Using AutoLisp programming and AutoCAD to draw Gears

[RichD]

Has anyone out there tried to automate gear drawing in 3D? I've been toying with a few routines to draw:

• Spur gears (internal and external)
• Helical Gears
• Miter gears

The basic idea so far has been utilizing the concept of "straight tooth" hobbing to generate the tooth profile...

• Draw the hobs using a rack pattern and extrude into a solid
  • (for helical gears I extruded at a 20degree angle in the "Z" axis)
  • Make the "Z" values way thicker than you'd ever need
  • center the tip of the middle tooth on the 0,0,0 x.y.z point and then save the hob drawing (it gets inserted into the drawing file as the teeth are being generated)
• For the miter gears your solid hob gets extruded with the correct taper i.e. 45degrees (I haven't done this yet, but I'm sure it's not a problem)
• All hobs are drawn at 1 diametral pitch, then scaled down automatically depending on the user input for a particular gear
• A solid cylinder is drawn based on user input...no of teeth, DP, shaft size, thickness of blank using 0,0,0 coordinates for the center.
• User is asked for number of passes (imaginary number of flutes of the hob) to simulate the simultaneous rotation of the blank and movement of the hob.
• Program starts by inserting the hob at the cutting depth, then subtracts the hob from the cylinder. This reveals the first "cut" and the hob disappears.
• The blank is rotated (1 tooth/no. of passes)... say 1/12th of a tooth if you chose 12 passes.
• Next insertion point for the hob is calculated to be ...1/12th of the circular pitch if you chose 12 passes. For example...On the screen you see the blank rotate and the hob appears/disappears for each "cut".
• This repeats until the gear is finished. It's about as slow as a hobbing machine, but it's cool to see the teeth being formed as the blank rotates. You can also run the program in a "perspective" view to see it in 3D if you want.

I'll make the "lsp" files and hob drawings available when I get it all debugged. I'm using AutoCAD release 14. I thought some of you CAD guys might find it useful.

Rich

 

[gbritnell]

Thanks Rich,
I'd be interested in the process and numbers when you get them done.
Thanks,
George

 

[RichD]

George,
Here's another screen shot.

The program works well and I've been altering the hob(s) based on research and trial and error. My focus has been centered on reducing "undercut" on low-count pinions. I have an older program I wrote that draws theoretically/numerically correct involute gears. What I do is first draw a gear with the old program, then run the "hobbing" program right on top of the first gear. Then I zoom in to study the tooth form.
Rich



Enlarged view...

 

[RichD]

It's raining here, so a good time to draw.
This picture shows a perfect gear with a hobbed gear laying on top of it. The hob used was a 1 DP 14.5 degree PA straight tooth hob with the tips filleted to .157" radius. 5 passes (or flutes) were simulated here.

This is the plan view of the previous image. A single pass without any lateral movement of the hob between indexing. Note the rougher edges on the teeth


Very Rough single pass detail...


Example of an 11 tooth gear hobbed with 25 passes...

Same gear rendered in 3D...

 

[RichD]

A 60 tooth 20 degree PA gear hobbed with 3 passes and Rendered in 3D...

Same gear zoomed in...


So I guess these examples show that with some additional passes you can produce acceptable gear tooth form using a straight hob as shown by "Hobbynut" on Youtube. You just need to move the hob a little and rotate the blank by the same amount. You just need to divide the circular pitch and the normal indexing angle by the same number. That's all my program does.

Rich

 

[Cogsy]

5 passes (or flutes) were simulated here.

This is the plan view of the previous image. A single pass without any lateral movement of the hob between indexing. Note the rougher edges on the teeth

I have made and used a 5 flute hob like this one (spinning as a cutter in a mill, which I assume you have simulated here) but I'm confused with this picture. You have the second last flute doing very little and the last flute is cutting air. From what I've read (and it works for me) is to have the middle flute on the centerline of the blank. Then every flute cuts and the tooth form is more correct.

I make one cut then index to the next tooth. All up I index the blank the number of teeth on the gear, plus 3 to finish the final teeth on both sides. I know it's not 100% perfect, but for timing gears on IC engines they seem to work fine, and they're not even overly noisy once they bed in.

 

[RichD]

You're right Al about the hob position, but this screen shot was taken after I interrupted the program. The hob starts out as you describe but then moves along with the gear as it rotates in unison. To better describe the motion...I try to move the hob along as if it was a threaded type of hob. When you tell the program the number of flutes (or passes) to use, let's say..."4", the hob is moved laterally (circular pitch/4) and the blank is rotated (indexing angle/4). So in effect it is as if a 4 fluted threaded hob has turned 1/4th of a turn.

This has all been just an experiment to see if the notion would work. The computer doesn't care how many moves to make, but a guy running a machine would likely tire of the number of moves needed.

Now getting back to the hob position being wrong, the program returns the hob back to the starting position when the 4th (or last) flute or pass has finished. Since the program doesn't crash the hob into the work like a real machine would, it doesn't matter. When I drew the 20 degree hob I included 7 teeth instead of the 5 teeth I used on the 14.5 degree hob for kind of the same reason you mentioned. In practice though, either way gives the same result.

If I ever try this on a milling machine, I will likely index the blank normally for a full turn, then make one small move of the blank and cutter. Go back to the normal indexing interval for a full turn, then repeat until done. I have cut gears using the straight hob method, but I just used a single pass, nothing fancy. Those gears are still running fine 3 years later.

 

[RichD]

Here's a sample output...



Edge view same gear...