Announcement

Collapse
No announcement yet.

Facet cut gears

Collapse
This topic is closed.
X
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Facet cut gears

    On seeing the Gearotic program introduced by John S. I decided to figure out how the process was implemented. A little research shows that it isn't a new idea although the implementation by Art Fenerty in Gearotic is new. The earliest example I could find was a machine patented about 40 years ago that performed the same process via a completely mechanical system instead of a CNC machine.

    I determined the method by which facet cut gears are made by investigating the way that an involute is approximated in most CAD programs. The end result in the several programs I checked is a collection of line segments instead of an arc or spline curve primitive. This lends itself to easy description mathematically which is why it is done that way.

    The investigation resulted in this initial description of the angles and offsets required to produce a 14 tooth 9DP 20PA gear with an OD of 2 inches.



    Here is the setup I used to make such a gear. The G-code is hand written using a template I created that will allow the easy plug in of variables for any gear. At the moment I am contemplating making a simple stand alone program for the creation of spur gears only. It will be a while before I get to that because I am expecting the new 10" mirror for my next telescope to arrive shortly. The material is black acetal with a 1/8 solid carbide cutter at 5000 rpm.



    Note that the gear was selected to emphasize the facets created. Gears with more teeth will have much less or even invisible faceting.



    The run time of the process was about 30 minutes. The G-code is far from being optimised and the time can be reduced very considerably. The machine is also capable of cutting an acetal gear like this about twice as fast but that makes it hard to see what is happening.

    Here is a video of the process. Sorry about the shakycam but it didn't lend itself to using a tripod while making a gear in real time.



    Next post is to show some simple low angle herringbone gears that I made by this method.
    Free software for calculating bolt circles and similar: Click Here

  • #2
    Herringbone gears:

    It turns out that these were easier to make than I thought. Even with this low helix angle the teeth should much stronger than a straight cut gear.They roll together nicely by hand. I havn't yet jigged them up on power to find out how they sound.





    Free software for calculating bolt circles and similar: Click Here

    Comment


    • #3
      Nice work. Most of the people I know that produce straight spur gears cut them on a water jet, how accurate a tooth profile do you think can be produced that way? Not that it really matters, I am sure the gears produced by the methods mentioned here would be fine in any practical application.
      James Kilroy

      Comment


      • #4
        Well done Evan. I enjoyed watching this process. your set up is very well thought out, did you make the milll yourself ? Alistair
        Please excuse my typing as I have a form of parkinsons disease

        Comment


        • #5
          The accuracy of the gears is limited only by the accuracy of the machine since any number of facets may be used to approximate the curve. In this instance I used ten facets which mathematically results in no deviation greater than .0007" on this particular curve. As the curve grows less acute the accuracy with ten facets grows better. The total run time to make a gear will not increase in direct proportion to the number of teeth. As the number of teeth increases the number of facets required to approximate the tooth decreases.

          Yes, Alistair, I designed and built the mill.
          Free software for calculating bolt circles and similar: Click Here

          Comment


          • #6
            Nice job Evan! I like the term "Facet-cut Gears."
            "Twenty years from now you will be more disappointed by the things that you didn't do than by the ones you did."

            Comment


            • #7
              Originally posted by Evan
              On seeing the Gearotic program introduced by John S. I decided to figure out how the process was implemented. A little research shows that it isn't a new idea although the implementation by Art Fenerty in Gearotic is new. The earliest example I could find was a machine patented about 40 years ago that performed the same process via a completely mechanical system instead of a CNC machine.

              I determined the method by which facet cut gears are made by investigating the way that an involute is approximated in most CAD programs. The end result in the several programs I checked is a collection of line segments instead of an arc or spline curve primitive. This lends itself to easy description mathematically which is why it is done that way.

              The investigation resulted in this initial description of the angles and offsets required to produce a 14 tooth 9DP 20PA gear with an OD of 2 inches.



              Here is the setup I used to make such a gear. The G-code is hand written using a template I created that will allow the easy plug in of variables for any gear. At the moment I am contemplating making a simple stand alone program for the creation of spur gears only. It will be a while before I get to that because I am expecting the new 10" mirror for my next telescope to arrive shortly. The material is black acetal with a 1/8 solid carbide cutter at 5000 rpm.



              Note that the gear was selected to emphasize the facets created. Gears with more teeth will have much less or even invisible faceting.



              The run time of the process was about 30 minutes. The G-code is far from being optimised and the time can be reduced very considerably. The machine is also capable of cutting an acetal gear like this about twice as fast but that makes it hard to see what is happening.

              Here is a video of the process. Sorry about the shakycam but it didn't lend itself to using a tripod while making a gear in real time.



              Next post is to show some simple low angle herringbone gears that I made by this method.
              Brilliant piece of work!
              Mike

              Comment


              • #8
                drmico60: lol, Next time just hit reply insted of quote.. Or at least, remove the several pages of images from the reply.
                Play Brutal Nature, Black Moons free to play highly realistic voxel sandbox game.

                Comment


                • #9
                  Nice work Evan. It's amazing how accessible easy gearcutting has just become to the home cnc machinist in a very short time.
                  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

                  Comment


                  • #10
                    Can anyone explain how any of this (Evan's method and the Gearotic method) is different that what JS was doing with this demonstration?

                    http://bbs.homeshopmachinist.net/sho...t=helical+gear

                    The Gearotic stuff is pretty interesting in that it simplifies a lot of the work, but I liked the D-cutter idea as it allows making gears that have less than 1/8" tooth separation (limitation of the endmill diameter).

                    I don't imagine there's any reason Evan's calculations or the Gearotic software can't work with tapered D-cutters. This is the standard base circle method used by shapers for cutting an involute with a rack form cutter except using multi-axis CNC.

                    Comment


                    • #11
                      Originally posted by dp
                      Can anyone explain how any of this (Evan's method and the Gearotic method) is different that what JS was doing with this demonstration?

                      http://bbs.homeshopmachinist.net/sho...d.php?t=24252&
                      It's all the same process -- you're CNC milling a gear from the solid. If you don't want to generate the G-Code directly, the Gearotic is doing, and Evan is doing by hand, you can just download a DXF file and do the same thing on any CNC mill. Evan has done that before too (milled a DXF gear model directly).

                      Evan, you sure turn your calculations into a Perl script that generates G-Code. That would be very popular
                      Last edited by lazlo; 10-15-2010, 12:00 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


                      • #12
                        nice

                        Evan

                        Impressive, nice work

                        Leesr

                        Comment


                        • #13
                          Evan, you sure turn your calculations into a Perl script that generates G-Code. That would be very popular
                          I intend to do exactly that. However, I am easily distracted. Enormous optically perfect shiny objects tend to do that. The mirror I ordered is accurate to 1/20 wave!
                          Free software for calculating bolt circles and similar: Click Here

                          Comment


                          • #14
                            Enormous optically perfect shiny objects tend to do that. The mirror I ordered is accurate to 1/20 wave


                            Evan, don't drop it!
                            Location: The Black Forest in Germany

                            How to become a millionaire: Start out with 10 million and take up machining as a hobby!

                            Comment


                            • #15
                              Originally posted by Evan
                              I intend to do exactly that. However, I am easily distracted. Enormous optically perfect shiny objects tend to do that. The mirror I ordered is accurate to 1/20 wave!
                              Very cool - did you get a conical mirror? I've been thinking about that option.

                              Comment

                              Working...
                              X