GEOTOL help

When you are using legacy, the features (both the circles and planes) will be calculated using whatever algorithm is specified in the feature itself. By default this is Least Square, which is an averaging method.

The legacy position call out is using these calculated features (centres, vectors and sizes) in its analysis.

If you're looking at the measured values of the features themselves and seeing correlation with the position calculations this is why, because it's looking at the same values.

When you are using GeoTol, the previously calculated features don't actually come into play. All the hits of both datum and considered features are recalculated.

The primary datum will be constructed external to material on the top plane.

The B datum will be constructed perpendicular to datum A then external to material on the B face.

Datum C will be constructed perpendicular to both A and B external to material on the C face.

The hole will be created external to material, much like Max Inscribed (but not exactly).

If you construct a primary datum plane from your A plane, a secondary datum line from your B plane and tertiary datum point from your C plane, then do a 3-2-1 alignment from those. Then change the calculation for for the circle to Max Inscribed and report using legacy you should get very similar numbers to the GeoTol output.

That said, it's likely the block isn't that far out, more likely that form of the circle or the datums (maybe a hit on a burr or bit of dirt) swaying the results.

If you've measured only 3 points on the datums you won't see form error, but if you check perpendicularity you will likely see them showing not square.

Alternatively, if the GeoTol diameter is showing small, it's likely this causing the issue.

When you are using legacy, the features (both the circles and planes) will be calculated using whatever algorithm is specified in the feature itself. By default this is Least Square, which is an averaging method.

The legacy position call out is using these calculated features (centres, vectors and sizes) in its analysis.

If you're looking at the measured values of the features themselves and seeing correlation with the position calculations this is why, because it's looking at the same values.

When you are using GeoTol, the previously calculated features don't actually come into play. All the hits of both datum and considered features are recalculated.

The primary datum will be constructed external to material on the top plane.

The B datum will be constructed perpendicular to datum A then external to material on the B face.

Datum C will be constructed perpendicular to both A and B external to material on the C face.

The hole will be created external to material, much like Max Inscribed (but not exactly).

If you construct a primary datum plane from your A plane, a secondary datum line from your B plane and tertiary datum point from your C plane, then do a 3-2-1 alignment from those. Then change the calculation for for the circle to Max Inscribed and report using legacy you should get very similar numbers to the GeoTol output.

That said, it's likely the block isn't that far out, more likely that form of the circle or the datums (maybe a hit on a burr or bit of dirt) swaying the results.

If you've measured only 3 points on the datums you won't see form error, but if you check perpendicularity you will likely see them showing not square.

Alternatively, if the GeoTol diameter is showing small, it's likely this causing the issue.

Thank you Badger. This helps a lot.

Yes I am looking at the same values. This is how machinists and management also is looking at it. They can take the block set it on the rock take a point with the height gauge on the top plane(datum a) and find the center of the hole. That is deviation 1. They then do that to the face(datum c) so you now have x and y. They plug that into the position formula and get a value. That value is very close to legacy but very different then GEOTOL. That value they trust. The CMM they do not trust. CMM should always be able to match other methods that have been used for 100 years if not why have a CMM. So now it has to be PCD or the CMM that is wrong.

how and why would both the feature and datum be recalculated? A point is a point.

So if a cad model is used for that top plane and its nomials are 0 but it measures .0006 positive in the Z does it calculate the datum as 0 or the measured value of .0006? If it is not using the perfect plane of zero or the cad model why? the CAD is the MBD and it what the contract/PO says is go to.

YOu lost me here. Why is it created externally?

I will try this. Should all circles( i use cylinders) by Max inscribed? I have always used them as LS

I do agree with this. I have tried many times to tell them our machines and parts are very dirty. We also have huge temperature swings all day. As far as points I have to take a very minimum amount of points. They do not want the CMM to be running parts long. If chips are not flying we are not making money. They do not accept long run times on parts.

If the diameter is small why would the center point change? the center is the center. Size won't change that. The form would change that.

I would say 3 hits is unacceptable for a datum, its fine for dcc alignment even manual alignment, POST THE CODE. PLEASE I'm sure that will explain the problem. Put it this way Space-X has actual requirements for the programs we make (amount of hits per size of feature and full alignment again after DCC alignment), I'm sure your company would be very upset how long it takes those programs to run haha.

mike my alignments are almost always my datums. Why would I want to probe the same thing twice to get an alignment and a datum from the same feature. Plus up until 2022 I could program with less points normally 4 for a plane and have it give me results I could trust and verify.