Transposing gears, the crazy way

I cut some pretty good gears with a spindle that was running out 15 thou. I was assuming (wrongly) the cutters were at fault when it was actually my mill needed fixing... Don't worry about it! The form cutters only cut an approximate tooth profile in an ideal world anyway.
 
Thanks. Yeah, logic dictates that the .002 is nothing compared to the deflection that's going to happen in the arbor, the unsupported gear mandrel (no tailstock) and the 10" diameter gear blank itself. The gears will be low speed. 800 RPM lathe, and threading will be a lot less than that.


Onward, trusty steed!! We have chips to make!
 
First gear. Well second. I had a failed blank... I tried to get 3.5 diameter out of a 3.42 piece, that I used for a practice run. Good thing too. It seems I have trained myself to the pattern of unlock axis, move axis, relock axis as a single two-handed process. But here a I need disengage power feed, crank table back 2 inches, move dividing head to the next index, engage power feed. Several times I followed the unlock-move-lock pattern with the table feed. Then tried to re-index only to discover the table moving. So now it's disengage power feed, stick right hand in pocket, crank table with left hand....


28-stud.jpg



It is amazing how light scratches look like gouges in the pictures. Those are file marks from knocking off the burr from the keyway broaching.
 
And..... we're done. *

we-four-gears.jpg



The four gears I need to cut the metric threads I'll actually use are made. The final was the 127


127-gear.jpg


I needed to cut a tooth every 40th hole around the 127 hole plate, for 42 laps around the dial. Since my sectors aren't big enough for the circle, and I didn't feel like counting off forty 127 times, I counted off and marked 40 three times. Then I made 3 piece of mechanics wire that spanned 39 holes. Thus you had 39 spaces for a wire, an empty hole for the index pin, another 39 wire, another empty hole, another 39 wire, another empty hole. And from there it should be 7 holes back to the start point.


dividing-head-127.jpg

I developed a pattern where I would advance the sector to the next hole and start the cut. While the cut was progressing, I would move the just-left wire to the front, leaving the empty hole and then making sure there were eight empties (one plus 7) between it and the currently trailing wire. Thus I always had a wire either side of the sector pin, and moved the wire so that next point was surrounded, if that makes sense. That way I had double references in case a wire got bumped out of place.

Next up is to figure out how to pin the double idler together and bore and install a bush.


* I'm thinking about remaking the 120 tooth. I had the mandrel drawbar lightly nipped down, which worked just fine for the smaller gears. But the larger 120 tooth put up enough cutter vibration that it worked loose. It took me a bit to realize why I had picked up more vibration / wobble in the blank... a few teeth are a bit wonky. I did remake the drawbar as a stud and nylock nut before doing the 127.
 

Attachments

  • we-four-gears.jpg
    we-four-gears.jpg
    48.6 KB · Views: 21
It looks fabulous. Now you get to try it out. Why the two smaller gears?

joe
 
The stock Southbend set includes something like 8 stud gears to be able to do all known plus a few theorized metric pitches. 26, 28, 32.....46. This is on top of the stock 24 and 48. The 120/127 with the stock 24 stud converts the 8TPI setting to the metric 3.0. From that you can reach 1.50, 0.75 easily with the first tumbler, and 6.0 with the 48 stud. I need to be able to do 0.80, 1.00, 1.25, 1.50, 1.75, and 2.75. I did a chart of all the combinations of stud gear and tumbler settings, and found I could get to everything I need with those two added gears. I don't have the charts in front of me. As I remember, everything I wanted except one could be done with one gear in addition to the stock 24. The other gear is only to do 1.75.
 
Okay, at home with the notes. The nice thing about this lathe setup with the 120/127 is the math is really really easy. Number of teeth on the stud gear divided by the indicated TPI on the gearbox gives the metric pitch. While Southbends charts only use positions 1 and 3 on the right tumbler, there are many ways to get to most pitches. The 40 tooth is to do 1.25 (40 / 32tpi). The 28 tooth is to do 1.75 (28 / 16 tpi) 1.0, 1.5 etc can be reached with several gears. 0.8 I figure I can do with a die. And that covers pretty everything you'll see in automotive apps.
 
When you take the most obstinately prime number, and marry it to an extremely anti-prime number, being divisible by 2,3,4,and 5, you get this hybrid


127-assembled.jpg



Together they go out to bridge the communication gap[s and attempt to unify the engineering world



127-installed.jpg



And everyone lived happily ever after.



(yeah right)
 

Attachments

  • 127-assembled.jpg
    127-assembled.jpg
    66.2 KB · Views: 24
In case anyone was curious....this is a factory dual tumbler set for a 16” SB . It also has a specific banjo that goes along with the kit.
 

Attachments

  • 175EF1AE-9CE6-40F3-AEE5-CA961F089651.jpeg
    175EF1AE-9CE6-40F3-AEE5-CA961F089651.jpeg
    1.2 MB · Views: 23
Yes, thanks for that. That's the first picture I've seen of the big lathe version. It verifies what I suspected SB had done with the banjo. I"m still considering two different paths to get more stability on the banjo. And the 64 tooth.
 
Back
Top