Bearing within a bearing?

thunk fired this volley in news: snipped-for-privacy@4ax.com:

You increase any inherent 'play' in the system. Also since the 'middle race' is no longer heat-sunk to a shaft or housing, risk the possibility of heating it without any way to remove the heat.

OTOH, you've distributed the wear over at least twice as many rolling elements, and if everything were perfect the balls and races would turn more slowly than a conventional bearing. IF everything were perfect.

Lloyd

I always thought that would be like two resistors in parallel as far as friction goes?

And play = vibrations, most usually. Also, how do you securely secure an outer bearing race into an inner bearing race? You can't squeeze either. You can't heat it to red hot to interference-fit it. And glue isn't strong enough.

I would think the smaller bearing would take most of the revolutions due to its lower friction surfaces. That might allow a faster spin, but with the added vibration from play, it might result in a truly spectacular catastrophic disintegration, too. I wouldn't want my hands/eyes/body anywhere near it when that happened.

I wouldn't do it.

For purely academic purposes its not that hard. You use a bushing in between the two races. Depending on the bearing type and application a double lipped bushing so when you apply force in one direction it presses against the inner lip of the bushing, and the outer lip pushes against the inner race of the outer bearing.

For practical reasons I wonder about the application. A high speed spindle of some kind? I imagine run out would be pretty bad, but it might be able to handle more load than a single ceramic bearing to run the same speed.

Another thought based on bearing type... If they are tapered roller bearings with axial pressure the central race could be tapered one way for the outer and the other for the inner. I suppose an ACB could be designed the same way as well.

Nope. That would be true if the friction in a ball bearing were proportional to velocity. It's not -- when everything is sitting still the two shafts act like they're connected with a spring. Then, once there's enough force to get the balls rolling (as it were) there's a relatively constant friction force. There is a small element of friction that rises with velocity, but it's not usually very big.

I know the gory detail on this from stabilizing gimballed platforms that have to hang from helicopters. One of the biggest paths by which the aircraft motion gets into the platform is through the bearings, and one of the biggest tradeoffs (once you're buying the World's Most Expensive Bearings) ends up in bearing strength vs. bearing friction -- basically, the more load the bearing can bear, the more friction it has.

Early Royal Enfield motorcycles had a connecting rod bearing that was free to move both in the rod big end and on the crankshaft.