calculating flatness

[[ch...]]

I created a plane using 16 points, my actual flatness of the plane constructed by 16 points is different from maximum and minimum value difference of the individual points x value. I use the plane for spatial rotation, I made one point x axis zero. No planer rotation , no z,y origine.This is my base alignment. My production people says that the flatness and points maximum and minimum value difference should be same. What is the science behind calculating flatness in calypso???

IMG_20200110_123014.jpg

[[Cl...]]

I think that depends on how you are evaluating the
plane? Some of the more knowledgeable guys in
here should be able to clarify that.

[[De...]]

Flatness is zone tolerance and therefore is the difference between the highest point and the lowest point on the plane.

Your Lowest measured point is -0.0617.
Your Highest measured point is 0.3348.

If you add these together you get 0.3965.

Your result for flatness is 0.3973.

It appears your results match with less than a micron of difference between the two.

That difference can be accounted for based on Calypso calculating flatness at the edges of the plane and not only where your points are taken.

[[Me...]]

Flatness uses a best fit of the points. Reporting linear X values does not unless you create a BF alignment that mimics the fit used for flatness.

[[ch...]]

Please sign in to view this quote.

My lowest point is point 9 it is -0.14, higher point is point 12 0.33
So flatness should be 0.47!!!!!

[[ch...]]

Please sign in to view this quote.

Ok, I will try in best fit...

[[De...]]

Please sign in to view this quote.

Sorry, I did not see that point when first looked at it.

You said you use the plane as your spatial in your base alignment with only a point for your origin, but what algorithm are you using in your base alignment?

When checking flatness it should be using minimum feature (which finds the smallest zone that all measured points will fit within), I would suggest since you are attempting to emulate the flatness check, that you should try minimum feature in your base alignment so your base alignment is aligning the same way flatness does.

If it is using Outer Tangential or Least Squares then it will actually align differently depending on the measured points than the flatness check will.

[[Er...]]

What does the base alignment algorithm has to do with flatness? 😱
The flatness is the same regardless of alignment.

Edit:
Reading the OPs post again, I think I understand you Derek.

Yes, your "x-values" in most likely not measure in the same direction as the planes vector, thats why your result looks strange.

[[Br...]]

The tolerance zone has nothing to do with the alignment. There are a variety of different algorithms that can be chosen to fit the data in the most optimal manner. the Minimum zone method for instance is supposed to be the most optimal but is susceptible to outliers so you need to account for that. LSQ will be more like an average best fit that wont be affected by outliers as much. Trying to recreate those algorithms with your single points without using PCM will be really difficult. using the base alignment would not even be close depending on how it is put together.

[[An...]]

Please sign in to view this quote.

For sure.
The delta of min.- and max. value will always be larger than the value of flatness.
(in 99.99% of the cases)
Flatness always uses the "Minimum Algorithm".

[[Br...]]

Another thing, Is the same feature being used to evaluate the flatess that those points came from? you would have to create a plane, and recall all those points into the plane, then evaluate that for flatness if you want an apples to apples comparison.

[[De...]]

Please sign in to view this quote.

In the first post in the thread OP stated that they aligned to the plane being measured by using it as their spatial alignment in their base alignment. They then placed points aligned to the base alignment and checked the X values of the points to verify the results they were getting from flatness. They wanted to know why the X values did not match the results of the flatness characteristic.

It is my understanding that the X values for any feature will reflect the X value in the base alignment unless they are assigned to some other alignment. That said, it seemed to me, that if you wanted the base alignment to align to the face the same way flatness did, that you would want to choose Minimum Feature as that is what the flatness characteristic uses. If Minimum feature is used and the outliers and filters are set the same way in the spatial alignment of the base alignment and the flatness characteristic I would expect then the X values reported would fall within the same zone as the flatness characteristic.

There is a possibility of course that a peak or valley that was not checked on the plane was checked with one of the points and would cause some difference in the results. In general I still think the concept would be ok to simply have a second method to verify the flatness results.

[[ch...]]

Please sign in to view this quote.

Actually it is a big die casting job, it is not a machined face.As ERIC told the vector direction of individual points might not be same as the vector direction of the plane constructed using those points.
I was ordered to check the flatness of the surface in addition to that variation of points with respect to one reference point, for understanding where is minimum and maximum. I can't do scanning, there is lots of ejector pin projection in the surface.

According to advise from this forum I think I should use minimum feature option and calculate the flatness using best fit alignment. I know flatness do nothing with alignment,but the X values definitely do change with respect to alignment.l got the point now, the vector direction of individual points and the resultant plane will not be same. Finally it become a problem in vector algebra.

Thanks for all the ideas....

[[Br...]]

Please sign in to view this quote.

Ok yeah i agree with that. I sort of breezed over the OP I guess.