Wood Thrust Bearing - Marbles & Ball Bearings

I wanted to make a flat thrust bearing (similar to a lazy susan) using marbles. The generic 9/16" marbles I had were white, yellow, green, blue, black and red in color. Each color marble was a slightly different diameter and none of them were exactly round.

The track was routed into the wood using a 5/8" diameter core box bit. The white marbles are slightly over 9/16" diameter while the smallest red ones are well under 9/16" diameter.

Being slightly out-of-round or egg-shape means they will not roll uniformly around the track. This can be easily demonstrated by just pushing one of the marbles around the track and observing how quickly it stops. The largest ones stop quickly and the smallest (black, blue, and red) make it all the way around with a gentle push.

I replaced the marbles with 5/8" diameter steel ball bearings, which are perfectly round and of exact diameter. The bearing now runs quite smoothly and much quieter, but the rotation can be a bit erratic because the bearings running into each other interfere with their rotation. Looking at one bearing traveling left to right, the rear of the bearing is turning upward and the bearing directly behind it is turning downward. When they contact each other they effectively are trying to zero out the rotation between them.

Spacing the bearings evenly around the track and preventing them from running into each other is accomplished by using a cage. I made the cage from tempered hard-board. The holes in the cage are 11/16" diameter.

Then I decided to try the bearing cage with the different size (color) marbles. The cage improves the bearing but the marbles are still out-of-round and run rough.

Sometimes you can get ahead of yourself in a project or change your mind about how something should work. I already cut the center out of the large wood section, and then later decided to add the bearing race.

Before I could route the bearing track, I needed to accurately relocate the center in this large wood workpiece. The following video explains a solution I came up with for relocating the center.