CMM Inspection Approach: Teach CMM geometry by manual probing or program from nominal

[schnuer]

The issue has often come up where I work about the best way to perform inspections on our parts. Without getting to technical there are two opposing views: either
1.) write a complete and thorough program (in "edit mode") based on the nominal geometry of a part
-this takes (at best 30%) more time than option #2
-programming time can be minimized by piecing together segments of previously written code
-logic contained in the permanent record is easy to follow and review in case of measurement error/verification purpose
-allows for program to be easily reused, documented, debugged and modified in the future

or
2.) teach the CMM the geometry ("in learn mode") by manually probing on all the required features and report out the callouts requested by Engineering
-alleviates the need for rigidly defining the print geometry before inspecting the part
-probings are not vectored properly, ie. normal to the surface and chance of measurement error and/or inconsistency is much higher
-the sequence of commands used in the measurement session can only be saved as an informal record because each probing point is not individually defined, ie. the commands are executed by the operator
-if someone questions the measurement or the technique you don't have a definite record of how things were done, and you could end up remeasuring a different way each time:

Fortunately, I've built up my reputation enough to have a little sway in this, but pressure to squeeze the timelines is always a factor.

 

[True Position]

Re: CMM Inspection approach: teach CMM geometry by manual probing or program from nom

Is your machine DCC? Are people manually taking points? I know with Calypso even if you're manually touching a part to get a starting alignment for a pre-created program it go's back and measures the alignment features again using the programmed points/settings. Your vector question leads me to believe your machine is DCC, if you want, do a comparison of results with both methods on a simple part, two diameters a set distance apart, two tooling balls perhaps? Shouldn't take more then an hour or two, if the results are very different you have your evidence.

 

[granpagus]

Re: CMM Inspection approach: teach CMM geometry by manual probing or program from nom

We use method 1 for all programs that generate data that is retained or used in reports, are used in checking finished product, or otherwise considered important.
They do go through a (rudimentary) validation process, are fully documented and both hard copy and digital copy are archived.

Method 2 is used for 1 of's, quick checks etc. They are not validated, documented or archived and are deleted after 30 days. They do seem to proliferate rapidly.

And method 2 programs do not migrate to method 1. Too many headaches.

 

[mboucher]

Re: CMM Inspection Approach: Teach CMM geometry by manual probing or program from nom

Take the time to write a program. If you have a DCC CMM you will want to do this so that the features are measured with the same probe force, same vector direction, etc... each time. Make sure your drive vectors are correct and this will elminate cosine probing error.

It is okay to have a manual origin setup but make sure to repeat it in DCC mode so that the origin and the remaining features are measured with the same parameters.

There is some info on this at the magazine site, I publish CMM Quarterly, and I encourage you to view the materials there.

Mark

 

[John Nabors - 2009]

Re: CMM Inspection Approach: Teach CMM geometry by manual probing or program from nom

There is a third option that I like best assuming you have this option (I do with Calypso): extract all of your features from a solid model. The majority of our parts (aerospace/defense stuff) are made from either Catia V or Pro-E solid models. I can generate a program from a solid model in at least half the time it takes to generate one by 'teaching'. Also, with Calypso, 'line by line' programming is not an option.

 

[mboucher]

Re: CMM Inspection Approach: Teach CMM geometry by manual probing or program from nom

. Also, with Calypso, 'line by line' programming is not an option.

I not sure what your saying here. I use Calypso and I am able to program from and model and also with a model using the joysticke to create a program line by line.

Mark Boucher

 

[John Nabors - 2009]

Re: CMM Inspection Approach: Teach CMM geometry by manual probing or program from nom

I not sure what your saying here. I use Calypso and I am able to program from and model and also with a model using the joysticke to create a program line by line.

Mark Boucher
CMM Quarterly

With other CMM software I have used in the past you could simply type in a program without ever touching the joysticks. That is what I mean by 'line by line'.

 

[mboucher]

Re: CMM Inspection Approach: Teach CMM geometry by manual probing or program from nom

Ok, that makes sense. Yeah, no access to the code is rough sometimes when you could just pop in and change something really quick.

Mark Boucher

 

[True Position]

Re: CMM Inspection Approach: Teach CMM geometry by manual probing or program from nom

But why would you ever want to? How often would manually editing a program be faster then just changing the feature/char. inside Calypso without the risk of editing something wrong and doing something terrible.

I gave up trying to be smart and manually editing programs when I scrapped a part out on an EDM I was running doing that.

 

[granpagus]

Re: CMM Inspection Approach: Teach CMM geometry by manual probing or program from nom

I do a lot of offline programming without CAD (PCDMIS). CAD is faster.
I program at night. Minimally trained personell run the programs during the day.
User interface, program sequence and documentation must be complete and consistant.
It's been a while since they called me at home.