how thin is too thin?
 

I'm designing a gear to go on the Chiaphua shaft. Specifically, it is a 17 tooth 32 pitch with a length through bore of maybe .75" or so. It will have a standard 2mm keyway in it.

This size of gear, with the .315" bore to go on the Chia and then the 2mm keyway leaves only about .028" between the top of the keyway and the bottom of the "valley" between teeth. Is this too thin?

Re: how thin is too thin?
 

Might depend on the material you cut the gear from, and the orientation of the keyway w.r.t. teeth. Stress will concentrate near the corners of the keyway. Since your tooth pitch along the 8mm bore is 8*pi/17 ~= 1.48 mm and the keyway is 2 mm wide, you cannot avoid putting the stress in a thin area. Do you have enough axial space to include a hub on your gear?

Re: how thin is too thin?
 

Excellent thought and yes I do, but I wasn't planning on actually cutting gear teeth. I was going to make it from pinion wire or spur gear stock which would not give me a larger diameter hub. For material, it most likely would be carbon steel.

Re: how thin is too thin?
 

Compare the shear strength of the key with the tensile stength
of the cross section of the gear that you are concerned about the
thickness of, assuming generic material strength values for carbon steel.

As your second cut, you can make a setup to perform a strength
test, using a shaft that you make so you don't waste a CIM motor.
If the gear tooth breaks first, you are done.

Finally, with all the concern about set screws coming loose and all
the safety measures suggested as a backup, and the fact that we
lost our short set screws, etc..., we TIG welded our spur gears on
the end of the motor shafts this past season. Cut the shaft so it
is flush with the end of the gear, wrap the gap between the gear and
the motor with a well soaked rag, and turn your welder loose on
the problem. The welder needs to be both good and quick at getting
the job done, and needs to be equally quick with another wet rag to
cool off the metal before much heat travels into the shaft.
The keyway could be optional...

Re: how thin is too thin?
 

Too thin to do WHAT?

If you are not going to load the motors at all, then it will be just fine (I doubt that air resistance acting alone will break the gear).

If you are going to load the motor to stall, I suppose you can get by too. BUT... ...if you are going to impact load the motor, I suppose all bets are off.

The general rule of thumb I try to follow if I can is to have at least a tooth thickness below the root diameter and 1.5 of that number if I can get it.

In this case the tooth thickness is 1/(2XDP) = 1/64 = .064" You look like you are only at about 1/2 of what I WANT to see. But... ... anyone who's read the NBD paper knows that I give up my rule of thumb when pressed.

Is it possible that you can only only have a keyway going through part of the gear?

How about this? Could you have a keyed collar that you mate the gearto (weld or press for example)?

Joe J.

Re: how thin is too thin?
 

Would it work to just have a plain bore through the gear, but have a hub of diameter of .375 with 4 slots cut in it at 90 degree intervals and then use a split lock collar to clamp around that to affix it to the motor shaft? Would that hold (not slip) at motor stall? I suppose I could just try it, but I don't have a motor. I could do a little mockup with a similar shaft to the motor shaft and then use a torque wrench or something to move it while holding the clamping collar in a vise.

Re: how thin is too thin?
 

Yes, Sanddrag, you can certainly clamp a gear on the CIM shaft safely. We, and many other teams have done it. I think clamping is one of the most under used ways (in FIRST) to attach stuff to shafts. I wish I had some numbers, but from my experience, except when it comes to moving arms or other extremely high torque stuff (and maybe even then with a sufficiently large diameter shaft/collar), everything can be clamped on a FIRST robot.

I have also heard it recommended to use extra-strong loctite when clamping to improve the amount of torque you can transfer. Personally, I just make sure the metal surfaces are clean, maybe wipe it with rubbing alcohol.

Re: how thin is too thin?
 

1/64" ~= 0.0156" and 3/2 x 1/64 ~= 0.0234" so by your rule-of-thumb sanddrag has enough thickness (0.028") below the root diameter above the keyway.

Still it seems thin to me. I would not want to shock load it.

Re: how thin is too thin?
 

You are right, my mistake -- I am mildly dislexic 1/16 and 1/64 seem occupy the same spot in my head. At times I am sure that 1/16 is ~0.064 (right) and at others I am sure that it is ~0.016(wrong). The fate of 1/64 is equally confusing to me. Sometimes it comes out ~0.016 (right) sometimes it comes out ~0.064 (wrong).

In any case, I am just flat out wrong on this one on at least 2 accounts. First because I mixed up the decimal version of 1/64. Second because the formula for tooth thickness was just wrong.

Let me see...

Allow me to point out my mistake:
Pitch Diameter is Nteeth / DP
Distance around Pitch Circle is Pitch Diameter X Pi = Nteeth X Pi / DP
Tooth Thickness is Distance around Pitch Circle/ (2 X Nteeth)
= (Nteeth X Pi / DP) / (2 X Nteeth) = Pi / (2 X DP)

So...

I was off by a factor of Pi in my formula.

For 32 DP, the circular tooth thickness is then:
Pi/64 = .048"

.028 is STILL too thin for my rule of thumb but my initial answer was right for the wrong reasons.

Joe J.

Re: how thin is too thin?
 

Hm... .028" scares me slightly. The smallest gap ive run between the hub and keyway of a CIM gear is .041", thats the gap you get with a 12 tooth 20DP tooth with a 2MM keyway on the CIM shaft.

Now if its totally necesary for you to use a 17 tooth 32DP on the shaft, and are not willing to make a whole lot of spares, I would make a broach that instead of cutting a keyway, forms the key itself, so you dont need a key on the shaft. Or even better if you have an EDM, just wire burn the sucker.

I dont know if you can weld a key into the keyway to get more material in there without warping the gear, I would not trust my welding on it.

I would go with broaching, that would fix it, and it would actually not take that much time. All you do is cut as much of the pattern out with an end mill at first, then make a broach and slam it in there, of course its a lot more complicated, but you get the idea. It sounds scary, but I've done it before, its not that bad.

Re: how thin is too thin?
 

This is a great idea. We have several sources in the Detroit area that will wire burn what ever shape you want in a part for $20-30 per part.

Joe J.

Re: how thin is too thin?
 

I like this idea, but would recommend a 7/16" hub outer diameter and just two slots 180 degrees apart. This gives (7/16" - 8mm) / 2 ~= 0.062" hub wall thickness. Then the hub stress is carried by a steel thickness that is 0.062/0.048 = 1.28x the tooth thickness -- not up to Joe's 1.5 standard, but much better than you'd get using a 3/8" hub.

SDP-SI shows 17 tooth 32DP pinion wire in carbon steel, p/n A 1C 9-N32017 out of stock now, but the dimensions look right.

McMaster has nice 7/16" collars for a couple of bucks; e.g., 6435K53.

Re: how thin is too thin?
 

Press fit, press fit, press fit or Loctite, Loctite Loctite!

This is a classic case for a press fit gear. We have been press fitting a gear on the CIM shaft since 2003. If you are nervous about press fit, then use Loctite 680. You can machine a pilot diameter all the way through your gear, then counter bore with a diameter slightly larger to take the Loctite 680. The Loctite web site has all the technical information you need.

To put your mind at ease regarding a Loctite 680 joint, FANUC Robotics (the company I work for) has a paint robot called the P-200. We have more than 1,500 of these robots installed in automotive paint shops all over the world. One of the the shoulder joint gears has a Loctite 680 interface and it never fails. This robot is designed for 10 year life running 16 hours per day, six days per week. It sees shock loading daily and holds up magnificently. My point is that a Loctite 680 designed joint will get you out of all your troubles for the least cost.

-Paul

Re: how thin is too thin?
 

Is Loctite 680 a permanent solution or can it be "undone"?

Joe J.

P.S. I know the answer but I thought I would give myself a soft ball to hit out of the park ;-)

McMaster.com
Loctite 680 Data Sheet: