Reporting surface profile check with CMM

[prdave00]

How are surface profile measurements typically reported? The metrology lab always reports the absolute value of 2x the maximum deviation from the basic profile. In other words if the equally disposed profile tolerance zone is .010", the calculated value would be compared against .010+0" and .010-.010". They indicated this is preferential for most of their customers as opposed to reporting a positive or minus deviation from the basic profile. Can anybody comment? Is this typical?

 

[dgallup]

That's normal for a symmetrical tolerance. It's the same as a true position tolerance zone where you calculate the deviation from the basic dimensions and multiply by 2.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[prdave00]

Thanks dgallup. That was the second part of the explanation from the lab. I was just surprised because I'd want to know which direction the maximum deviation was in. I would think a shop would want this kind of feedback to assess whether there's a mean shift in the data.

 

[pmarc]

Reporting 2X maximum deviation from true profile is a very typical way to go (just like for position tolerance), but sometimes not useful.

If for instance complex outline of a seal is inspected, such information can be meaningless because it would be much more important to see/know which 'areas' of the outline are close to profile tolerance zone boundaries, which are at + and which are at - relative to true profile, which are out of the tolerance zone, etc. Having maximum deviation only would tell almost nothing in this case.

On the other hand simple YES/NO assessment might be enough, if one is only interested in checking whether the actual surface is within profile tolerance zone or not.

 

[dgallup]

Metrology labs typically just report if the part is good or bad. If you want process control or want to understand a process that's a whole different story.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[ptruitt]

I have been getting inspection information in a graphical format. The inspection service puts the point cloud into "Smart Profile" (a software application) and emails the results for me to view in the Smart Profile Viewer. The viewer has limitations. But I do like being able to see (visually, see) how well the part complies with the profile requirements. I am still trying to decide how best to use the information to ascertain process capability in a meaningful way, though. I wonder what others might suggest with respect to process capability.

Peter Truitt
Minnesota

 

[prdave00]

I recently came encountered an "issue" with a supplier. A 3D surfaced feature with a profile tolerance is being checked using one of 2 CMMs depending on what's available. I should note that the entire sample population from each batch is checked using the same CMM (i.e., both CMMs cannot be used to inspect the same feature). One CMM reports the "deviation" as 2x the absolute max deviation from the surface and compares it against the total profile tolerance. The other CMM reports the "deviation" as either the maximum deviation if all points in the cloud fall on the same side of the surface or, if some points in the cloud are on the inside and some on the outside of the basic surface, it reports the additive as the deviation (e.g., one point is -.001, the other is +.002, then the deviation is .003). For the latter cases, the deviation is compared to half the total profile tolerance. I've become used to the former method of comparing 2x the max absolute to the total tolerance. My biggest issue is that, regardless of the methodology, they are transferring the "deviation" to their inspection form and therefor, without having the printout available, I can't blindly compare the results of multiple lots to calculate capability indices, etc. since the deviations may be calculated differently.

The CMM operator maintains that the latter method is correct based on ASME Y14.5.1M-1994 and comparing twice the absolute deviation to the total tolerance is therefor incorrect. I've read the applicable section 6.5 of the standard, but all it does is make my head hurt. Anyone better versed in the mathematical definition standard have an opinion?

 

[DeanD3W]

ASME Y14.45 will be a new standard that will define content options for measurement data to address GD&T specs. We're hoping to get it released within about a year (maybe a bit more). When it's released I hope some of the issues like the one prdave describes will get fixed.

Yes, 2X the absolute value of the maximum surface deviation is a common profile value for an equal bilateral profile. The surface deviations for each measured point are commonly negative if less material is present compared to the true profile, or positive if more material is present.

The difference between the most positive and most negative surface deviation value is not a correct way to calculate a profile value. The point of the 2* |max SurfDev| method is to find the tolerance zone thickness that will just contain the worst measured point, while keeping that tolerance zone centered about the feature's true profile. That other approach the cmm operator prdave describes yields a number of a zone that fails to stay centered on the true profile.

For Y14.45 we're working towards one method that also will work for unequal profile tolerances as well as for equal bilateral. The method also needs to work for unconstrained profile tolerances, which have no datum feature references.

Both Evan and Don Day are part of the Y14.45 effort. Either of them may have more to say about this. Whatever we have at this point may be revised before the standard is released, so it won't be much help to describe what we have so far... I just want to mention that a method that I hope will clarify profile data reporting should be available in an ASME standard fairly soon.

Dean

http://www.d3w-engineering.com