pic: Snap Rings Or Shaft Collars?

[GDG 2337]

loading photo ...

[Cory]

More info is needed on how the shaft was loaded and what caused the failure.

[FrankJ]

It does not look like fatigue to my admittedly under-trained eye. What are the details around the failure?

[sanddrag]

2024 or 7075?

[ratdude747]

Originally Posted by FrankJ

It does not look like fatigue to my admittedly under-trained eye.

Looks like it might have been... too little detail to tell for sure.

IMHO snap rings are the way to go unless what is going on the shaft already is in essense a shaft collar (like a keyed sprocket).

What WILL cause severe problems under torque is drilling a hole through a shaft and putting a bolt/cotter pin. Just ask anybody who saw the 1747 bot at Cage match 2008 before elims. We sheared a steel shaft that drove the slide lift on the manipulator right where it was drilled for a drive sprocket. had we keyed the shaft and used the gear like a shaft collar it wouldn't have failed.

[GDG 2337]

Took the broken axle shaft to a Metallurgist this morning, he described a normal cycles to failure type of break. Two cracks formed 150 degrees apart from irregularities in the snap ring groove and spread toward the center. He suggested changing alloy from 2024 to 7075 for added strength. We’ve played a lot matches in the Michigan District system, Championships and two off season events. Bot got a new set of pneumatic tires for IRI which increased load on an already weakened axle shaft.

[Brandon Holley]

One understated feature of IRI is robot lifetime. By the point the competitors get to IRI, some robots have seen 100-120+ matches and are now playing some of the hardest matches they've ever seen. Things start to fail that you don't necessarily expect and team's aren't necessarily ready for.

Thanks for sharing. Hopefully your team got something out of a rough situation.

-Brando

[JamesCH95]

If the snap ring groove was cut in a torqued or stressed part of the shaft it was bound to fail and the design should be reconsidered.

As some have said, more context is needed, and a clear picture would be nice (macro mode FTW).

[Rob Stehlik]

I would just switch to a steel shaft and call it done. I don't think the minimal weight savings you get from aluminum shafts is worth the risk. We have used aluminum shafts in the past, and while we didn't see any catastrophic failures like you show here, we bent a few of them.
In response to your original question, usually snap rings are more reliable than shaft collars, but they do weaken the shaft. You want to avoid transmitting torque past a snap ring groove if possible. Generally I like to use snap rings on the ends of the shaft to keep things on, and spacers in the middle of the shaft where it will see torque loading.

[Madison]

I am presuming, based on the description, that this is:
http://www.andymark.com/product-p/am-0396.htm

I don't recall how the Toughbox gears are arranged, but I don't believe that they transmit torque across the snap ring groove -- the short end is there for a bearing and for the encoder. It's probably difficult to find a different application for this shaft that puts torque on both sides of the groove.

[Joe Finkel]

Totally agree with Rob:steel shaft way to go. Also, when grooving for snap rings
one should refer to machinist handbook.

[trilogy2826]

This general topic has been discussed several times before. General theme: Use the right tool for the job. If you understand what material types can be used where, you can get away with a robot that has no business weighing less than 120 lbs. Not saying we have it all figured out. We still have a lot to learn, but here is what we have learned during our short time in FRC.

We just started using the FEA tools embedded in Inventor this last year and this helped shave our sheet metal thicknesses by at least 20 %. We also found a couple problems with fatigue and were able to fix it before we even manufactured a single part.

2828 converted over to 7075 AL hex shaft in 2010. We have used .375, .5, .625 and .75 for anything from gearbox shafts and cantilevered axles to arm joints and shooter axles. All of these had e-clip style snap rings as the main retaining fastener. We had one issue this year that was similar to the OPs shear at the groove, but it turned out to be that the material used was accidentally 2024 instead of 7075. After the switch, no problems.

I highly suggest purchasing from onlinemetals.com. They have a great selection, fast shipping and prices that are hard to beat.

If anything, please don't "just go to steel". In most cases it isn't necessary.

[R.C.]

Originally Posted by trilogy2826

If anything, please don't "just go to steel". In most cases it isn't necessary.

Agreed, there are very few instances where you need a steel shaft in FRC.

-RC

[IanW]

Originally Posted by trilogy2826

2828 converted over to 7075 AL hex shaft in 2010. We have used .375, .5, .625 and .75 for anything from gearbox shafts and cantilevered axles to arm joints and shooter axles.

Where do you get your .375" 7075 hex bar? Onlinemetals does not supply any in this size. Fry Steel's website says that they have it, but when I called them about a month ago, they said that they didn't have any in stock and didn't have any plans to stock it.
Thanks for your help!

[DampRobot]

One of my favorite ways to retain elements on shafts is to just use spacers stacked up between them. Yes, you usually have to manufacture them yourself, but they can make assembly and disassembly much easier. They maintain components the correct distance apart without a long alignment process (shaft collars) or weakening the shaft (retaining clips). In my designs, there is usually just a stack up of spacers and components captured between two bearings, or there is a shaft collar that you just press up against one end of the stack up.

On a side note, if you do decide to use shaft collars, use two piece ones. It can be a life saver to be able to drop a shaft collar off a shaft to get at something rather than disassemble a whole mechanism to get it off.