hexagon logo

Questions about DCC accuracy after initial manual alignment setup.

Hello everyone!

So, I've been lurking for some time on these forums, and finally feel I am experienced enough to post. I'm relatively new to programming and using our CMM in day-to-day use, but I've been familiar with them since about 2012. I work in aerospace, and work is getting to the point where I need to have more output from our CMM (I work in a small shop and am the only QC person here). I've been told by QC guys that I know from a few local businesses that it is NOT acceptable to use DCC to inspect parts after a Manual alignment is ran. Basically, I have a small fixture where the X and Y of the part does not move much if at all. I run the manual alignment on the first part for the day/ if I am just setting this part number up, let the program run the DCC and then the rest of the program. After I get the results, I record/save the .PDF file for our records, and then I clear execution list, slap another part on the table and bolt it down, and run the program from where the DCC starts. Essentially, the QC guys I know say that the repeatability on this is atrocious, and not good calibration practices. I've been doing some snooping around on this forum and it seems that quite a few experienced users here think the opposite: that as long as the fixture is good and your probes are calibrated after every change that things should be smooth sailing. I keep my touch and move speeds the same as cal too. I'm super new to all of this, and I'm not *exactly* sure about all the values either, and even though I've looked through this forum, don't really know the difference between LF/SF/MF/EF type probes or the differences between touch speed and all that. I try to read up on as much of this stuff as I can, but changing some of those small parameters doesn't seem to change much...but then again I'm a rookie here, so I probably wouldn't know the difference. If anyone could point me towards a guide where it might help me understand the small nuances it would be greatly appreciated. This is especially important because we have many parts that have Total Positioning tolerances of as tight as .010, so repeatability is my biggest concern. (well, that and getting off as many parts as humanely possible from our CMM)

Also, I looked around for an answer for this next question...but how do I change parameters for to automatically calibrate a single tip in the program? I finally figured out how to change probe tips in a program, the user has to manually change it (we don't have an auto changer), and then the program can calibrate the newly inserted tip. It took me FOREVER to figure out that I had to draw the sphere because info was kinda scarce/confusing. When I insert the Calibrate Single tip command, it works great..but how do I change the settings for the auto-cal? Changing any values in the very first probe loadout in the program does nothing to change the values for the auto-cal. The auto-cal always takes 13 hits from my sphere, and then stops without going up which I find odd. I just add in +6 inches for an incremental move after that just to be on the safe side.

Thank you for any information that y'all can provide. This place is awesome, thanks for having me!

​​
  • , this is what my measure probe dialog box looks like...??? What am I doing wrong? Pc-Dmis is fairly user friendly, but some things I'm really clueless as to where to go...


    You change your settings to match what you want, then in the program code, add in the Auto Cal code and put the newly named parameter set in and it measures the part to that set. So you can select different probe angles for a specific set, if need be.




    Code Example:


  • OMG...I cannot +1 +1 +1 this post enough!!!! I'm still not quite that familiar or comfortable with command mode...I have a simple custom hotkey made so that I can toggle between Summary and command. A QC dude that I know locally comes over and he always has to change it back to command mode for him. I'm just not "there" yet/haven't toyed around with the command mode enough to know these tricks. THANK YOU THANK YOU THANK YOU!!!
  • Okay, that makes sense...you want to keep your speeds the same as cal and every other OP to maintain accuracy, gotcha. The retract and prehit are to eliminate false positives? You're saying on the second DCC alignment, the distance is closer to the surface of the part than the first DCC? thank you again! This is going to be so helpful for me in the future! ^_^ Everything I have read here in the forum has allowed me to cut down on time without cutting corners.
  • , I love where you are going with this! The buddy of mine over at a neighboring aerospace shop uses readpoints somewhat frequently in his day-to-day, but I have yet to get that far. I've been writing my newer programs with the following steps: A full manual alignment, then coming back with an initial rough DCC alignment that mirrors the hits from the manual align (doing the typical plane, line, point if possible on the part), then coming back in with a DCC alignment that takes 10-20 hits on the 3 setup surfaces, thus creating 3 planes that are more accurate that I use for taking my measurements off of. I've been doing experimenting with it, taking 1 part that I know is good, and doing a lot of runs on it. It has made me way more confident about the repeatability of not only my machine, but the process as well. For a general rule of thumb, I'm probably going to try and show repeatability every time I set up a part for mass-volume checking parts through our CMM...just for my peace of mind? The thing that really makes me worried (besides being new to all of this), is that a lot of the sign-offs and quality control stuff will come back on me if there is anything wrong with the product down the line. I really want to be able to sleep better if you know what I mean. Some of the tolerances for even a simple rivet assembly on these parts is .002" tolerances from a rivet hole to rivet hole, on top of the rivet assembly having total positioning on the middle hole. I guess I'm just not there mentally to be able to say "hey, this data is great"