No announcement yet.

Rotary table vs. dividing plate

  • Filter
  • Time
  • Show
Clear All
new posts

  • Rotary table vs. dividing plate

    I'm cutting gears ('wheels' to the horologist) for a clock. So far they have been of tooth counts that make it easy to perform the indexing with the rotary table. However, I will have to make a 21 tooth escape wheel, and 21 divides into 360
    17 1/7 times. I'm prepared to makke a 21 hole indexing plate if need be, but I got to thinking about how accurate one could be just using the rotary table. Here are my calculations:

    Wheel diameter: 2 inches.
    RT is calibrated to 2 minutes, with a vernier to get to 10 seconds.

    By eye, I believe I can come within 1/2 minute without even using the vernier.

    So how accurate is 1/2 minute? If the circumference is 3.141 x 2 = 6.282 inches, then there are 6.282/360x60 inches/minute of arc, which comes to 0.291 mils/minute,
    or .145 mils/half minute.

    If I haven't blown the math or made any dumb assumptions, then I'm thinking that I can easily be as accurate indexing with the RT as I can making a plate. I'm not at all sure I can drill to .15 mil precision.

    Any thoughts?

  • #2

    I don't know what brand of rotary table you have but my 10" has the same 10 second vernier and specifies an accuracy of only 30 seconds or a half minute. You should check the specs on yours.

    30 seconds at 5" (the radius of my table) is 5 X sin(1/120*) or 0.00073". At smaller raidi, the error is proportionally less. About 0.000146" at your 2" diameter. I believe your calculations are essentially accurate.

    You should also consider the possible errors in an indexing plate. How close can you drill a hole to the specidied location? Will you ream it? And how would you make the plate? If made with the rotary table, it can't be better than it unless special techniques are employed.

    I assume you are talking about using the plate directly and not with the rorary table to achieve the 1/7 divisions. In this mode, it will have the same error as the table PLUS any of it's own errors. It would be better to make plates for use with the table as their errors would then be divided by the worm ration of the table (1/90 or 1/72 or whatever it is). This greatly increases their accuracy.

    If you do use the table directly, by all means, use the vernier. The fact that the table's basic accuracy is only 30 seconds does not mean that you can set the dial within that amount and still get that accuracy. The 30 second stated accuracy is from the actual dial setting so if you are off by 30 seconds in setting the dial and the table can be off by 30 seconds from that dial setting, then you could be off by 30 + 30 = 60 seconds or a full minute. I don't think you want that.

    Paul A.
    Paul A.

    Make it fit.
    You can't win and there is a penalty for trying!


    • #3
      I think your accuracy with the rotary table dial will be more than satisfactory. Clock gears typically have a "rattling" fit anyway, for absolute minimum drag, so you'll be fine.

      People used to make clocks by marking out and filing the teeth by hand! Don't worry about it.
      Try to make a living, not a killing. -- Utah Phillips
      Don't believe everything you know. -- Bumper sticker
      Everybody is ignorant, only on different subjects. -- Will Rogers
      There are lots of people who mistake their imagination for their memory. - Josh Billings
      Law of Logical Argument - Anything is possible if you don't know what you are talking about.
      Don't own anything you have to feed or paint. - Hood River Blackie


      • #4
        Your accuracy for a clock or a gear for a lathe would be fine.You only have to be extremly accurate if you are building gears for the space shuttle. I had some teeth breack off a lathe gear once, I TIG welded them and then ground them to shape with a hand grinder using my calibrated eyeball to size them. That was 5 years ago and they still run fine. If you want to make an accuate dividing plate and you have a lathe try this. Cut a large cardbourd or wooden circle (as large as you can to attach to the rear of the spindle on your lathe and as accurattly as you can)and attach it to the back spindle (as centered as you can)(how to attach it I leave up to you as long as it's sturdy). Put a string all the way around the circle and cut it where the ends meet. Take the string off and measure it, then divide by 21. Then mark the string or use a tape measure and mark the graduations around the circle. Put a piece of tape with a mark on it for your start point. You will also need a drill holder for your tool post on your lathe or just a bar of metal (heat treated is better) mounted in your tool post with a hole in it to act as a drill bushing (note that the drill bushing should be long enough to keep the drill straight). Now rotate(with the lathe in the lowest gear and everything else locked)by hand your new indexing plate that is mounted on your chuck on a mandrel(or in a collet for the purists)and drill/ream the holes(if drilling use a split point new drill). The larger the circle the less error you will have. If you do try this, make some other hole circles at the same time while you have it set up. Enjoy.



        • #5
          Thanks for the replys folks. I'm going to go ahead with just using the RT unless I figure out some reason not to.

          Mike: yeah, that's about how I would have made the 21 hole plate. I would have drilled 23 holes across a thin strip of shim stock (2 holes for overlap attachment) and fitted the resultant ring around a wodden disk with a groove in it; the holes in the strip accept a detent during indexing.




          • #6
            Looks like prices have gone up - or this guy is just hopeful, but here's one on ebay for $59. This is the exact model I use, and it's as good as the Blake IMHO.