did anyone made a gear with his 3axis cnc

[Servo Wizard]

max_imum,

I did not keep files for the indexer because it was a One Off production. It's a relatively simple design but it would be difficult to convey the details in words. The indexing lever rotates on a sprague bearing which facilitates free wheeling when it returns to the initial position. The indexing detent pin is also a spool valve which supplies the pneumatics for the indexing cylinder. Don't feel bad if you don't comprehend those details, I have 40+ years of experience and day dreaming behind my projects.

The tool path is what I refer to as a simple Zig Zag Stick Tool path, however the geometry is not simple as it involves 32 pairs of involute arcs per cogg. If you study the previous pictures you can see how the tool path functions

Servo

[miljnor]

so basical a "kellering move" over the gears tool path with a set stepover and this is with NO indexer movement while cutting???

I was mainly interested in "does the indexer move while cutting" or is it stationary and just rotates after every segment. The later i would pressume

[Servo Wizard]

miljnor,

Yes you have it right, the indexer remains stationary while the cutting tool cuts the group of coggs.

Those sprockets were manufactured for use on High-Tech road racing Gokarts that are powered with motorcycle engines that have a 6 speed gearbox. Several of the Left coast Gurus had attemped to manufacture the sprockets but the splined gearbox output shaft presented a problem for them. Every engine came equipt with a 13 tooth chain sprocket which did not fit into the gearing scheme so I found use for all those tool box sprockets. I no longer have pictures of the back side of the flanged sprockets but the attached tool path picture tells the rest of the story.

BTW, You're revealing your age when you use words like "kellering".

Servo

[miljnor]

what!

darn I will have to stop using antiquated sayings!

besides I am very young! NO realy!

well maybe not, but damn I still look good!

[JBV]

Made a gear with my 2 axis (so far :P) converted mill and a 5 mm endmill

[max_imum2000]

i assume this is a foum cutter ?
whats the dimenstion of the smallest gear u think u can make ?

[JBV]

Thats a 10 module, stub involute 20 deg gear with 10 teeth (100 mm pitch diameter) made in 19 mm thick MDF I don't know how small module i can cut for a given endmill ! I just loaded the gear into cad and measured

Just made it because i only have 2 axes converted on my benchtop mill so far.. And a gear seemed like the most interesting 2d part to make If i had a rotary table i would use cnc to grind a cutter with the exact tooth form for the intended gear (generated from a program) and then index and cut each tooth one by one.

[smoregrava]

Do I understand you rigth Servo wizard? You milled a pocket at the backside of your Belt sprocket that fitt outside the 13 tooth chain sprocket. And you used screws to hold the chain sprocket to the belt sprocket.? Then you did not need to make that inside spline.

Anyway a beautiful work

[gt3073b]

I'm trying to make sense of the whole theory of operation on this tool path. "Kellering" didn't show up in either Google or Wikipedia, so here is what I think I understood from the zip files. Forgive me for the long-winded post that follows....

Since a belt sprocket would still mesh properly just sitting on a flat belt (ie the belt not wrapped around the sprocket) then you can cut a profile the shape of the belt with the teeth in a straight line tangential to the sprocket. Mutiple teeth are cut since more than tooth is inside the pitch circle of the sprocket. This would allow you to do a single tooth design and linearly array it several times without the hassle of cutting a round part in cartesian coordinates. Basically its like using a rack to cut the pinion.

Now about your indexer...
If N is the number of teeth on the sprocket, then you could either index 360/N or 360*M/N where N is not divisible by M. Cutting every Mth tooth and indexing the cutter enough times around would then eventually get every tooth cut. So when you cut the first tooth, you can cut it anywhere on the circumference. Next, rotate until that tooth catches your stop set at M teeth away. Repeating this using a simple air motor with the cylinder operating the stop would allow you to rotate until a tooth catches the stop. Cut and rotate until every tooth is cut.

For example, if N is 5 (for a 5 tooth gear) you could manually index and cut tooth 1,2,3,4,5 in order or you could use a stop catching a tooth M teeth away (for example 3) and cut 1, 4, 2, 5, 3. If you wanted to cut a 7 tooth gear with an M of 4 (adjust the stop screw for the cylinder and put a larger part on, offsetting the tool path for the correct radius) you would then end up cutting 1,5,2,6,3,7,4.

Am I even close?
Bryan.

[Servo Wizard]

gt3073b,

Look at the indexer disc and note that the spacing encompasses more degrees then required for a single tooth. Look at the tool path and you will see that there are multiple tooth profiles. Tooth counts with odd numbers required a special number of degrees and that accounts for the odd indexing.

Each sprocket had it's own unique tooth profile which consisted of between 64 and 128 pairs of involute curves. The reason for using the involute curves was to insure maximum tooth to sprocket contact while preventing heel scrub as the belt tooth leaves the sprocket tooth.

I realize that my text explanation is brief but when combined with the prictures it should be comprehendible.

Servo