584,812 active members*
5,286 visitors online*
Register for free
Login

Thread: Gear cutting

Results 1 to 11 of 11
  1. #1
    Join Date
    Feb 2007
    Posts
    29

    Gear cutting

    Hi, I'm new to this forum and to cnc and I'm just in the final throws of scratch building a small but sturdy 4 axis cnc miller. One of my main interests is restoring antique watches and small clocks and one of the uses I want to put my machine to is cutting new wheels and pinions for these. As it is extremely expensive to buy commercial cutters for every wheel or pinion size and, for most of the wheels and pinions used in antique watches, commercial cutters of the right size don't exist, I would like to find a way to cut these gears by generating the shape. What I envisage is using something like a small slitting saw, of less than the tooth space thickness, in the Z-axis spindle, having the blank in the A-axis rotary parallel to the X-axis and moving the X, Y and Z axes in harmony to generate the wheel tooth form (which is cycloidal [church window shaped] for the wheels and semicircular topped with radial sides for the pinions). I have searched as well as I can on the 'net and in the literature I have access to but can find nothing even remotely relevant. Can anyone help or give me a lead to literature which may be relevant please?
    I have the demo version of Mach3 and access to Bobcad and Rhino at the moment. Many thanks.

  2. #2
    Join Date
    Sep 2006
    Posts
    245
    Several books mention using single tooth flycutters. This practically forces you to make a new cutter for each gear you plan to make. The result should be as accurate as the cutter.

    I like your method, since it does not require separate cutters. Do you think the results would be smooth enough? Or would you polish the teeth after cutting them? Friction needs to be kept to the absolute minimum.

  3. #3
    Join Date
    Feb 2007
    Posts
    29

    Gear cutting

    Hi Steve, I don't see why wheels generated in this way shouldn't be good enough to use straight away although it, of course, depends upon the size of steps. In practical terms, I would expect to have to polish the teeth of pinions but If these are smooth I wouldn't anticipate any problem with brass wheels straight from the machine. They will, of course, smooth themselves off very quickly in use. The problem is just how to get all 4 axes dancing in synchrony to generate the curves smoothly.
    As to making cutters, those for plain wheel teeth are not too difficult although, of course, size is a problem and getting both sides of a cutter to cut 60 teeth on a 1/4in diameter gear exactly equal can be a pain. The real problem comes when trying to make pinion cutters in these tiny sizes. Not only do you have to try to get relief on the cutting edges which are quite tight curves, but there is a real problem in measuring the root width of the cutter where the sides of the cutter taper towards each other. I have a very expensive shadow projector, travelling microscopes and other sensitive measuring equipment but even these aren't much good for practical 'nanoscale' engineering!!<G>
    Ian

  4. #4
    Join Date
    Mar 2003
    Posts
    4826
    I believe Bobcad had some kind of a gear tooth drawing function, it was not all that great, and required a bunch of trimming and perhaps redrawing the base of the tooth. Then, the tooth had to be copied and rotated to make a complete gear outline.

    However, for working as you describe on a 4th axis, you might be better off to use a simple subroutine for each tooth profile, and then simply work out the number of repeats to equal the number of teeth in your gear.

    You would first draw the tooth profile full scale, using, I guess, what Bobcad has to offer, which is only a standard involute profile. Then, offset the profile for half the thickness of the slitting saw. That would become your toolpath. It could be interpolated into tiny segments and translated into the proper YZ plane.

    For one tooth at a time, I do not think it would be necessary to have the 4th axis performing live simultaneous motion. You can simply perform YZ interpolation along the path. Then, retract the tool out of the cut, and simply index the 4th to the next position.

    Rather than use a slitting saw for tiny work, I think maybe an abrasive cutoff type wheel would be a better tool choice and these will cut pretty good in brass. You would run the slitting saw on a small dremel tool type spindle to get enough rpm to help retain cutoff wheel life.

    For the purposes of simplication of the process, it might be good enough to assume that the shape of the edge of the cutoff wheel is a 180 degree radius, or perhaps you can dress it to that It will likely wear in to that pretty rapidly, I'm guessing.

    It is important for the sake of running the offset toolpath, that the edge of the wheel be a 'hemispherical shape' so that the offset is a valid representation of the travelling center of the wheel's edge radius.
    First you get good, then you get fast. Then grouchiness sets in.

    (Note: The opinions expressed in this post are my own and are not necessarily those of CNCzone and its management)

  5. #5
    Join Date
    Mar 2006
    Posts
    2712
    watchman, You might snoop around and find a diagram of how a gear tooth is generated by hobbing or shaping. The involute form is generated by a number of cuts by a straight rack form. Its actually a series of flats. The more passes, the more flats the closer the final tooth form to a true involute. If the gears in question are spur gears, you should be able to position the spindle axis and rotary axis then cut the teeth along the axis parallel to the gear centerline. The more cuts per tooth, the better the tooth form. Stack multiples of any gear and place extra stock on ends to prevent excessive burrs or damage at start and finish breakout. If they are flat, personally I,d have them cut on a wire EDM. Mitsubishi has a great gear program as an option.
    DZASTR

  6. #6
    Join Date
    Feb 2007
    Posts
    29

    Gear cutting

    Thanks guys, all good information but I think you may have missed the info on the type of gears I'm looking at cutting. these are gear wheels and pinions for watches - something like 60 teeth on the wheel of 3/8inch diameter! The tooth form has to match the existing wheels which are always cycloidal in form not involute. Cycloidal teeth have straight parallel sides up to the pitch circle and then a semicircular curve from each side meeting at a point forming the addendum. I have drawn out the path of a tooth and the path a cutter would have to follow in the Z and Y axes whilst the work turns about the X axis but it isn't a constant amount. that is, I couldn't just tell the machine to rotate the work D degrees while moving the Z to E and the Y to F. Is it possible to use mathematical expressions in G-code for the Y and Z values or could I perhaps ..... its just occurred to me - I guess that if I just draw a curve plotting Z against Y in a CAD program and generate the G-code from this I could just edit in a rotation of the A-axis to generate the curve of one side of the tooth, repeat it mirrored for the other side of the tooth and then use this as a subroutine to cut the right number of teeth - would this work?? Thanks..

  7. #7
    Join Date
    Mar 2004
    Posts
    1542
    Quote Originally Posted by watchman View Post
    Cycloidal teeth have straight parallel sides up to the pitch circle and then a semicircular curve from each side meeting at a point forming the addendum.

    ..... its just occurred to me - I guess that if I just draw a curve plotting Z against Y in a CAD program and generate the G-code from this I could just

    Thanks..
    You just described the problem. Each side of the tooth is a linear move in Z and Y, followed by a circular move. Use Cad to solve for your values and make a subroutine to cut each tooth. Rotate A and repeat.

    Karl

  8. #8
    Join Date
    Mar 2003
    Posts
    4826
    Wire EDM sure sounds like a good solution, not that I've had any experience with them. No stress on the part while cutting, simple 2d profile to cut, no cutter wear etc. Thats where I would start looking.

    On a tiny tooth like that, I doubt that you'd really see the difference between a cycloid and an involute. The difference between the curves would probably be so minute as to be within the normal error range of the tool doing the cutting.
    First you get good, then you get fast. Then grouchiness sets in.

    (Note: The opinions expressed in this post are my own and are not necessarily those of CNCzone and its management)

  9. #9
    Join Date
    Jul 2005
    Posts
    12177
    watchman; you stimulated me to go Googling and I found this;

    http://www.csparks.com/watchmaking/C...rs/index.jhtml

    and stole an excerpt from it:

    First of all, B.S. 978 calls for using the arcs of circles instead of true epicycloids as the addendum curves. The circular arcs are chosen to closely match the epicycloidal form. Needless to say, making a circular form tool is easier than generating an epicycloid. Another key dimension provided by the standard is the addendum height for the wheels. This height is chosen to be 95% of the theoretical addendum height. For pinions, the standard calls out extra spacing between small pinion leaves. It also specifies simple addendum shapes for the pinion leaves.

    Which seems to suggest that complicated calculations for cycloidal curves are not essential.

    That link also links to another page describing in great detail the making of the cutters. I don't think I would have the patience to be a watchmaker.

  10. #10
    Join Date
    Oct 2005
    Posts
    859
    Go to this thread on the Bobcad forum to see info you may use.

    http://216.117.147.20/bobcadsupport/...1583#post11583

  11. #11
    Join Date
    Oct 2005
    Posts
    859
    Quote Originally Posted by HuFlungDung View Post
    Wire EDM sure sounds like a good solution, not that I've had any experience with them. No stress on the part while cutting, simple 2d profile to cut, no cutter wear etc. Thats where I would start looking.

    On a tiny tooth like that, I doubt that you'd really see the difference between a cycloid and an involute. The difference between the curves would probably be so minute as to be within the normal error range of the tool doing the cutting.


    We use wire EDMs to cut gears all the time. It is very accurate and the gear module in our wire software can get accuracy settings to .00000001 in calculating form. Not that any machine could cut it this close but the calculations are as close to perfect as anyone would ever need. However we can cut very accurate gears on our wire EDM. (Let me know if you need some.)


    Bobcad can generate very good gear forms though.

    Follow the link above to see how to make that one tooth form cut the entire geare profile in bobcad.

Similar Threads

  1. Gear Cutting for a begginner?
    By Owain1602 in forum MetalWork Discussion
    Replies: 53
    Last Post: 01-10-2007, 05:19 AM
  2. Cutting a planetary gear...
    By WilliamD in forum MetalWork Discussion
    Replies: 11
    Last Post: 11-09-2006, 09:56 PM
  3. gear cutting software
    By DrStein99 in forum CNC (Mill / Lathe) Control Software (NC)
    Replies: 2
    Last Post: 02-21-2006, 05:29 PM
  4. Gear Cutting
    By chuckknigh in forum DIY CNC Router Table Machines
    Replies: 12
    Last Post: 06-02-2004, 12:30 AM
  5. Gear cutting
    By danswen in forum MetalWork Discussion
    Replies: 1
    Last Post: 11-04-2003, 06:27 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •