hexagon logo
Register Sign in
  • Replies 8 replies
  • Subscribers 31 subscribers
  • Views 102 views
  • Users 0 members are here
Options
Related

Profile of a cone

xxKJxx
Greetings everyone....my name is Kevin, but I go by KJ. I'm hoping someone might be able to assist me with a problem I've been having regarding the ROMER arm I've recently started using. The main issue I have is when I measure a cone on a part and then try to dimension it's profile my results are extremely skewed. Since I work at a company that prevents me from disclosing any particulars on what we manufacture, I am limited in what I can show/link to these forums.

With that said, I feel a lot folk deal with cones on parts so I think I'm okay so far. Anyways, I've worked with PC-DMIS for awhile (learned the DOS version when I first started using a CMM) and I've been to the Level 2 class at Hexagon for training. There are things I still need to learn I'm certain, but for right now I'd settle on being able to figure out why my results always end up inaccurate.

So far, I have tried using an auto-feature to construct the cone I want to profile and then measuring the cone. After I try to dimension the cone (datums are defined), my results are either 0.000 or show up as being several inches out of tolerance. While I have never been afraid of zero, I doubt the parts I'm checking (forged) are that smooth. For example my tolerance FCF might say | Profile symbol | .010 | A | B | the measured value shows 2.3119 or something similia. Way outside of tolerance.

When I run the program offline my results always come in nominal so I don't think I'm programming it incorrectly. Also when ran on a CMM, the results are more believable (if not in tolerance, than at least not inches out of tolerance).

My ultimate question, since I guess i haven't asked one yet, is this - Does the ROMER need special instructions for PC-DMIS to accurately report the profile of a cone?
  • Do you have a CAD model imported into PC-DMIS? If so, the best thing is to just collect a number of vector points on the cone surface, combine to a SET and dimension profile on that (or look at individual T values). If you don't have a CAD model, how should PC-DMIS know what the THEOs for each point should be? And what should it do when you don't measure at that exact point? Maybe autotrigg can help placing the points more accurately, but being manual you will always measure somewhere else than the THEOs say...
  • Just like Andersl says, align to your datum scheme and then take auto-vector points..report the set as well as individual T values. You'll need a model to do this.

    The only way to do this without a model would be for you to program it by manually probing your datum surfaces, keying in the correct nominal XYZIJK & size values for them, and then having your eng dept give you the XYZIJK nominals for target points for you to probe in that cone. The reporting would be the same, tolerance the set as well as T values. BUT you're using a romer arm so you wouldn't be able to do this either =/
  • Do as AndersI said, take individual points. There is some thing messed up with the auto cone. A couple of years ago I had an AE confirm this and he said he submitted it to be fixed, but I do not know if it ever got rectified.

    https://www.pcdmisforum.com/forum/pc-dmis-enterprise-metrology-software/pc-dmis-portable/390053-auto-cone-extracted-from-a-cop
  • Without a CAD model, individual points are useless as this is a ROMER arm, and I'm not even sure an auto feature cone would work. You're never going to touch off at precisely the point it's looking for.

    I haven't used an arm in over a decade, but from what little I remember, I don't think you can do profile without a CAD model with any manual machine.
  • I am using a model for my program so I have that going for me. I have tried to use vector points and constructed a set from those points to give me the profile but as was mentioned I don't ever get the correct location when trying to repeat the measurements.

    Is autotrigg where it keeps making high pitched sounds as I get closer to the theo of the vector points? I used that once but couldn't remember what it was called not how to enable it again.

    One the things I run in too when using vector points is a consistant lack of repeatability. I haven't looked at T values so I'm going to do that today.
  • Can you do a distance or time based scan? Do that with find nominal turned on.
  • I am using a model for my program so I have that going for me. I have tried to use vector points and constructed a set from those points to give me the profile but as was mentioned I don't ever get the correct location when trying to repeat the measurements.

    Is autotrigg where it keeps making high pitched sounds as I get closer to the theo of the vector points? I used that once but couldn't remember what it was called not how to enable it again.

    One the things I run in too when using vector points is a consistant lack of repeatability. I haven't looked at T values so I'm going to do that today.

    Can't help you there, We primarily use our arm with the laser scanner. We seldom do any real hard probing.
  • Can you do a distance or time based scan? Do that with find nominal turned on.

    I've attempted to use scans but I've failed so far to get them to function for me. The options on not the same on the ROMER as they are for the CMM. I just lack the knowledge to use them correctly I guess.

    I did however, use T values on the auto vector points I created to figure out which points where causing the Set Scans to be out of tolerance. By going back to re-measure them using the point auto trigger feature and then re-dimensioning the scan, was able to get accurate results. Verified using CMM, so thank you all for the assistance. I really appreciate you all taking the time to help me out.
© 2024 Hexagon AB and/or its subsidiaries. Privacy Policy | Cloud Services Agreement