Announcement

Collapse
No announcement yet.

Shop Made Tools

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

  • Originally posted by rcaffin View Post
    ...

    Nice old photos, but I have to confess, I cannot see what the guy is actually trying to achieve there. Perhaps you could explain it in some detail to me?

    Cheers
    Roger
    Even I get it

    If the bar is parallel, then if you put the base of the indicator stand against the bottom of the bar and the indicator at the top, then repeat for the other side of the bar, the difference in reading will be twice the error in squareness.

    The key is that the base holding the indicator has to touch the bar as well as the indicator itself. If everything is parallel and square, then the indicator will read the same number from both sides of the bar, at the same height. If its out of square, then the reading will be different side to side.

    Yeah, I had to stare at it for a while, and read the text, to see what was going on.

    David...
    Last edited by fixerdave; 11-08-2016, 01:56 AM.
    http://fixerdave.blogspot.com/

    Comment


    • Originally posted by rcaffin View Post
      That's how we make optical flats: lapping three units against each other in a cycle. They routinely get down to lamba/10 that way.
      You can also make right angles that way, by lapping three units against each other. I never said it was easy though.

      Roger
      Yes, there's a tried and true process for developing flat surfaces using three plates. And further, to develop perpendicular faces by comparison among three pieces. But the key to producing the correct geometry accurately and efficiently is a methodical process using analysis to determine where material must be removed to come closer to the goal.

      The process was developed fairly early in the industrial revolution in the effort to produce more accurate machine tools than the ones used to produce the parts. In other words, if you had a machine with erratic ways used to make parts for the next machine, that new machine would have exactly the same errors as the one that produced it. Lapping was used initially to address the problem, but it's hard to control the process. In other words, you have two parts that are worked against each other to try to get more accurate surfaces by averaging. The problems is that you can't ensure that only the high spots on each part will be worked down. To some extent, the high points of one part will further lower the adjacent surfaces on the other part.

      So, to get back to the three plate method, by using a regular rotation of the plates changing which plate is the master and which other two plates are scraped down to match it, you can get to very flat surfaces on all three. But there are potential pitfalls if the process isn't understood and used properly. For example, if you used three rectangular plates rather than square ones and always lined them up along the same edges, even reversing them won't prevent the possibility of a curved Pringles chip shaped surface. To eliminate that prospect you must also compare the plates using 90 degree rotations rather than always 180 degrees.

      Using the methodical process will also get you accurate squares. However, if you aren't using it properly you can perfectly well get three squares that are all alike but not 90 degree angles. The knowledge has been around a long time and reading up on Maudslay, Whitworth, and books on machine tool rebuilding will bring you up to speed. It's a good deal more than just lapping three plates together to get a flat. If you've worked with grinding lenses or mirrors you'll know that lapping is also used to produce nice (and accurate) concave and convex mirrors and lenses.
      .
      "People will occasionally stumble over the truth, but most of the time they will pick themselves up and carry on" : Winston Churchill

      Comment


      • Using the methodical process will also get you accurate squares. However, if you aren't using it properly you can perfectly well get three squares that are all alike but not 90 degree angles.
        I always understood that comparing three squares against each other in turn was a reliable test of squareness and I cannot understand how it could not be.

        Comment


        • Originally posted by The Artful Bodger View Post
          I always understood that comparing three squares against each other in turn was a reliable test of squareness and I cannot understand how it could not be.
          I'm drawing on several things there and to be honest, the statement as it stands is probably an oversimplification. Example one. An article in a woodworking magazine on purchasing tools, advised readers to check squares in the store before purchase. Comparing several squares to one another you might find most match but an occasional outlier. Don't buy that one. An astute reader pointed out that this only tests whether they are the same without actually distinguishing whether the large group or the single different square were accurate.

          In another instance someone, perhaps actually Forrest Addy, pointed out that in a shop environment where toolmakers typically make some of their own tools, there's tendency to borrow something from another worker when making one of your own. The assumption is that the other guy's square is a good one and thus your own will be as well. There's a chance to propagate errors around the shop if no one goes back to first principles to verify that the master you're working to is a good example.

          Thinking through the process of generating three squares. First scrape all three to flat surfaces by comparing to a surface plate. They won't be accurately square, but just two approximately square flat surfaces on each. If one is taken arbitrarily as the master and number two and three scraped to match, comparing two and three to each other will show in what way square one is off. The second two squares will touch each other at the top or the bottom. So pick one of them, say square two, and scrape the high end making a guess about how much, re-establish it as a flat surface against the surface plate, and take this one as the master to which the others (squares one and three) are scraped and then compare squares one and three. There will probably still be a gap at top or bottom, but smaller than before. Keeping up the rotation and comparison cycle will develop accurate squares without resorting to an outside master. When all three can compare to each other, you can be assured that they're all square which is what you were saying.
          .
          "People will occasionally stumble over the truth, but most of the time they will pick themselves up and carry on" : Winston Churchill

          Comment


          • Just a friendly suggestion - maybe this discussion could be moved to a new thread, instead of filling up the Shop Made Tools?

            Comment


            • Good point, I will start another thread.

              Comment


              • Not a general method

                If the bar is parallel, then if you put the base of the indicator stand against the bottom of the bar and the indicator at the top, then repeat for the other side of the bar, the difference in reading will be twice the error in squareness.
                The key is that the base holding the indicator has to touch the bar as well as the indicator itself. If everything is parallel and square, then the indicator will read the same number from both sides of the bar, at the same height. If its out of square, then the reading will be different side to side.
                I understand that, but it makes several very restrictive assumptions:
                That the blade of the square is dead parallel
                That you can get to both sides of the blade in equal manners
                That you handle the indicator very carefully to avoid any tilt

                For many it might be a lot simple to put two square against each other on a surface plate and to measure the gap between them with feeler gauges. I believe one can detect by eye a gap between two edges like that down to the sub-micron level.

                Cheers
                Roger

                Comment


                • Vise crank handle

                  Not home-shop-made, but a government job.



                  I made a dozen of these as Christmas gifts for co-workers a few years ago.

                  Comment


                  • Made this:
                    Click image for larger version

Name:	7E7148F6-43A7-46F2-B44B-ABAAA6303A6C.jpeg
Views:	436
Size:	41.8 KB
ID:	1835194
                    To make this:

                    Click image for larger version

Name:	49B725CA-EF37-4A4F-BC4C-2F3D2340E814.jpeg
Views:	431
Size:	47.6 KB
ID:	1835195

                    To fix this:
                    Click image for larger version

Name:	7F39C3D1-C961-4972-A7F2-B6DEA423B9B4.jpeg
Views:	432
Size:	55.1 KB
ID:	1835196
                    Last edited by Boats69; 10-24-2019, 12:46 PM. Reason: Replaced defunct Photobucket link!

                    Comment


                    • That looks suspiciously like a rotary broach. What angles did you put on the holder and cutters? The one in the picture looks like a little more than 2 degrees, but it's hard to tell for sure. Did it cut freely or did you have to push hard?
                      Kansas City area

                      Comment


                      • It's amazing how far us HSM types will go to save something like a 4 jaw....

                        Lovely job on the broaching tool. If you're willing to share I'm sure a lot of us would like to see the details of the joint that provides the oscillating action.
                        Chilliwack BC, Canada

                        Comment


                        • I borrowed freely from this site ... it is close to 1.5 degrees, don't have all the fancy tooling yet so basic math and shims to get the angle right. The 6mm square went surprisingly easily into both silver steel (drill rod?) and EN8 - I would say medium pressure....

                          Comment


                          • I am lazy with pics. Will do a strip and post some photos when I have a bit of time.

                            Comment


                            • Some pics of the internal workings, pretty straightforward: a thrust bearing on the one end and a ball bearing on the other, added 1.5 deg on the face of the rear mount and used a commercial MT2 shank. The only somewhat tricky bit was to get the right clearance on the nut for the preload on the thrust bearing.

                              Click image for larger version

Name:	6593A9C5-07FA-4AEC-9746-8DE44E9870D4.jpeg
Views:	378
Size:	64.4 KB
ID:	1835192


                              Don't critize the over-engineering - the bearings are what I had spare and the rest of the dimensions followed from there.

                              I am happy with the end result.
                              Last edited by Boats69; 10-24-2019, 12:40 PM. Reason: Replaced defunct Photobucket link!

                              Comment


                              • Using bearings like this in a way aids in ensuring that the broach doesn't try to rotate during use and produce a wandering hole. So really I'd consider it as good choice vs over engineering.

                                Plain bearings could be used too. But the drag from them likely would require a anti rotation arm that runs against a fixed point to prevent the broached hole coming out with a slight spiral to it. So a decision which was made for reasons of convenience and to use what was on hand likely turned out to be optimum for at least one other reason.

                                It's great to see the pictures and read the experience about angle and ease of use with light to moderate pressure. I think this will inspire more than a few of us to follow in your footsteps.

                                And a hearty WELL DONE! for this very nicely made unit.
                                Chilliwack BC, Canada

                                Comment

                                Working...
                                X