 'blend surfaces' between tools or whatever is creating the entire profile shape.
Is this part CNC machined? Is it cut in-house? If so, show this graphic to whomever cut the part... They can use it to adjust their tool offsets, and fix the errors on the part.
If it's extruded or molded, I'd be looking closely at the dies.

Another thing to note, your displayed variation is exaggerated by default on a dimension, by a factor of 10x so the visible representation on the graphic likely isn't to scale. Say your spec is profile to .005", this is showing less than .0025" variation at the areas you boxed, nothing you can catch well with your naked eye.
Cancel
Vote Up 0 Vote Down

Sign in to reply

Cancel
UKCMM over 7 years ago

Beside the points already raised you may need to reduce the scan speed as your highlighted areas are all at change in scan direction and to high a speed can cause this kind of error
Cancel
Vote Up 0 Vote Down

Sign in to reply

Cancel
RandomJerk over 7 years ago in reply to bigtallanddopey

I'm on board with this explanation - if the part deviates too much, and it's a thin wall, your nominal is being granted from the back side of the wall.
Cancel
Vote Up 0 Vote Down

Sign in to reply

Cancel
AndersI over 7 years ago in reply to bigtallanddopey

This is one thing I don't understand - the back side has opposed direction, why should FindNoms match that surface at all?
Cancel
Vote Up 0 Vote Down

Sign in to reply

Cancel
DJAMS over 7 years ago

On the right side, there are suspicious -looking joggles in the tolerance bands. Was the initial scan performed in MASTER mode? If so, your nominals are not from CAD
Cancel
Vote Up 0 Vote Down

Sign in to reply

Cancel
dhoughton over 7 years ago in reply to bigtallanddopey

This helped a lot. Lowering the find noms and selecting best fit brought these areas together.

Thanks for your help!
Cancel
Vote Up 0 Vote Down

Sign in to reply

Cancel
dhoughton over 7 years ago in reply to DJAMS

When you say MASTER mode are you referring to the Nominals Method? If so then no, I'm in the FINDNOMS mode.
Cancel
Vote Up 0 Vote Down

Sign in to reply

Cancel
DJAMS over 7 years ago in reply to DJAMS

Exactly what I was referring to.
Cancel
Vote Up 0 Vote Down

Sign in to reply

Cancel