I've trained many people with different technical background. A very typical thought from those who has worked as cnc or edm operator is the work piece should be placed “square” to cmm axis otherwise you won’t get good results, which is understandable because that’s how they do the set up on cnc or edm. That we all know is not really true on cmm because it’s the software that takes care of aligning the work piece not the set up. But this is a good habit for cmm inspection though because lining parts up “square” to cmm axis will minimize probe “shanking” especially for complicated castings or machined parts which requires all kind of different probe angles.

 But some other ideas about alignment are very wrong and “dangerous” to cmm inspection. One of them is “I just want a quick and dirty measurement so I don’t need DCC alignment.” I’ve even seen a company where none of the technicians do DCC alignment in their inspection or programming because they think that “saves” time. Maybe it’s not common anymore but for me it’s a good opportunity to discuss a little bitter deep on the importance of DCC alignment.

 DCC stands for Direct Computer Control so this topic doesn’t apply on manual cmms. When we check parts on cmm we all start off with creating a manual alignment (there’re cases where manual alignment might not be necessary and we’ll talk about it at the end) and then followed by a DCC alignment to “fine tune” the part alignment. It’s obvious that for parts with curves, 3D surfaces, casting drafts you have to have a DCC alignment to ensure correct measurement, but how about simple parts only require typical plane-line-line alignment? 

I think for most of you the alignment will be as easy as 1-2-3: 4-point-plane for datum –A- and 2-point-line for datum –B- and another line to intersect with –B- to create –C- and then to “save” time can we just skip the DCC alignment? Probably you say “no”, then “why”? Because it’s “better” or won’t “hurt” to have a DCC alignment. That’s true it’s always “better” to have one but for me it’s “imperative” that DCC alignment should be run before performing any measurement. And here’s why:

1, Probe compensation is not 100% accurate in manual probing. Even you’re a super video game player you won’t be able to probe the points perfectly square to part surface, which affects the probe compensation calculated in software. Only when the probe is driven by computer with specific IJK vector the probe radius is correctly compensated. The larger the probe you use, the bigger error this could cause.

 2, Datum targets won’t be repeated on another piece or same piece but measured again. Can you probe exactly at the same spot if I take this part away and then put it back on your cmm? On the drawing planes, lines are perfectly flat and straight but not in the real world. So depending on the real feature form error, how many points taken and where you probe those points, your alignment will be slightly different and that difference will mostly affect your measurement results. Even on the same part, if you probe the front –B- edge differently you will definitely get different readings for profile R30 then depending on the tolerance you could end up getting a “different” part. Same applies on hole positions.

 3, Easy verification when measurement is in doubt. Let’s say you come back from a coffee break and get some weird readings and you want to make sure the part is not moved so just simple run your DCC alignment. That really “saves” time.

 4, Absolutely saving time when running a batch of parts. This is so easy to understand but what most people don’t understand is why they should do DCC alignment if only 1 piece is required to check. Because this piece or more pieces could come back for check any time and other people might challenge your report any time and you want to be the “good” cmm guy.

 Now let’s think about checking flatness only (sounds even the manual alignment is not necessary) on a plate, do we need a DCC alignment? For all the above reasons, the answer is “Yes”. Because flatness is determined by every individual point and every point reading is affected by their XYZIJK which is controlled by your part alignment. I stated in last issue in my article about scanning a flat plate that a manual alignment is not even necessary but now you know at least a plane alignment is needed to control the depth of scanning so it will repeat on other parts.