Reported position of cylinders

1) Yes, if you are able to physically measure the holes as a cylinder then you should.

2) With the Geometric Tolerance command (and previously XactMeasure), the wort end of the cylinder axis is reported for position.  Legacy dimensions give you the option of start, end, axis average or worst end.  However, since position applies to the entire length of the feature, you should always report the worst end.

1) Yes, if you are able to physically measure the holes as a cylinder then you should.

2) With the Geometric Tolerance command (and previously XactMeasure), the wort end of the cylinder axis is reported for position.  Legacy dimensions give you the option of start, end, axis average or worst end.  However, since position applies to the entire length of the feature, you should always report the worst end.

Thank you Neil.  Helpful as always.

Follow up question.  If the holes are threaded, can that potentially throw off the results?  I'm dealing with that myself and the starting point for all the holes is reporting in position, but the end point is out on some of them.  But it's wildly inconsistent.  Some of the worst ends still look ok, and some are way off.  In some cases, the starting point is actually worse than the end point.  Could the threads be giving me mixed results?

 Neil Challinor 

When we measure holes on thin material, and we care about TP of holes. Which is the best algorithm for this : Least_SQR  Max_insp or Minimum Separation ..?

ESchertz said:

Follow up question.  If the holes are threaded, can that potentially throw off the results?  I'm dealing with that myself and the starting point for all the holes is reporting in position, but the end point is out on some of them.  But it's wildly inconsistent.  Some of the worst ends still look ok, and some are way off.  In some cases, the starting point is actually worse than the end point.  Could the threads be giving me mixed results?

Are you applying the pitch when measuring the cylinders?  Also, how are the threads produced?

RobOntario said:

When we measure holes on thin material, and we care about TP of holes. Which is the best algorithm for this : Least_SQR  Max_insp or Minimum Separation ..?

If you are using Geometric tolerance commands for the positions then it doesn't matter how you measure the features because they get re-calculated.  You can choose either LSQ (least squares) or DEFAULT - default means that the math will be in accordance with the GDT standard you have chosen (ASME or ISO).  For example, DEFAULT datum math for ASME Y14.5 is constrained L2 whereas DEFAULT datum math for ISO 1101 is constrained min/max.  Ultimately, it is up to you to decide which math is most suitable for your application, based on the manufacturing method, types of features, accuracy of your machine and type of sensor.  This is discussed on the  following page of the help file which recommends users perform their own studies in order to determine which math type is most suitable for them (see extract below)...

PC-DMIS Help Center - 2023.1 (hexagonmi.com)

Frequently, you may need to use different algorithms during verification than the specification uses. Because of this, geometric tolerance commands give you math options that allow you to control what algorithms are used for the verification. You may find it difficult to choose the best math options. The only way to really be certain you’ve chosen the best math options is to do a careful study.

Recommended Steps for a Careful Study

1. Take several actual parts that represent the variety of errors the manufacturing process can produce.

2. Densely measure all the parts with lots of cross sections, and use equipment that offers much less measurement uncertainty than the form error.

3. Choose math types that closely approximate the specification.

4. Measure the same parts in the way you expect to actually measure your parts in production. Use the same sensors and measurement strategies you plan to use.

5. Choose a wide variety of math types, and compare how well those math types approximate your dense-and-high-accuracy measurements. That lets you pick the math type combination that most closely approximates the specification.

Typically, the best math option depends on the ratio between your measurement uncertainty and your form error. If the measurement uncertainty is much larger than the form error, then you can’t measure the real form error with your sensor anyway, and it’s best to choose something simple like plain least squares for the datum and feature math types. On the other hand, if the measurement uncertainty is much less than the form error, then it’s best to choose math types that closely approximate the specification.

Neil Challinor said:

Are you applying the pitch when measuring the cylinders?  Also, how are the threads produced?

I'm not sure I know how to do that to be honest.

So, if it was a M8x1.25-6H thread, the pitch would be 1.25.

Finding and Verifying Pitch Diameter - PC-DMIS for CMMs - PC-DMIS - Nexus Community (hexagon.com)

Thank you.  I don't remember this being covered in PC DMIS 101, if it was.

If it's in English, then do I put in the threads per inch instead?  Like 16 if I had 3/8-16-2B threads?

1 divided by threads per inch. So for your example 1/16 =.0625 

If your taking positional measurements from threads, I would suggest using a TP simulator as positions from threads can be very inaccurate in my experience.