Announcement

Collapse
No announcement yet.

39 tooth splines?

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

  • 39 tooth splines?

    Well a friend calls me other day and asked me if i could machine a shaft with 39 splines? Whats the best way of doing this. I guess a dividing head. Kinda aka Torker post . Now what do i need to do 39 splines guys ?? I have a old dividing head and also a rapid dex but it only does even divisions. I was (Had been) working on a electronic divider system. David Cofer helped me a bunch. I had it mocked up with a huge ground extension plate made for my shaper and so on,.. ( Dave built a PLC unit for it worked very nice) I just dont have room in my shop to set it all up. Anyhow maybe later. Thanx Mike

  • #2
    Buy this and built a quicky indexing head. Then put it back on eBay.

    Comment


    • #3
      You could make a 39 hold plate for your dividing head. Not as hard as you may imagine.

      As I have explained before, dividing heads and rotary tables can be used to make successively more accurate versions of their own plates. You just have to make several generations.

      Make three blanks. Lay out one by hand as accurately as you can. The only thing really necessary is that it have 39 holes. Now use that one to make a second one using the head. You will need to skip over as many holes as your head ratio indicates. So if you have a 40:1 head you would skip over 40 holes or one full rotation plus 1 hole each time. Now use that second plate to make a third. Use the third as a permanent 39 hole plate.

      The head essentially acts as an accuracy amplifier in this procedure. Each successive plate has an error factor that is as many times smaller than the previous one by the head's ratio so if you have a 40:1 head, each successive plate's error is 40 times LESS. The third plate is 40 x 40 = 1600 times as accurate as the original hand marked plate you first made. Thus if you made a 1 degree error on that one, the third one would have only a 1/1600 degree error. That's pretty good and probably as accurate as your head itself. And the work you make with it will have another 40 X reduction in the error. In other words, the errors in laying out the original plate have almost completely disappeared and using a third generation plate will produce errors no greater than a perfect plate would. The head's own errors and normal machining tolerances will now be the major factors.
      Paul A.
      SE Texas

      And if you look REAL close at an analog signal,
      You will find that it has discrete steps.

      Comment


      • #4
        The dividing head will do it but a rotary table won't with out a special plate. I had to cut a gear for the power feed in the apron that had 39 teeth. It would be easier with a horizontal mill,but it can be done on a vertical.

        Comment


        • #5
          Why wouldn't a rotary table do it? And index/spacer wouldn't, but a rotary table would. Just work out the angles and use the vernier. If that's not close enough, do what Paul suggested and make a plate on the rotab instead, then run it through the dividing head to get the accuracy you need. Seems like a plan to me...
          Russ
          Master Floor Sweeper

          Comment


          • #6
            A good stand-by

            Good thinking Paul.

            But for those with a rotary table and a vernier on the hand-wheel (deg:min:sec or decimal degree)Mark Klotz's "Rotary" utility does it in a minute and is hard to beat.

            (Straighten/"align" the number into columns if you wish for yourself):

            Number of divisions = 39

            DIVISION degdec deg min sec
            0 0.0000 0 0 0
            1 9.2308 9 13 51
            2 18.4615 18 27 42
            3 27.6923 27 41 32
            4 36.9231 36 55 23
            5 46.1538 46 9 14
            6 55.3846 55 23 5
            7 64.6154 64 36 55
            8 73.8462 73 50 46
            9 83.0769 83 4 37
            10 92.3077 92 18 28
            11 101.5385 101 32 18
            12 110.7692 110 46 9
            13 120.0000 120 0 0
            14 129.2308 129 13 51
            15 138.4615 138 27 42
            16 147.6923 147 41 32
            17 156.9231 156 55 23
            18 166.1538 166 9 14
            19 175.3846 175 23 5
            20 184.6154 184 36 55
            21 193.8462 193 50 46
            22 203.0769 203 4 37
            23 212.3077 212 18 28
            24 221.5385 221 32 18
            25 230.7692 230 46 9
            26 240.0000 240 0 0
            27 249.2308 249 13 51
            28 258.4615 258 27 42
            29 267.6923 267 41 32
            30 276.9231 276 55 23
            31 286.1538 286 9 14
            32 295.3846 295 23 5
            33 304.6154 304 36 55
            34 313.8462 313 50 46
            35 323.0769 323 4 37
            36 332.3077 332 18 28
            37 341.5385 341 32 18
            38 350.7692 350 46 9
            39 360.0000 0 0 0

            Comment


            • #7
              CNC Mill? How quick do you need it? I just saw a post that Evan offered just such a service and I will throw my hat in the ring also. I only have a Tormach but it will do it.

              P/R

              Comment


              • #8
                With 39 teeth I am assuming it is an involute spline*. A hob is still the best way to do this. But if one is not available or no gear shop is nearby maybe the "rack hob" method that was discussed last year some time might be viable. Basically you have a cutter with a rack form (similiar to a thread mill) with the centerline of one tooth at the same height as the center line of the work. Make a pass. Back the tool out index, cut again. The teeth of the cutter above and below will also cut part of the tooth profile. Perfect , no. But the videos the guy had posted looked like it should work alright.
                Forty plus years and I still have ten toes, ten fingers and both eyes. I must be doing something right.

                Comment


                • #9
                  Tools required.

                  As stated something to Index the 39 divisions, and a cutter.




                  Ash Gear used to carry these. They also had the flat root cutter series.


                  Cheers,
                  Les H.
                  The Impossible Takes Just A Little Bit Longer!

                  Comment


                  • #10
                    Originally posted by oldtiffie
                    Good thinking Paul.

                    But for those with a rotary table and a vernier on the hand-wheel (deg:min:sec or decimal degree)Mark Klotz's "Rotary" utility does it in a minute and is hard to beat.

                    .........
                    I'm well aware of the use of tables but he said he had a dividing head, not a rotary table. Many (most?) dividing heads do not have calibrated dials. Besides, it can get very tedious setting all those angles to the second or nearest 5 or 10 second point: a very easy way to make mistakes. I have a RT and I don't know if I would attempt 39 divisions that way.
                    Paul A.
                    SE Texas

                    And if you look REAL close at an analog signal,
                    You will find that it has discrete steps.

                    Comment


                    • #11
                      Perhaps

                      Originally posted by madman
                      Well a friend calls me other day and asked me if i could machine a shaft with 39 splines? Whats the best way of doing this. I guess a dividing head. Kinda aka Torker post . Now what do i need to do 39 splines guys ?? I have a old dividing head and also a rapid dex but it only does even divisions. I was (Had been) working on a electronic divider system. David Cofer helped me a bunch. I had it mocked up with a huge ground extension plate made for my shaper and so on,.. ( Dave built a PLC unit for it worked very nice) I just dont have room in my shop to set it all up. Anyhow maybe later. Thanx Mike
                      Perhaps not so Paul.

                      That
                      rapid dex but it only does even divisions
                      seems an awful lot like a rapid indexing rotary table (usually 8" - or more) to me - perhaps not. Many of those rapid indexing (RI) rotary tables have a very comprehensive set of dividing plates as well as a 90:1 worm:wheel ratio.

                      The tables for my "Vertex" 8" Quick-indexer are at:


                      90/39 = 2 + 12/39 = 2 + 8/26 = 2 + 29/91

                      So it is quite easy to make just the one 39-hole indexing plate - if the OP has a 90:1 "Quick-indexer" or similar.

                      If he does have the QI-er and dividing plates he can make the gear with the QI rotary table - without making another dividing plate for the dividing-head.

                      Perhaps the OP can let us know.

                      Comment


                      • #12
                        Noobie Question

                        Would it be possible to use a digital readout with a bolt hole circle function to make a dividing plate? I have only done this with an even number of holes but it seems to me that my unit will cut up to 100, perhaps an uneven number would be possible.
                        VitŮŽria, Brazil

                        Comment


                        • #13
                          No reason why not. it's just XY co-ordinates after all. I can't see any reason why odd numbers are excluded up to whatever the DRO can read.
                          I regularly do 3 and 5, larger number I tend to CNC but that's only because I have one and it's easier.


                          [Edit] Just been out and tried 127 on my Onyx DRO and it just spat the co-ordinates out. Bit time consuming but at least it can be done and is accurate.
                          .

                          Sir John , Earl of Bligeport & Sudspumpwater. MBE [ Motor Bike Engineer ] Nottingham England.



                          Comment


                          • #14
                            Hmm - yeah - why not?

                            Well, again DRO wasn't mentioned as an option, but to be fair it wasn't ruled out either.

                            I've had a look at the manual for my yet-to-be-installed DRO for my mill, and there seems to be no limits at all on the pitch circle or the number of holes (odd or even).

                            So far as I can see the rest of it depends on the DRO. Some may only give you the ordinates and you have to sort them out. Mine says it gives successive relative X and Y to "next hole" (ie just move to the screen zero for X and Y). It shouldn't get easier than that. Its all too easy to get lost in successions of absolute ordinates.

                            I am pretty well catered for with my 90:1 rapid indexer and comprehensive dividing plates.

                            I have no problems printing out a table from Marv Klotz' utilities and just using the vernier on the hand-wheel. The trick is to align or start the job at the zero of the table scale and the hand-wheel vernier.

                            Another possibility is to draw the plate in CAD and pick off the co-ordinates of each hole.

                            I have to say that CNC is looking better by the day generally and by the minute with some posts here today!!

                            I had better "get my finger out"!!

                            Comment


                            • #15
                              39 is a standard dividing hole pattern in the 3-plate 40:1 Brown & Sharpe "standard." So my Ellis, and Bridgeport dividing head plates both have 39 holes.
                              "Twenty years from now you will be more disappointed by the things that you didn't do than by the ones you did."

                              Comment

                              Working...
                              X