Creating an internal gear for servo horns

[Huub Buis]

A rotary broach would likely do the job, and much more quickly than a shaper. The challenge for you would be making the cutter -- it would be much easier with a mill and I get the impression that you don't have one.

Didn't thought of that. I use a rotary broach (self made) on the lathe to broach hex holes. I drill the holes 2 mm deeper than the broaching depth.
My first rotary broach was a piece of stock, clamped in the tool post at 6° (using feeler gauge at the back) and than drilled/reamed an 8 mm hole using a drill/reamer in the chuck. Put a bearing ball in the hole and made a round broach. Clearance angle on the broach is 3°. That broach was performing just as well as my second broach made with real bearings.
A hex broach is easy to make but that spline broach is more complicated.

 

[JPMacG]

My experience in the radio control airplane hobby is that aluminum servo horns are available and not prohibitively expensive. Could you buy one, turn it down to the center socket and press the socket into a new piece that is made to fit your need? You could maybe Loctite and/or pin it in place.

 

[whitmore]

... broaching wouldnt work as that requires the broach to go through the entire piece, for horns this wouldnt work as they are capped so they can be screwed to the end of the servo shaft, so it would have to be done with a tool that did not require to go all the way through the piece.

In a pinch, you could use a broach to make the through-hole spline, then rivet/bolt/weld it into the final assembly.
The only clean way I know to machine a blind-hole spline is with EDM.

 

[djb-systems]

@Huub Buis @homebrewed I dont have a mill and wouldnt know where to start creating a rotary broach.

@JPMacG If the servo horns were easily available it wouldnt be an issue but these are 18T servo horns and the only place I can find them is on Aliexpress / Alibaba and after ordering 2 sets that never turned up, I am not doing that again. If they were the normal 25T I would have lots of options.

@whitmore I was looking at DIY wire EDM machines last year when I was looking for milling machines never gotten around to getting either, haha The wire EDM would have been perfect for this I think.

 

[fixit]

@fixit

Thats an idea, would have to look at my design to see if I could do that and test it for any slippage. Hmmm good thought, although I still want to try and do my own internal spline but at least your suggestion would get me going.

Shouldn't see any slippage if the set screws went into the spline on the shaft.

 

[djb-systems]

@fixit Yeah I have even thought about just removing the spline from the output shaft of the servo, but a small keyway in it and use that method, although my servos were not the cheapest in the world so dont want to destroy it.

 

[djb-systems]

Hey All

So here is what I am going to try this weekend, make my own single point broaching tool at the same size of one of the servo teeth. Then using a really simple indexing system to do the teeth one at a time. I am going to try this in Delrin to begin with just to see if I can get it to work, if I can then awesome, if I cant then well @fixit to the rescue, haha.

 

[homebrewed]

If you have a mini lathe with a 16TPI lead screw you can print a 27-tooth change gear that will index your spindle to exactly 18 degree increments. But you also will need some way to drive the lead screw (as in, a handwheel attached to the end of the LS).

The way this works: The change gears establish a ratio between spindle rotations vs. LS rotations. The start of an indexing scheme, eh? If you choose A = 27 B = 57 C = 80 and D = 40 that would cut a 16.888888.... TPI thread, but what you're really after is getting the right turns ratio, and that's what the given change gear set achieves.

To use this to cut your splines you would start with the LS crank at a known position, broach one tooth, then rotate the LS exactly ONE turn and repeat until done.

I didn't come up with this scheme to cut gears or splines, but to scribe divisions on dials. FYI printing a 47T change gear will create 100 divisions with a .0082% error. A dial with 50 divisions can be scribed with 0% error, but not with a 47T gear -- it requires a 34T gear. I haven't explored the possibility of using any allowable gear ratio yet but expect it's likely possible to get zero error for 100 divisions that way. The minimum-possible gear is 21T, maximum may only be somewhat greater than 80T (due to mechanical interference).

I have found the LS-mounted handwheel to be a useful addition to my mini lathe. I made a plinth I can use to replace the compound slide, which greatly increases the rigidity of the machine for doing parting off operations, and I use the handwheel to precisely locate the cut off tool. It also makes it much easier to smoothly turn longer pieces, compared to trying to slowly & evenly turn the apron's handwheel. Finally, it can be used to cut coarser threads, where the lead screw's mechanical advantage (MA) exceeds the MA from the spindle side of the gear chain.

This scheme should work for other lathes as well, but the actual change gear to-be-printed would depend on what gears you already have and what TPI (or mm pitch) your lead screw is.

 

[djb-systems]

Hey All

So did attempt to make my own tool, didnt turn out great but I guess its just practice, had to buy myself a bench grinder to give it a go, its quite difficult trying to make a tool with a tooth that small. However indexing the lathe was easy, found a vid on youtube where you basically print out a disc with the correct divisions, no way to lock the chuck using this method, but I was able to hold it still enough.

 

[Huub Buis]

Hey All

Its quite difficult trying to make a tool with a tooth that small.

Flats (Spline tooth) can be grind using a jig. I made a stackable jig so i can easily select different angles. The grinding disk is a cheap (€10) "diamond" disk that lasts a life time. Regrinding a parting or threading tool takes a few seconds.