Slot minor (diameter) dimension too large?

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

  • Slot minor (diameter) dimension too large?

    I'm currently laying out a part with many relatively narrow slots (2.02 mm x 25.5 mm.) I need to measure the slot minor (diameter) and slot major dimensions (length) and also slot to slot dimensions. I'm using a 1mm cylindrical probe on a brown and sharpe global CMM with 2012 version of PC DMIS. Initially I created the slot feature using an auto feature, but when I reported the measured slot diameter it came out over tolerance, it deviated from nominal by about 0.2 mm, which is huge considering the tolerance is 0.05 mm. I checked the slots with some pin gages and they don't seem to be anywhere near that big. I then measured the width of the slot using two vector points, and it came out almost right on nominal, pretty well in line with what I was finding with the pin gages.

    It's technically a round slot, but due to the narrow width I'm using a square slot feature. I have the radius set so that it takes the hits along the long end of the slot well out of the rounded ends. I have tried it with the auto feature "measure width" option both selected and unselected and get the same too wide results. The slot major dimension and slot to slot seem to be coming out fine.

    Any thoughts on why my slot features are coming out with such a big "diameter"? I searched the forum but couldn't find anything that related to this issue. I know how I could get what I need just taking vector points, but I have 200 slots on this piece and would like to avoid having to keep track of 1000+ vector points in order to get all the dimensions I need. Maybe there is some other way to skin this cat that I'm not able to come up with??

  • #2
    Tried measuring the two slot radius using autocircle (one autocircle per radius)?
    PC-DMIS CAD++ 2o19 R1 SP11

    Comment


    • #3
      Really thin part ? Is that why you are using the cylinder probe ? Have you had any other size issues with this probe ?

      Comment


      • #4
        I found useing auto square slot with find center more repeatable for size & location. But I would do the auto circle to verify the radius. The round slot seems to be erratic for the radius.

        Sent from my ALCATEL ONE TOUCH Fierce using Tapatalk
        Horizontals - Dual Arm 40-17-21 / Single Arm 30-14-16 - PCD 2013MR1 (64bit)

        Comment


        • #5
          This is a stamped part. Material is 1.6 mm thick. This part is almost like a perforated corrugated plate, with the slots in the "valleys". It's a stamped part. I use the cylindrical probe so I don't have to fiddle too much with my program to get my probe at the exactly correct depth to take the hits. It also ensures I pick up the most narrow point through the depth of the slot. Sometimes these parts can have a bit of bow to them as well. I can't take sample hits because the slots are not in a flat surface.

          I don't think the issue could be with the probe since the same probe gets a fine measurement on the width when I determine it using vector points.

          I will try using the find center option and report back if that changes the reading. I wasn't using it since my slot to slot distance is pretty consistent, and I thought it would speed up the run time of my program if I didn't do a find center on each of the 200 slots.

          I assumed trying to measure the circle would be a less reliable way to determine slot width, but I can give it a shot. I'm not certain how true our radii are at the ends of the slot, the radius is just a reference dimension.

          Comment


          • #6
            Even with the non normal surface, I was taught to set depth to the center of metal to capture the snap of the punch. But I would be concerned with the cylindrical probe unless your perfectly parallel to the metal, which I rarely was. That helped me match my gage pin verification. But most of my slots were +/- 0.25 unless it was a datum, then it was +0.1/-0.0mm.

            Sent from my ALCATEL ONE TOUCH Fierce using Tapatalk
            Horizontals - Dual Arm 40-17-21 / Single Arm 30-14-16 - PCD 2013MR1 (64bit)

            Comment


            • #7
              If you are measuring the diameter of the slots using the radius on both ends, it is quite possible you will get a different result from measuring, for example, the distance between two lines in the same slot. This would be caused by imperfect form on the ends due to manufacturing practices. This is common with stamping, milling, etc. The solution is in the shop. I suspect they will try to behead you for reporting your findings, or at the very least shame you into admitting it's YOUR method that is all wrong. I will say that the cylinder probe is not the most precise, so they may have a point.

              Comment


              • #8
                You need to measure the shank diameter of your cylinder stylus, historically we have been using a 2mm cylinder and the shank is fractionally smaller than the ball diameter, if the shank is significantly smaller on your stylus then this will give you oversize readings.
                Last edited by AlexA; 04-25-2014, 09:45 AM. Reason: learnt to spell
                Mnfctr: Hexagon,
                Model: Global B Perf,
                Hardware: Control FBII,
                Software: PcDmis CAD 2010MR3
                PH10MQ, TP20, Leitz equipped

                Comment


                • #9
                  slot width.jpg

                  I've attempted to attach a screen shot of my code and a graphical view. I did this with a 1 mm spherical probe to address the concerns some of you had with the cylindrical probe, but the results are the same. To summarize:

                  1. I'm using a square slot auto feature.
                  a. Measure width is selected (although the results are the same when it is not)
                  b. Find center option is selected (although the results are the same when it is not)
                  c. Changing the corner radius input does not change the results.
                  2. I'm creating a line on each side of the slot.
                  a. the hits for the line are taken at the exact same depth as the slot.
                  3. The width of the slot per the Auto slot feature is 2.285 mm. Based on pin gage measurement this is way too big.
                  4. The width of the slot by taking the distance between the two lines is 2.044 mm. Based on pin gage measurement this appears to be the correct width.

                  Any explanation?? Or is PC DMIS just doing something that messes up the auto slot "D" measurement?

                  Comment


                  • #10
                    Sorry that picture came out terrible, here is the code:


                    MOVE/CLEARPLANE
                    SLTS1 =FEAT/CONTACT/SQUARE SLOT/DEFAULT,CARTESIAN,IN
                    THEO/<17.97,29.15,-8.8>,<0,0,1>,<0,1,0>,2.02,25.5
                    ACTL/<17.965,29.295,-8.774>,<0,0,1>,<-0.0006788,0.9999998,0>,2.28,25.548
                    TARG/<17.97,29.15,-8.8>,<0,0,1>,<0,1,0>
                    MEAS WIDTH=YES,RADIUS=0
                    SHOW FEATURE PARAMETERS=NO
                    SHOW CONTACT PARAMETERS=YES
                    DEPTH=0
                    SAMPLE HITS=0,SPACER=3.5
                    AVOIDANCE MOVE=NO,DISTANCE=0
                    FIND HOLE=CENTER,ONERROR=NO,READ POS=NO
                    SHOW HITS=NO

                    DIM LOC1= LOCATION OF SLOT SLTS1 UNITS=MM ,$
                    GRAPH=OFF TEXT=OFF MULT=10.00 OUTPUT=BOTH HALF ANGLE=NO
                    AX NOMINAL +TOL -TOL MEAS DEV OUTTOL
                    D 2.020 0.050 0.050
                    2.280 0.260 0.210 --->
                    END OF DIMENSION LOC1

                    MOVE/CLEARPLANE
                    LIN6 =FEAT/CONTACT/LINE/DEFAULT,CARTESIAN,BOUNDED
                    THEO/<16.96,19.1,-8.8>,<16.96,39.1,-8.8>,<0,1,0>,<1,0,0>,<0,0,1>,20
                    ACTL/<16.947,19.106,-8.8>,<16.944,39.09,-8.8>,<0,1,0>,<1,0.0001328,0>,<0,0,1>,19.984
                    TARG/<16.96,19.1,-8.8>,<16.96,39.1,-8.8>,<0,1,0>,<1,0,0>,<0,0,1>
                    SHOW FEATURE PARAMETERS=NO
                    SHOW CONTACT PARAMETERS=YES
                    NUMHITS=2,DEPTH=0
                    AVOIDANCE MOVE=NO,DISTANCE=0
                    SHOW HITS=NO

                    MOVE/CLEARPLANE
                    LIN4 =FEAT/CONTACT/LINE/DEFAULT,CARTESIAN,BOUNDED
                    THEO/<18.98,19.1,-8.8>,<18.98,39.1,-8.8>,<0,1,0>,<-1,0,0>,<0,0,1>,20
                    ACTL/<18.99,19.094,-8.8>,<18.989,39.093,-8.8>,<0,1,0>,<-1,-0.0000668,0>,<0,0,1>,19.999
                    TARG/<18.98,19.1,-8.8>,<18.98,39.1,-8.8>,<0,1,0>,<-1,0,0>,<0,0,1>
                    SHOW FEATURE PARAMETERS=NO
                    SHOW CONTACT PARAMETERS=YES
                    NUMHITS=2,DEPTH=0
                    AVOIDANCE MOVE=NO,DISTANCE=0
                    SHOW HITS=NO

                    DIM DIST1= 2D DISTANCE FROM LINE LIN6 TO LINE LIN4 (CENTER TO CENTER),NO_RADIUS UNITS=MM,$
                    GRAPH=OFF TEXT=OFF MULT=10.00 OUTPUT=BOTH
                    AX NOMINAL +TOL -TOL MEAS DEV OUTTOL
                    M 2.020 0.050 0.050
                    2.044 0.024 0.000 --#-
                    Last edited by olivia311; 04-26-2014, 03:04 PM.

                    Comment


                    • #11
                      Please see my previous response, and then re-read your #4. CMM = coordinate measuring machine. It is NOT a FMM, or form measuring machine. While one can dial the CMM in for form, it is, in my opinion, always a good idea to have secondary methods for form. Round slots are a prime example.

                      Comment


                      • #12
                        Originally posted by NUGUY View Post
                        Please see my previous response, and then re-read your #4. CMM = coordinate measuring machine. It is NOT a FMM, or form measuring machine. While one can dial the CMM in for form, it is, in my opinion, always a good idea to have secondary methods for form. Round slots are a prime example.
                        Your previous response addressed trying to measure the slot by measuring the radius at the ends. I'm not doing that. I'm using the square slot auto feature which doesn't measure the radius at the ends.

                        Comment


                        • #13
                          I know you were using the lines to re-verify the distance in question but I am assuming that the hits for the line and the hits for the slot are not in the same location. The auto features can be troublesome if one hit in the set is erroring for any reason. Especially if you incur an error in the first two hits as the demon adjusts the vector for the hits 3 & 4 to parallel a line created from hits 1 & 2. You can dimension the individual hits by typing in the SLTS1.HIT[1]. Repeat but replace the 1 with 2, 3... and so on. From there you can trouble shoot. I would trial different areas for hits 1-4 by changing the RADIUS value. Also from your previous comment, you stated you have moved the away from the radius but the code shows otherwise. Or did you change the locations manually?

                          I have the radius set so that it takes the hits along the long end of the slot well out of the rounded ends.
                          Code:
                          SLTS1 =FEAT/CONTACT/SQUARE SLOT/DEFAULT,CARTESIAN,IN
                           THEO/<17.97,29.15,-8.8>,<0,0,1>,<0,1,0>,2.02,25.5
                           ACTL/<17.965,29.295,-8.774>,<0,0,1>,<-0.0006788,0.9999998,0>,2.28,25.548
                           TARG/<17.97,29.15,-8.8>,<0,0,1>,<0,1,0>
                           MEAS WIDTH=YES,[COLOR="#FF0000"]RADIUS=0[/COLOR]
                           SHOW FEATURE PARAMETERS=NO
                           SHOW CONTACT PARAMETERS=YES
                           DEPTH=0
                           SAMPLE HITS=0,SPACER=3.5
                           AVOIDANCE MOVE=NO,DISTANCE=0
                           FIND HOLE=CENTER,ONERROR=NO,READ POS=NO
                           SHOW HITS=NO
                          Horizontals - Dual Arm 40-17-21 / Single Arm 30-14-16 - PCD 2013MR1 (64bit)

                          Comment


                          • #14
                            Problem Solved

                            I finally found the root cause of my issue a few weeks back and thought I would update for any out there who are struggling with a similar issue and stumble across this thread. It was a rookie mistake. My touch speed was too high. I was using the auto slot feature in DCC mode, but when I was taking the hits manually I was in manual mode. The hits I was taking manually I was taking slower and that's why those measurements were more accurate. These narrow slots are the only things I typically measure that have tolerances which are so tight, and so I did not notice the error on other parts I have measured with the same touch speed. Now that I've slowed down the touch speed I'm getting consistent results across all the different methods (auto slot, manual mode cmm measurement, pin gauges, cylindrical probe/spherical probe) they all agree. At least within enough of a range that the differences are insignificant with respect to my tolerance.

                            Out of curiosity, what touch speed do you all typically use?

                            Comment


                            • #15
                              I usually use 2 (percent of maximum on my CMM One) though there are times that I have to go down as slow as 0.5, particularly with features such as 2mm holes. Before including multiple hit speeds in a program I measure a gauge at both speeds several times to verify that variable hit speeds won't affect my results. With this particular machine and the fairly short probes I use I've not seen any deviation with touch speeds from 0.5 - 2.0. It is preferable to contact your part at the same speed you calibrate at.

                              The speeds I use are based on a day of testing various styluses at various speeds and various tip angles on a variety of gauges.
                              Last edited by Erikir; 07-15-2014, 05:12 PM.
                              2012 CAD++ MR1
                              Brown & Sharpe ONE
                              PH10T

                              Comment

                              Related Topics

                              Collapse

                              Working...
                              X