CAD++ Question

  • Filter
  • Time
  • Show
Clear All
new posts

  • CAD++ Question

    pc-dmis cad++ 3.7 mr1

    We're cutting several parts from 3D data on several different machine tools to determine which machine cuts closest to CAD data. The cavity pocket has various 3D contours, sharp external corners, and smooth internal fillets designed to test the high-speed machining capabilities of a new machine. I can visually see some surface irregularities in some of the parts. I have two questions, if you could enlighten me.

    1. What is the best way to inspect the entire part with minimal programming efforts? I was trying to do a perimeter scan but can't get it to work. Then I was going to do a patch scan. There's only 30 or 40 surfaces in the model. Is there an easy way to inspect each surface and output the points and deviation from CAD?

    2. What is the best way to compare the inspection results from one part to another? I was thinking of using points to show the deviation from CAD nominal. If I inspect all 3 parts, then adjust the tolerance so that most points are within tolerance on 1 part, then I could compare the results quickly by seeing the number of red out-of-tolerance points on the other parts. Maybe there's a more precise method?

    If anyone could help me with either question, I'd appreciate it.


  • #2
    Datapage is a good tool to use to compare your data.

    It would be eaiser to determine what approach if we could see a pic or print downlown.

    Also what type of head do you have on your cmm, that a big issue to on what or what not so scan and or take points. What is the size of the part?

    But I would probly go with 4 points per surface. And enough edge point like maybe 3 points per staight edge and in every radius hit the hi and lows
    sigpicSummer Time. Gotta Love it!


    • #3
      I'm new to the forum and not very experienced in the software. I'll try to figure out how to get a pic uploaded.

      The part is about 4"x6"x3"deep. I was hoping to patch scan + perimeter scan just to get a bunch of points quickly. Maybe picking several points on each surface wouldn't be too bad in program mode. I'll try.

      I have Datapage but have never used it. Is it easy to use?


      • #4
        Less is more.

        In the forum to upload a pic you must click on the Go Adnanced, then scroll down and click on Manage Attachmenta, then click on browse, and search for for in, that easy.

        As the the Less is more thing, you do not need a ton of points to compare different cut parts, its not like you are checking if for Final Inspection. you just want to compare cuts in steel. And if you or them think you need more than add them. Its always better than over kill
        sigpicSummer Time. Gotta Love it!


        • #5
          See attached picture of part

          looking for 'quick' way to inspect to

          1) compare to CAD nominal

          2) compare with multiple parts made on different machines
          Attached Files


          • #6
            I know this is not really an option but, if you could get it laser scanned that would be the best way of getting the most info about the entire part. The next best thing to do would be to just take point on it and dimension out the "T' values. Or you could do linear open scans along the y axis and then some along the x axis. and then ask for the "profile" of those scans.

            James Mannes


            • #7
              Seem like you should be able to do some U & V scans from the surface model
              there are some very good people on this forum that should be able to help with the scanning end of this
              I do not have a lot of experience in the scanning feild just small amounts
              But if i rember you just pick the surface and tell it how many points in U and how many in V and a distance or somthing like that


              • #8
                thanks, i'll try that


                • #9
                  It is kind of picky when it comes to UV scans. Usually you can only pick surface and do that one. Sometimes, depending on how smooth the surfaces, you can pick more than 1 and geta good scan out of it.
                  Again, I think linear open scans would serve the purpose.

                  James Mannes


                  • #10
                    would he then use the linear open as masters and check the other cuts back to the 1st one he created or can they be done in point mode so to speak
                    so it will compare aech point to Math data
                    Like I say I am no expert at scanning so anything I can pick up I appreciate


                    • #11
                      Wow! That's the first time I've tried the UV scan. That seems to give me close to what I want. I don't mind picking one surface at a time.

                      I did have to change the start and end to generate the points where I want them.

                      What is the difference between the surface and vector hit type? looks like they do the same thing.

                      Based on what I want to accomplish, should I check the Single Point to output points or can I dimension the scan to show deviations from CAD for each point measured?

                      Thanks for your quick help! That's amazing...


                      • #12
                        Profile is the way you want to go(at least that is what I say, others may say to send out the point info. See this thread concerning profile...


                        James Mannes


                        • #13
                          I think you need to use the single point method for what you are doing
                          Glad the UV thing worked out


                          • #14
                            If what you are looking for is which machine did the best cut, why do you need all them points? One comparing that many point is overwhelming. I thought all you want is to know which one was the best. And a couple of hits on a surface will tell you that. You should not need to take 100 hits to compare one part from another. Matt please give me your opinion on this, and let me know which way you would do this. If I am wrong I need to know.
                            sigpicSummer Time. Gotta Love it!


                            • #15
                              I kind of understand what you're saying about fewer points. I am using the UV scan in single point mode to select automatically a 3x3 grid on each surface. Longer surfaces I'm increasing to 6x8 or so to space out the points sort of evenly across the part but at least 4 points on each surface, too.

                              The point method is my best guess at comparing one part to another. I don't know how to dimension scans or profiles. I'll check out your link, though. I think if I show the deviation only on all the points and adjust the tolerance so that most are 'green', then I can inspect the other parts and see how many are red compared to the 'best' part. When I finish gathering the points, I'll read the posts on Datapage and see if I can use that to compare the parts.

                              I am keeping an eye on these posts, so if there's a better way, please let me know. I appreciate all the input so far. This forum is great!


                              Related Topics