Radial backlash in bearing retention removal with requirement for hand disassembly

Hi Guys,

I have a 304 steel handle assembly that rotates on two ball bearings. The bearings are secured in place with low clearance c-clips. The handle slides over the bearings and is secured in place with a set screw. On the opposite side of the set screw, a wave-spring provides tension and removes the axial backlash.

I am straggling to remove also the radial backlash. Normally I would press-fit the bearings or use a bearing retaining compound but I have the requirement for the assembly to be disassemblable by hand, with only an Allen key for the set screw. This means that the bore interfacing with bearing's OD should be a few thou larger than the OD of the bearings. See picture below.

The handle will take a maximum load of 20 lbs (10kg) with very low RPM and used sparingly.

Would a low-strength retaining compound work? Could it be disassembled manually? What else could be done?



Thanks in advance!
Dimos

Any kind of retaining compound will DEFEAT your axial spring.Your problem seems to be,too much clearance in the handle bore.A simple solution would be .Eliminate axial spring.Add another set screw at rear of rear bearing. Use full pointes set screws. If properly located these set screws would COCK the bearings a bit and force the bearings off center to take up slack in the bore, It will also retain the handle.This is not good practice,but its only a handle.It should last for years. Or maybe keep your present design and coat the handle bore with no 2 Permatex . This stuff is gooey and might take up the slack in the bore. Good luck Edwin Dirnbeck

2nd above.

Drill and tap set screw at the outside edge of the bearing race.

Pointed set screw.

As screw comes in it shifts the bearing.

Additional set screw with soft tip in center of race for final securement



Sent from my SAMSUNG-SM-G930A using Tapatalk

Tune up your google fu and search for "tolerance rings"

One thing is that the wave spring doesn't remove backlash - it spring loads it. This means that the pointed setscrew only sets a preload on that spring and radial bearing. The the extent of the spring constant and how tight the setscrew holds the spring, you have springloaded backlash. Given that this is a handle, it may be that you just want the setup to "feel" like it has no backlash, so this may be ok.

If I'm reading what you're doing with this correctly, you want to eliminate any play the user would feel in moving the handle in use. I'd replace the wave spring with spacing washers. This would allow the setscrew to tighten everything up a little bit more. Maybe not enough, though. You could use a centering spacer on both bearings. that is, a tube with OD the same as the handle ID, with say a 45 degree bevel on the inside that bears against the very edge of the bearing. You could have one on both ends - both bearings.

Two last suggestions (which may veer into an dimension were costs are not significant, and the perfect solution is the only one considered - engineeringland). If you preload the deepgrove radial bearings they will wear and you'll end up having to adjust the thing. So switch to annular contact bearings.

The single setscrew puts force on only one point of the bearing race. Sure, it works, but not good practice. Cut ID threads in the handle and a ring with od threads. Use that instead of a setscrew. Cleaner, and gives you less backlash.

Finally, you could just use the design the bicycle pedals used for the past 100 years. A handle spindle would screw into whatever you're cranking. It would have an AC bearing race (cone) machined into it. The handle would have two cups, one at each end. Grease the cups, insert bearings, put the handle on the spindle, and thread another cone on the outboard end of the spindle, which is threaded. Adjust for preload, apply locknut. Then cap it. You could also do this with two AC cartridge bearings. But with the raw cup/cone setup, you end up with negative backlash (e.g. preload) in axial and radial dimensions.

I've had bikes 30 years that didn't need any adjustment, and still had zero play in either direction.

This is what spindle, outboard cone, and locknut (with a keyed washer, btw) look like:

edwin dirnbeck said:

Any kind of retaining compound will DEFEAT your axial spring

Click to expand...

Duh! Sure. Thanks for your input. Interesting options I will consider!

Thanks bosleyjr! Now I have to make more prototypes

Thanks Tony, good ideas!

deemoss said:

Thanks bosleyjr! Now I have to make more prototypes

Click to expand...

Hey, no problem deemoss. I'm an engineer - making more work for the machinist is my job.

BTW, the spindle above, with bearing races machined in, is an older design. Most modern designs use cartridge bearings. IIRC, with wheels they often uses deep groove radial bearings, but for pedals I THINK they use AC. In bike bottom brackets (the thing the crank spins in) there are usually offerings of both. I have angular contact bearings in my main bike.

Some of those old bearings were amazingly silky and smooth. And, per your quest, you can adjust them to zero or slightly negative (e.g. preloaded) play.

deemoss said:

Hi Guys,

I have a 304 steel handle assembly that rotates on two ball bearings. The bearings are secured in place with low clearance c-clips. The handle slides over the bearings and is secured in place with a set screw. On the opposite side of the set screw, a wave-spring provides tension and removes the axial backlash.

I am straggling to remove also the radial backlash. Normally I would press-fit the bearings or use a bearing retaining compound but I have the requirement for the assembly to be disassemblable by hand, with only an Allen key for the set screw. This means that the bore interfacing with bearing's OD should be a few thou larger than the OD of the bearings. See picture below.

The handle will take a maximum load of 20 lbs (10kg) with very low RPM and used sparingly.

Would a low-strength retaining compound work? Could it be disassembled manually? What else could be done?

View attachment 306563

Thanks in advance!
Dimos

Click to expand...

Something about the cross sectional view of your bearings makes me think you are using angular contact bearings? If so, they need to be mounted iin opposition to one another, or they will seem very loose. I can't imagine typical ball bearings placed that far apart would give you any particular amount of radial play, especially if you think you need to have 'several thousandths' of looseness to the fit.

Bore some relief clearance in the area between the bearings so that the fit doesn't need to be difficult to do on assembly. But make the actual fits about a half a thousandth oversize if you want a push fit without things rocking and wobbling around.