No announcement yet.

How small is a thou- must be a million of 'em in an inch

  • Filter
  • Time
  • Show
Clear All
new posts

  • How small is a thou- must be a million of 'em in an inch

    I heard that somewhere. Kind of funny in a way. Here I am today getting my new ball bearing indicator worked out, making an interface piece to go from the lever to the indicator tip. Got that done and the basic assembly together, then a bit of testing. According to my math, pivot point distances vs lever length, etc, it should multiply by five, which it does exactly. One thou on the indicator is 2 tenths.

    So I'm playing with it, trying to determine if there's stickage or 'hysteresis', and I find that it's very responsive. I discovered also that it's easy to feel a couple tenths of movement with my fingers. I can slide the base back and forth to the point where I can split one mark into four, which corresponds to 50 millionths, and I can feel the base moving that little bit. I know that fingers are very sensitive, but I didn't realize they could be that sensitive.

    I suppose size is all relative. 1 thou is quite a large distance, even in a home shop. In terms of being able to machine to that tolerance, it seems rather crude to be able to achieve that. A couple of tenths should be relatively easy, and measurement to that tolerance should be relatively easy as well, although much care would be needed to measure repeatably and consistently. I've been interpreting my dial caliper to within a half-thou, which I thought was pretty good.

    I haven't taken the opportunity to check out the effects of hand heat on a measuring instrument, but in strictly mechanical terms it shouldn't be a stretch to be able to differentiate one tenth quite easily. I'm going to suggest that the ability to do this is 90% operator skill, more of a study in the precise manipulation of measuring instruments, rather than the limitations of instrumentation itself.

    I'm not addressing the issue of when you should work to tighter tolerances- that has been discussed elsewhere. I'm interested though- for those who do often need to work well down in the sub-thou range, do you have any special procedures that you follow, etc, other than holding your mouth right?
    I seldom do anything within the scope of logical reason and calculated cost/benefit, etc- I'm following my passion-

  • #2
    Sound like a good addition to you metrology kit , could you post a picture of how you have set it up?

    If I am getting down to the measurements that you have described I try to prevent any heat transfer from my hands to the measuring tool ,it can make a lot of difference at those dimensions.



    • #3
      What modification did you make to multiply the sensitivity of your indicator by 5?
      Peter - novice home machinist, modern motorcycle enthusiast.

      Denford Viceroy 280 Synchro (11 x 24)
      Herbert 0V adapted to R8 by 'Sir John'.
      Monarch 10EE 1942


      • #4
        Batteries on charge, pics tomorrow. The 'front end' is a ball bearing, mounted on the end of a lever arm. The arm pivots at a point about 5/8 from the bearing axis, and actuates the dial indicator at a point 3-1/8 from the pivot point. That's what gives the multiplication. The 'front end' bearing, or contact bearing as I've referred to it, has been selected to give an undetectable amount of runout so it doesn't negate the effect of having this multiplied sensitivity. Once I get the mounting base made and mount it on the lathe, I'll be able to check the runout of the bearing to a finer degree. Because it will roll against a workpiece in the chuck, rather than scratch along, it might be more susceptible to minute debris- but that would show up as distinct from eccentricity and should pose no problems. I expect to see the dial doing a fair bit of dancing, but it will still give a perfectly good indication.
        I seldom do anything within the scope of logical reason and calculated cost/benefit, etc- I'm following my passion-


        • #5
          This should clear things up a bit- this is the assembled unit ready for mounting on an adapter base which I haven't made yet.

          This is the guts

          and a couple details of the interface piece between the lever arm and the indicator stem.

          The interface piece has a bit of a knife edge filed into it, and the stem has a notch ground into it to allow the pieces to nest together. The piece of foam rubber keeps the parts from separating- spring pressure from the indicator is all that keeps the pieces together. Worth noting is that the normal spring pressure is amplified at the front of the contact bearing by the same factor as the magnification, 5 times. This should not be a problem, but I do have the option to counter that by adding a spring. At the moment, my thinking is that I won't need to- the added pressure of the bearing against the workpiece might even be a benefit.

          Last edited by darryl; 03-09-2013, 08:59 PM.
          I seldom do anything within the scope of logical reason and calculated cost/benefit, etc- I'm following my passion-


          • #6
            I like the ingenuity that thing shows, but...I am wondering, when measuring to tenths, if you need to worry about the fact that the motions are all arcs, and not linear motion. Also, because the front bearing moves on an arc, it will be moving up and down relative to center height....


            • #7
              There is no need for any absolute measurement when centering something, so if there are errors it won't matter. I did think about the errors due to the arcs- at the point where the lever actuates the indicator, I made that as close as possible to a linear translation. It's quite linear there anyway. The larger error is at the bearing and pivot end, but again it isn't much, and is concentrated over a small arc- I have about 100 thou of total range there, so if I center the range and have about 50 thou of motion either way, the error is small. That's more than enough for use as a centering indication.

              My bigger concern was hysteresis, or lost motion not measured by the movement of the dial pointer. During testing I found that I could go back and forth an almost imperceptible amount and the needle would follow. Getting a repeatable motion on the needle through a quarter of the distance between dial markings was easy. Since that corresponds to one twentieth of a thou, I think I'm going to live with that

              The worst thing affecting the system is due to the indicator itself- the stem has a sticky spot in the first 200 thou or so of its travel. I don't understand how such a precise instrument from a prestigious company like Power Fist can be like this - ha ha. This particular indicator is a bit drunken, having a periodic error of almost two thou in repeatable spots, and it also is out by 3 thou at a distance of 1 exact inch. I noted this and haven't used it for anything, except for this project where all that doesn't matter. I am going to try polishing the stem to get rid of the stiction though. I can use it in a range where it doesn't stick, but it bothers me-
              I seldom do anything within the scope of logical reason and calculated cost/benefit, etc- I'm following my passion-