Iterative alignment(help)

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Iterative alignment(help)

    Pc-dims 3.7mr2 cad++
    I have tried to do an Iterative alignment , using the points on the attached print, with no luck. I get an error every time.
    Any help would be great.
    Thanks and thanks agian
    Last edited by Clay_Man; 04-17-2007, 03:56 PM.

  • #2
    Well, if you are only using the 3 on the print, you will not have any luck. You need 3 points with vectors that are very similar for the level, 2 with vectors that are similar for the rotate that are as close to 90 degrees off from the first 3, then one more point that has a vector 90 degrees to all the previous points. Three points alone will not give you an iterative alignment unless they are true 3-D points, meaning spheres or holes (with a minimum of 3 surface sample hits). Just using 3 vector points will not do it as there is nothing to controll the lateral movement of the points. Meaning, if they have a 0,0,1 vector, there is nothing to keep them from being off in X or Y and there is nothing for Pcdmis to grab in those 2 axis to locate the part.
    sigpic
    Originally posted by AndersI
    I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

    Comment


    • #3
      What Matt said.
      When the going gets weird, the weird turn pro. Hunter S. Thompson

      Comment


      • #4
        Originally posted by Matthew D. Hoedeman
        Well, if you are only using the 3 on the print, you will not have any luck. You need 3 points with vectors that are very similar for the level, 2 with vectors that are similar for the rotate that are as close to 90 degrees off from the first 3, then one more point that has a vector 90 degrees to all the previous points. Three points alone will not give you an iterative alignment unless they are true 3-D points, meaning spheres or holes (with a minimum of 3 surface sample hits). Just using 3 vector points will not do it as there is nothing to controll the lateral movement of the points. Meaning, if they have a 0,0,1 vector, there is nothing to keep them from being off in X or Y and there is nothing for Pcdmis to grab in those 2 axis to locate the part.
        I am also using E, F and G

        Comment


        • #5
          Well, the vector for E and F are 180 opposite to each other and are NOT going to work for the rotation of the part. They give you 'stops' in the same axis, but are in-line to the other axis (they are in line in X axis, but have to be used to clock the part in the Y axis) and the G 'stops' the part in the Y axis. So, the E & F are 90 degrees off for what you need to locate the part. A fixture could not even be made to hold that part, it can and will twist about the 'face' of the part using E, F, & G, which is what is causing the error you are seeing.
          sigpic
          Originally posted by AndersI
          I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

          Comment


          • #6
            you probably have a vector wrong. would need to see program.

            Comment


            • #7
              Mat is correct the datum scheme is pretty bad. Your going to have to do somthing a little different then what is shown on the sketch. (nothing new to me).

              Comment


              • #8
                I don't know that you need this but I thought it couldn't hurt. This is originally fromm \v/inston...


                COLLECTING THERORETICAL AND MEASURED DATA

                1. From the “Auto Features” toolbar, select the “Vector Point” icon.

                2. On the Vector Point screen, be sure to have the “Misc.” selected to “Nominals”. Using the mouse, select a point from the CAD model on the top surface. Make sure you have “Surface mode” selected. Pc-dmis will pierce the CAD model and insert the values for this measurement.

                3. Once the point has been selected, verify all of the calculated values, paying close attention to the Normal Vector. This vector for the top surface should be logical in relation to the CAD coordinate system.

                4. Name the feature PT1. With the “Measure” box unchecked select “Create”. The CMM will now create the theoretical code for that point.

                5. Repeat steps 2-4 to create the remaining points as follows:

                ∙ PT2 and PT3 on the top surface. (A total of 3 points)
                ∙ PT4 and PT5 on a surface that’s ~75° to 90° from the top.
                ∙ PT6 on a surface ~75° to 90° to PT4, PT5 and the top.

                6. Once the 6 points have been created theoretically, its time to collect the measured data for these points.

                7. Mark All.

                8. Say Yes to “Ok to mark manual alignment features”.

                9. Execute the program. Follow the instructions to measure all 6 points. Once the execution is complete, the measured data has been collected in order to proceed to the Iterative Alignment.

                CREATING THE MANUAL ITERATIVE ALIGNMENT

                1. From the “Insert” menu select Alignment\New. From the alignment screen select “Iterative”.

                2. Select the points from this screen as follows:

                PT1, PT2 and PT3 as the Level (set drop down menu to default).
                PT4 and PT5 as Rotate.
                PT6 as the Origin.

                (Click “Select” after each, Level, Rotate, and Origin)

                3. Select “Meas All Once”. Verify that the “Point Target Radius” is set to approximately 5mm, (this means you should to take your hits within 10mm diameter of where you clicked on the CAD). Set your “Fixture Tolerance” to ~0.05mm (this is basically the best fit tolerance for the iterative alignment, if you get an iterative alignment error then try bumping the fixture tolerance up in 0.01mm increments). Select OK. Set the jog box to run SLOWLY. Say YES to the message “Measure all iter features now”. The CMM will prompt you to position the probe above the first point. Keep in mind that the machine is still in manual mode. Continue to follow the prompts through the rest of the points.

                4. Once the last feature has been measured, the Alignment screen will appear. Select OK. Watchingthe probereadoutwindow, notice that the origin is off the table (Car Position). Also notice that the axes are following the CAD model, not the machine axes.


                DCC ALIGNMENT:

                COLLECTING MEASURED DATA

                RULES:


                1. Insert an operator comment “Prepare for DCC Operation”. Then from the toolbar turn on the DCC mode.

                2. Due to the fact that the CAD origin is off the table, move points will be the most efficient way to establish clearance above and around the part for free state. If you are programming using a fixture then a good way to establish clearance is to add a “Move Distance” to your auto features, this can be found under the “Advanced” button in the “Auto Features” dialog box. This distance can be set to your liking 10mm to 500mm> depending on if you are trying to clear a clamp or whatever. When the CMM goes to measure one of the features it will start at the distance you set (say 500mm) above the feature, keep in mind that the probe will follow the approach vector and will start and finish at the distance you enter along the approach vector, so this is best used with an I,J,K in your machines Z direction.
                The use of clearance planes can also achieve the same result; moving from one point to another without hitting the piece being inspected.

                3. It is now time to collect the measured data for the DCC iterative alignment.

                4. Since you have the part locked-in relatively close with the manual iterative alignment you can now program the DCC alignment using auto features with the “Measure” box checked. Don’t forget your “Move Points” or “Move Distances” and make sure yourNormal Vectors” are in logical relation to the CAD coordinate system.

                *Before hitting “Create” set the jog box to slow to avoid a crash*

                The six point iterative is good for the manual alignment because it makes it faster to do your initial part set-up and it is not important to use datum features. All you are worried about, for the manual alignment is getting the CAD model locked-in relatively close so the DCC iterative can run through hitting the datums, precisely snapping the CAD to the part. The DCC Iterative Alignment is where the CAD model gets accurately locked-in to the part/fixture.

                If you programming a part in the free-state, then you will want to use your primary DATUM features for your alignment. Typically you will Take 3 points on datum –A-, measure a circle for datum –B- and a circle or a slot for datum –C-, typical but not always the case (important* if you have to use a slot then you will want to construct a point or circle at the slot to use in the iterative).

                If your are programming using a fixture then you will want to use the circle method iterative, for both the manual and DCC alignments. You will most likely use either three circles (cylindrical inserts), three spheres (tooling balls) or two circles and a point (two cylindrical inserts). You will want to enter the XYZ locations of the tooling balls/cylinder inserts in the auto features box for auto circle/auto sphere location. This information will be stamped on the fixture.

                CREATING THE DCC ITERATIVE ALIGNMENT

                1. After you have your measured data for the DCC iterative alignment, follow the steps above from CREATING THE MANUAL ITERATIVE ALIGNMENT the only difference is you will check the “Measure All Always” box for the DCC alignment and you may be using features other than just points.

                2. Once the last feature has been measured, pc-dmis will evaluate the measured points and decide if it needs to re-measure the points and “Iterate” the alignment. Most likely it will and will run through the alignment a second time.



                sigpic

                James Mannes

                Comment


                • #9
                  Originally posted by Matthew D. Hoedeman
                  Well, the vector for E and F are 180 opposite to each other and are NOT going to work for the rotation of the part. They give you 'stops' in the same axis, but are in-line to the other axis (they are in line in X axis, but have to be used to clock the part in the Y axis) and the G 'stops' the part in the Y axis. So, the E & F are 90 degrees off for what you need to locate the part. A fixture could not even be made to hold that part, it can and will twist about the 'face' of the part using E, F, & G, which is what is causing the error you are seeing.
                  Thanks
                  So using A1 thru A3 + E, F, and G how would you set your alignment??

                  Comment


                  • #10
                    Well, the 3 "A" point are fine, the E & F points need to be 90 degrees from their current target positions, and pointing roughly in the same direction. Then, since THEY will be 'clocking' in the 'Y' axis, the G point would also need to be rotated 90 degrees to get the last axis stop. Basically, in A-B-C terms, they gave you 3 A datums (level) 1 B datum (rotate) and 2 C datums (origin) when you really need 3A, 2B, 1C. 3 minimum for level, 2 minimum for rotate, 1 minimum for origin.
                    sigpic
                    Originally posted by AndersI
                    I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

                    Comment


                    • #11
                      My opinion:
                      Take three A points, two on E, one on G if most of your features are toleranced to A-E-G
                      Or three points on A, two on F and one on G if most of your features are toleranced to A-F-G

                      Comment


                      • #12
                        Those little tabs sticking down are WAY TOO small to use 1 of them to rotate the entire part to, no matter which one you use. The machine variation itself will make the part rotate at least 0.010" at the end away from the little tab used for the alignment.
                        sigpic
                        Originally posted by AndersI
                        I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

                        Comment


                        • #13
                          I agree the tabs ARE probably too small to rotate to
                          Question: What are the features toleranced to?

                          Comment


                          • #14
                            Originally posted by RussL
                            I agree the tabs ARE probably too small to rotate to
                            Question: What are the features toleranced to?
                            The bottom tab is tol. back to E and F, and trim surface is back to E,F, and G

                            Comment

                            Related Topics

                            Collapse

                            Working...
                            X