Announcement

Collapse
No announcement yet.

Cylindrical squares or blocks

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

  • #31
    Originally posted by oldtiffie
    This is getting tedious.
    Yes, it is Tiffie.

    Nope - wrong again. In both the case/s of Forrest and Harvey they both used known/proven squares, set their bases (in Forrest's case, the ball of the mast of his surface guage) to the base of the reference square and zeroed the dial indicator.
    No, that's completely wrong Tiffie.

    Jay has a wrist pin that he's trying to check for taper and lean. You can not prove the wrist pin by just checking the base. All that will do is show you the error at the base. Checking the base will not give you enough information to calculate the taper and lean of the cylinder.

    You must run a DTI along the length of the cylinder to determine the tilt and lean.

    That is wrong as well to the extent that any errors in "tram" or in the mechanical alignments and "fit/s" of and in the machine mechanisms - if not identified - will manifest themselves as "false positives/negatives" and show either as errors which do not exist in the item under test or are additive (+ve or -ve) and superimposed on the real state of the test piece.
    You still don't get it. Do you understand why you don't need a straight test bar to tram a mill head, or to do Rollie's Dad's Method? Any tram error is zero'ed out when you do the second measurement on the opposite side and add the two measurements together. That's the beauty of the approach!

    I suspect that your use of the example of Paul's was a "red herring" as it had nothing to do with Forrest or Harvey's methods. But having said that, I'd support the reliability of the accuracy of Paul's movement a V-way above just about any other on balance - for the reasons given.
    Sigh. You still don't get it. Paul's method is the same as the method I described -- instead of putting the cylinder you're trying to prove under the mill spindle and running the DTI along it on opposite sides, you can put the cylinder on a lathe faceplate and run the DTI along it on opposite side. It's the same test, just horizontal instead of vertical!

    Perhaps so - I plead guilty to that. But I must say that you are not innocent in that regard either - a case of "the pot calling the kettle black" perhaps?
    Paul, Forrest and I answered Jay's question on page 2. As usual, you didn't read the preceding posts, and instead jumped into the thread with a long tirade that did not show the OP how to test his wrist pins, and wandered about all over the place with unrelated links.

    Perhaps so, but TipTop is quite able to speak for himself without anyone else assuming his role and doing it for him - as it the case for anyone and everyone on the this forum.
    You notice, of course, that Jay thanked the original responders (myself, Forrest, and Paul), but was so turned off on the thread once you started posting that he hasn't responded since. Likewise, Forrest and Paul had better sense that to try to explain their responses to you...

    So let me try this one last way: Jay gives you one of his wrist pins to test. Show me how you're going to calculate a precise measurement of the taper and lean of his cylinder.

    Hint: you're going to need two measurements at the top and bottom of the opposite sides, so measuring the cylinder at the base isn't going to give you the answer.
    Second hint: the closed-form solution for the cylinder's taper and lean is (from my answer on page 2 ):

    Originally posted by lazlo
    Taper = (D1 + D2) / 2

    Lean = [(D2 - Taper) - (D3 - Taper)] / 2
    Last edited by lazlo; 06-17-2008, 10:55 PM.
    "Twenty years from now you will be more disappointed by the things that you didn't do than by the ones you did."

    Comment


    • #32
      Jay:

      Since now you're probably completely confused, here's an article by Lucas Precision, a Monarch, Lodge and Shipley, Natco/Carlson, and Oloffsson (high-end machine tools and CNC) distributor as well as metrology gear:

      Lucas Precision: Construction And Use Of A Cylindrical Square

      To properly check and use a Cylinder Square:

      With the square on the top of the machine table, put the spindle back gears in neutral and position the head and table so that an indicator held in the spindle contacts the side of the square near the bottom end of travel.

      Establish a 'zero' reading at the exact center of the square by moving the saddle back and forth to get the 'highest' reading, then set indicator 'zero'.

      Traverse the head up the column and note the error and weather it is 'plus' or 'minus'. Save this indicator reading and call it reading "A". Make a light pencil mark on the base of the square corresponding to the path that the indicator traveled.

      Without moving the cylinder square, re-position the indicator to read the exact opposite side of the square. Establish a 'zero' reading at the exact center of the square by moving the saddle back and forth to get the 'highest' reading, then setting indicator 'zero'.

      Traverse the head up the column and note the error and weather it is 'plus' or 'minus'. Save this reading and call it reading 'B'.

      Re-position the indicator to read the original side of the square. Rotate the square on the tabletop exactly 1/2 turn.

      Establish a 'zero' reading at the exact center of the square by moving the saddle back and forth to get the 'highest' reading, then setting indicator 'zero'.

      Traverse the head up the column and note the error and weather it is 'plus' or 'minus'. Save this reading and call it reading 'C'.

      Insert the 3 readings taken above into the following formula and solve for 'T'. Please note the order of the calculation as indicated by the brackets, and perform the addition in an algebraic manner (taking into account he 'sign' of the readings). This is the amount of taper built into the cylindrical square due to manufacturing error.

      T = (A + B) / 2

      Insert the original readings taken above and the calculated value of 'T' into the following formula and solve for 'L'.

      L = [(B-T) - (C-T)] / 2

      Please note the order of the calculations as indicated by the brackets, and perform the addition in an algebraic manner (taking into account the 'sign' of the readings). This is the amount of 'Lean' built into the cylindrical square due to manufacturing error.

      Insert the calculated value of 'T' and 'L' into the following formula and solve for 'L'.

      D = (T + L)

      Please note the order of the calculations as indicated by the brackets, and perform the addition in an algebraic manner (taking into account the 'sign' of the readings).
      Last edited by lazlo; 06-17-2008, 10:38 PM.
      "Twenty years from now you will be more disappointed by the things that you didn't do than by the ones you did."

      Comment


      • #33
        Let's carry on

        lazlo,
        Don't worry I am not getting anymore confused than I had made myself in trying to figure a way to Check these pins. Thats one of the nice things about being very ADHD, you can think about 10 problems and / or solutions at once. I see the difference between the two testing methods and I understand about your triangulation and trig of the differences in measurements. I would not trust my 1927 9" SB to have ways that would pass muster for this ordeal. I don't know if my Fray miller would either. So I believe that puts me in the bench test park with the other camp.
        oldtiffie,
        I have a surfac gage that was built by a machinist that obviously new what he was up to by the quality of his craftsmanship. At this time, I do not have an attachment for my Starret last word indicator or my regular style indicator. I know I need to make some. So I tried a V block type stand with my dial indicator to see what would happen. I clamped the V block with post and indicator to my surface plate. Then I set the cylinder in the V way and set the indicator (preloaded) at the top with the dial at "0". I then spun the cylinder around slowly and found that my indicator was showing .005 runout. Now this is showing a relationship between the base and the side of the cylinder at 5.25" off the surface plate and also the possibility of an egg shaped cylinder. So then I did the same thing with the cylinder sitting on its other end and had only .002 runout.
        Anyone,
        So correct me if my thinking is off here so far. I have read the bedside reader (in bed) about making cylindrical squares. When making one in just one set up, you releive yourself of errors. I thought I would cut a fat one and start with an existing product, but now I need to find where the error or errows are. Now I suppose I could set it up in my mill and do a a sweep around the top of my cylinder to center it and this would also tell me if it was in fact not cylindrical but egg shaped and that would at least answer one question. I do not believe they are egg shaped at the ends though just because of the interference fit in the pistons. I think my ends are off and possibly the middle could be (which I have not checked) where the rod runs. It may be harder to true a used wrist pin than making a new cylindrical square. But the saving grace is that I don't have to have this accomplished right away. I do own a Grenby cylindrical grinder, but it is awaiting my shop addittion. Hopefully that will take place yet this summer. Any more ideas are more than welcome, I might be flunking my self imposed test. Jay
        "Just build it and be done"

        Comment


        • #34
          Originally posted by tiptop
          lNow this is showing a relationship between the base and the side of the cylinder at 5.25" off the surface plate and also the possibility of an egg shaped cylinder. So then I did the same thing with the cylinder sitting on its other end and had only .002 runout.
          You're getting close to measuring the taper, but you don't want to move the cylinder between measurements at the base and the top, because the lean will distort the taper calculation. That's the big advantage of using a machine slide or quill.

          In other words, picture the Leaning Tower of Pisa: you can have a cylinder that's perfectly round on any horizontal slice (i.e., if you measure just at the base), but it tilts in some direction. Taper, of course, would be if the tower is perfectly straight, but smaller at the top than at the base. All cylinders will have a combination of the two errors.

          When making one in just one set up, you releive yourself of errors.
          You minimize the errors, but won't eliminate them. Even a cylinder turned in a 10EE won't be perfectly round/straight.

          By the way, that Lucas Precision link (above) also has a section on how to make a cylinder square. It's basically the same description as Guy Lautard's...

          I could set it up in my mill and do a a sweep around the top of my cylinder to center it and this would also tell me if it was in fact not cylindrical but egg shaped and that would at least answer one question.
          Notice I was describing how to measure the cylinder's taper and lean. Measuring roundness is much, much harder There's a whole chapter on measuring roundness in Moore's Foundations of Mechanical Accuracy. They start out the chapter by explaining that there's not even a metric or standard for roundness. Moore uses a polar plot showing deviation from perfect roundness, but there's no single "roundess" number.

          A rough way to check roundness is to rotate the cylinder in a V-Block. But Centerless Grinders usually leave odd numbers of "lobes" on the final product and if you have 5 or 7 lobes, the V-Block won't detect the error.

          This is why a plug gage won't fit into a ring gage of the corresponding size: plug gages are usually centerless ground, so they have 3 or 5 lobes, and ring gages are lapped, so they're very, very round. It's also why tool and diemakers will check a bore with both a 2-point and a 3-point bore gage -- they detect different numbers of lobes/out of roundnesses (is that a word? ).

          Moore shows their roundness checker, and there's a commercial product -- the Taylor Hobson "TalyRond", which is essentially the same thing:





          It's basically an exquisitely-precise rotary table, with a DTI on a mast that electronically raises up from the bottom to the top of the device under test. It outputs a three-dimensional roundness chart.

          For a home shop, you can do a similar measurement manually by adding more measurements to the method that Paul and I described: measure the error at several points spaced around the cylinder, from bottom to top. Map it all out and you'll have a three dimensional "roundness" chart.
          Last edited by lazlo; 06-17-2008, 11:33 PM.
          "Twenty years from now you will be more disappointed by the things that you didn't do than by the ones you did."

          Comment


          • #35
            Go for it.

            Thanks tiptop for the call-back to the OP and the wrist-pins (we call them gudgeon pins in OZ).

            First of all, go back to my post #17

            and you will see that you don't need any more than you've got to make or check a "master/cylindrical" square. Your don't need a surface guage either - as you've shown - just something that does the job of one. Your "indicator on a shaft in a Vee-block" is just fine.

            Now for the wrist pins. Test them for for straight, round and parallel with the wrist pin in vee-block and and indicator on the top of the pin. Keep the Vee-block and indicator stationary and move the pin on the vee-block and under the indicator. If you rotate the pin the indicator will show any "out of round". Move the pin along the vee-block and check for straightness and parallel in several positions (rotate the pin before moving it axially/along the vee-block.

            You might also run the pin under a good micrometer (a 0.001" is fine as you can easily judge "tenths" (by eye)). You should run the pin through or over your hand as it is very good at picking up "stuff" that an instrument or the eye might miss.

            If all is OK, clamp the pin in the vee-block, put the indicator on the (now) top (end) of the pin and move the vee-block (with the pin in it) under the stationary indicator. This will show the "out of square" (if any) on a good pin.

            While a good surface plate is ideal, any flat surface will do.

            If there is any "out of square" just mark where it is, remove the pin from the vee-block and use the side of the wheel on your pedestal grinder to remove a little of the "high" parts - re-check in the vee-block with the indicator - repeat until getting "very close" (say 0.001" to 0.002"). Then put a strip of "wet and dry" paper (used by Crash-repairers to finish paint-work) - in first "coarse" and then down to "fine/r" grades - with gas or turps. Put the pin in the vee-block and just "nip" it sufficiently to "grip" it in the vee-block. Keeping the vee-block off the W&D commence to "lap" the remaining "high spots" off while regularly re-checking with the indicator.

            It should be OK if within "two tenths" (0.0002") - but better is better of course.

            Making a new "master square" is relatively easy as I showed in a previous thread. I used precision cold-drawn steel tubing which was fine. Anything similar will do. I checked my material before I started by mounting it in my (as new "Chinese" - very good) lathe by indicating less than 2 tenths at each end and then running my lathe (on vee-ways) along its length in at least 4 positions.

            So, I'd guess that you've got all that is needed really.

            I used my "squares" to "true-up my angle plates as well - great result.

            I've seen and been very impressed with your machines and what you have done on them and likewise with your postings and have every reason to think that you can and will "do it".

            I have some other "issues" to deal with on this thread - as I guess you've noticed - but they should not affect or impact on what I've said here.

            If there is any way I can help please ask or say - here, or PM or email - your choice.

            Go for it.

            Comment


            • #36
              Boy this bunch goes on and on and on. All methods will work . Aint none of you building rocket ships and going to the moon. Are making Rolex watches. so settle down boys.
              Every Mans Work Is A Portrait of Him Self
              http://sites.google.com/site/machinistsite/TWO-BUDDIES
              http://s178.photobucket.com/user/lan...?sort=3&page=1

              Comment


              • #37
                Lane, my issue was that Tiffie, as usual, jumped in and hijacked the thread, without reading the solutions presented by myself, Paul and Forrest.

                Originally posted by OldTiffie
                All the previous posts seem to assume/presume that the wrist pins are new, in pristine condition, are round and straight and neither tapered nor worn.
                ???
                Last edited by lazlo; 06-18-2008, 12:06 AM.
                "Twenty years from now you will be more disappointed by the things that you didn't do than by the ones you did."

                Comment


                • #38
                  That's so.

                  Originally posted by lane
                  Boy this bunch goes on and on and on. All methods will work . Aint none of you building rocket ships and going to the moon. Are making Rolex watches. so settle down boys.
                  That's probably so Lane.

                  Taking your advice at face value (lots!!), then if all methods are equally valid then it is entirely - and only - up to each person to make his own judgment as to which method he will use consistent with his resources and requirements for an acceptable outcome.

                  Comment


                  • #39
                    About the Lucas Method of checking a cylindrical square . . .

                    The Lucas Method of checking a cylindrical square analytically combines the (Top - Bottom) differences of three pairs of dial gage measurements, each of which represents the physical summation of three error terms:

                    A) the difference in cylinder radius at Top and Bottom

                    B) "lean" of the cylinder

                    C) "lean" of the machine column.

                    Taken one at a time, the three Top - Bottom differences are essentially these:

                    Difference 1 = A + B + C

                    Difference 2 = A - B - C

                    Difference 3 = A - B + C

                    Three equations, three unknowns. Elementary algebra is sufficient to derive the three individual error terms.

                    Or so the theory goes.

                    In reality, each of the Differences is contaminated by some differing amount of non-repeatability in the position of the dial-gage-carrying machine head on the column. Lucas apparently believes that the head-on-column non-repeatability has already been reduced to some negligible amount by the time they are ready to check column lean, but I'd sure like to see its negligibility demonstrated by measuring a half-dozen or so Top - Bottom dial gage differences with the cylinder-square-on-table and dial-gage-on-column held constant.

                    Comment


                    • #40
                      You guys are sure making a simple but careful metrological procedure complicated.

                      I say this again: squares having parallel reference surfaces (ie: the opposite sides of the beam for a hard square or the opposite sides of a cylinder square) are self checking by the reverse error procedure. Don't need no algebra, machine spndles, or heated arguement.

                      All you need is a flat reference, a surface gage (or simulacum), and a dial test indicator.

                      My earlier post went through ALMOST all the steps. The one I omitted dealt with proving the parallelism and linearity of the opposite sides of the square. I thought that was a no-brainer.

                      You can use this procedure at any scale of prcision you choose. Thousandths in the casual home shop or parts of millionths for reference standard calibrations in the metrology lab. And it gives you an actual number and the sign of the error. And it's quick. And it uses nothing more than stuff from the tool box.

                      The first time you see the procedure in action you will Homer (holler "D'oh!" and slap your foreheads). I will make it a point to discuss this at the scraping class this weekend.
                      Last edited by Forrest Addy; 06-18-2008, 06:50 PM.

                      Comment


                      • #41
                        Originally posted by John Garner
                        The Lucas Method of checking a cylindrical square analytically combines the (Top - Bottom) differences of three pairs of dial gage measurements, each of which represents the physical summation of three error terms:

                        Difference 1 = A + B + C
                        Difference 2 = A - B - C
                        Difference 3 = A - B + C

                        Three equations, three unknowns. Elementary algebra is sufficient to derive the three individual error terms.
                        Yep, you got it John! Get the man beer!

                        Or so the theory goes.
                        Like I mentioned earlier, the method I described is ancient -- it's described in Moltrecht's Machine Shop Practice, there's Lucas Precision, and it's also the basis of Rollie's Dad's Method.

                        As long as the error on the mill/lathe you're using is repeatable, it will subtract out because the measurements are 180° from each other. So the mill or spindle can be out of alignment, and the chuck and spindle bearings can have runout, but you can't have any rocking.

                        "Rollie's Dad's Method of Lathe Alignment



                        What you need

                        • A round bar
                        The bar does not have to be completely straight.

                        • A dial indicator
                        Runout in the chuck is not a problem (for the same reason that a slight bend in the bar is not a problem).

                        What you DON'T need:

                        A tailstock, perfectly straight bar, a collet or precision chuck or any tool bits.

                        Why This Method Works

                        The bar acts as a circular cam. With a perfectly straight bar in a perfect chuck the bar is concentric with the spindle axis. Since we don't live in a perfect world there is almost always a slight offset between the center of the bar and the spindle axis. This offset varies from place to place along the bar due to slight bends and/or imperfect mounting.

                        At any place you pick along the bar the center of the "cam" is some unknown distance from the spindle axis. We'll call this unknown distance 'X'. As you turn the spindle axis the high measurement will be "Bar_radius + X" and the low measurement will be "Bar_radius - X".

                        Their average will be:

                        • ((Bar_radius + X) + (Bar_radius - X)) / 2 =
                        • ((Bar_radius + Bar_radius) + (X - X)) / 2 =
                        • (2 * Bar_radius) / 2 =
                        • Bar_radius

                        As you can see, the value and direction of the deviation have no influence on the final result. That is why it doesn't matter if the chuck is accurate or the bar has one or more slight bends.

                        If the bar is not the same diameter at both places we need to measure the diameters and adjust the readings. Averaging the high and low readings gives us a reading for the local bar radius. We convert that to a reading for the bar center by measuring the bar diameter and subtracting half the diameter (a.k.a. The Radius).
                        "Twenty years from now you will be more disappointed by the things that you didn't do than by the ones you did."

                        Comment


                        • #42
                          Just curious. . .

                          If you have multiple pins, how close could you get to proving the accuracy of square, straightness, and taper by working clean, standing them on end and shoving them together to see what kind of crack of light you get between two in different rotated positions?
                          Cheers,

                          Frank Ford
                          HomeShopTech

                          Comment


                          • #43
                            Damn Frank, where was you six days ago before Lazlo and Tiffie got into this pissing match? You could have save us all from a week of "hard headedness" By the way, How do you measure who has the hardest head?
                            Mel
                            _____________________________________________

                            I would rather have tools that I never use, than not have a tool I need.
                            Oregon Coast

                            Comment


                            • #44
                              Originally posted by lugnut
                              Damn Frank, where was you six days ago before Lazlo and Tiffie got into this pissing match? You could have save us all from a week of "hard headedness"
                              I was actually defending the solution I presented to the original poster, since Tiffie doesn't understand how it works.


                              But I get the message Lugnut. Tiffie hijacked the "What oil to use in my headstock" thread, the "Three Wire" thread, tried his best to hijack the "Oil Prices" and "Global Warming" threads (according to Tiffie all the posters are just "Pissing in the Wind"), and now this one. Next time I'll just let it go, and leave the OP dazed and confused.

                              I guess this is why the post-count is way down here on HSM, and we've lost contributors like Marv Klotz -- even Sir John rarely posts anymore. You can't even post an on-topic answer to someone's question without it turning into a flame-war.

                              To Frank:

                              Originally posted by Frank Ford
                              If you have multiple pins, how close could you get to proving the accuracy of square, straightness, and taper by working clean, standing them on end and shoving them together to see what kind of crack of light you get between two in different rotated positions?
                              Depends on how lucky you are Seriously, you're suggesting that with a large enough sample size, the average errors will zero-out. It may work, but I'd be willing to bet that whatever error the wrist pins have (in taper and lean), they're probably very consistent, due to the mass production method.
                              Last edited by lazlo; 06-18-2008, 10:21 PM.
                              "Twenty years from now you will be more disappointed by the things that you didn't do than by the ones you did."

                              Comment


                              • #45
                                There's merit to Franks proposal. It strongly resembles the three surface solution to proving flatness but aligning pins and observing the light leaking past the tangencies introduces many sources of error. Surface finish for one (haow would a smudge or a stain affect the tranmittance?). Repetitive abutment for another. And angle of view. Also, interpretation of the light transmitted through the gap. Yet another, quantifying the readings - that is determining the amount of error so that a third party can objectively attest the item meets specifications.

                                The three plate solution solves this problem neatly in that the blue indications can be taken up on vellum and preserved. The dimensional limits of blue transferance are well known.

                                The three pin solution proposed for proving squareness works in theory but fails in producing repeatable objective results. Two guys can look at the light leaking through a near-perfect tangency and honestly differ in the interpretation. Tensioned feelers is far more repetitive and thus more likely to provide acceptable proof of gap parallelism but feeler tension is dependent on the immobility of the gages element in the measurement system. Drag a smuge past the tangency and one have of the pins shift ... start over.
                                Last edited by Forrest Addy; 06-18-2008, 10:56 PM.

                                Comment

                                Working...
                                X
                                😀
                                🥰
                                🤢
                                😎
                                😡
                                👍
                                👎