Designing a Test Part to Assess CMM Operator's Knowledge

B

BBill

Hi everybody,

This is my first post but I have been reading many of the threads that pertain to the ISO 17025 standard over the last few days. However, I haven’t been able to find an answer to my questions. So, here is the deal.

My boss has asked me to design a small test part to be used for testing operator proficiency at using CMMs. The idea is to include specific features into the test part which would expose operator’s basic lack of knowledge.

That way, just by looking at the measured data, one would be able to assess whether the measurement errors (if there were any) were the operator’s fault of due to an un-calibrated machine. So my question is; What kind of features could be incorporated into the part which would test the operator’s proficiency and would allow the person assessing the operator to identify what the operator did wrong just by looking at the raw data.

Any help or ideas will be greatly appreciated on this one!

Thanks in advance.

Bill
 
T

True Position

Typically any measurement made with a CMMs main source of error will be within the machine. If the programmer can write the program in the first place is the issue.

Good tests are true positions, secondary datum structures, rotated alignments, alignment between two pins/holes with neither being the origin and them not lining up on either the X or Y axis, etc.

Getting different results for roundness and cylindricity is a good test for a Calypso programmer as well.
 
J

John Nabors - 2009

I can't remember working on a CMM that did not come with a sample piece as a training tool. If you do not have one perhaps you could check witth the manufaturer of your CMM(s) to see if they can provide one.

Hawat - speaking of Calypso, how do you like it? Zeiss was just getting ready to transition from U-MESS to Calypso the last time I worked on one.

-John
 
T

True Position

I absolutely love Calypso, it is the best CMM software I have ever had the pleasure of using. It's powerful, intuitive, and allows very easy program modification after initial programming.
 

gard2372

Quite Involved in Discussions
Hi everybody,

This is my first post but I have been reading many of the threads that pertain to the ISO 17025 standard over the last few days. However, I haven’t been able to find an answer to my questions. So, here is the deal.

My boss has asked me to design a small test part to be used for testing operator proficiency at using CMMs. The idea is to include specific features into the test part which would expose operator’s basic lack of knowledge.

That way, just by looking at the measured data, one would be able to assess whether the measurement errors (if there were any) were the operator’s fault of due to an un-calibrated machine. So my question is; What kind of features could be incorporated into the part which would test the operator’s proficiency and would allow the person assessing the operator to identify what the operator did wrong just by looking at the raw data.

Any help or ideas will be greatly appreciated on this one!

Thanks in advance.

Bill

Bill,

Your statements concern me a bit. The general tone of your Boss' direction seems to be directed at operator faults, lack of knowledge and what the oeprator did wrong. Am I totally off base here in reading the tone of your post? :(

The focus IMHO should be on a sound CMM training program. Every operator/programmer has to start somewhere and some pick up to code and "The Right Hand Rule" faster than others. :read:
 
T

True Position

It sounds more like he's looking for a method to validate the effectiveness of his training methods.
 

Kingsld1

Involved In Discussions
I would not be worried so much about the software. If I were to be testing a programmer I would be focused more on basic blueprint reading, GD&T and probably their surface plate skills.

Basic CMM courses generally take a week out of your life and are almost exclusively concerned with teaching the software. They also assume that you know your way around a print and could, if needed, do a setup the old fashioned way.

I have seen people "graduate" from a CMM course with little clue as to how a print is telling them the part should be setup.

I have worked with five different CMM languages. The general concepts are the same no matter what. With very little instruction you can figure out where they hide the command at this time. Learning how to read a print is the crucial difference between a good CMM programmer and a button pusher.

Another thing to remember is that there are no hard and fast rules about the number of points to take for a feature. You have to have some prior knowledge about how true a feature type typically is for a particular process. You also have to be aware of your tolerances and adjust the number of points accordingly.
 
T

True Position

I will agree with the previous person. The most important part of using a CMM is knowing how to do the same measurement on a surface plate, and knowing exactly what the print is telling you.

While my software background isn't as extensive as Kingsld1's, he is basically right, probing a circle is pretty much the same all around.

As to how many points? I've found a thousand per 25 mm of circumfrance per helix and 3 rotations minimum is generally quite accurate for bores.
 
J

John Nabors - 2009

I would not be worried so much about the software. If I were to be testing a programmer I would be focused more on basic blueprint reading, GD&T and probably their surface plate skills.

Basic CMM courses generally take a week out of your life and are almost exclusively concerned with teaching the software. They also assume that you know your way around a print and could, if needed, do a setup the old fashioned way.

I have seen people "graduate" from a CMM course with little clue as to how a print is telling them the part should be setup.

I have worked with five different CMM languages. The general concepts are the same no matter what. With very little instruction you can figure out where they hide the command at this time. Learning how to read a print is the crucial difference between a good CMM programmer and a button pusher.

Another thing to remember is that there are no hard and fast rules about the number of points to take for a feature. You have to have some prior knowledge about how true a feature type typically is for a particular process. You also have to be aware of your tolerances and adjust the number of points accordingly.

I completely agree with you. Back in my machinist days I saw that a person who had extensive experience with manual machines and a little bit of training on a CNC could easily show better quality and overall productivity than a person with extensive G-code training who was clueless about how to make a chip.

But then I am biased - I cut my teeth on a Archer knee mill and a Hardinge toolroom lathe.
 
Top Bottom