Around where I live there's these soapbox/gravity kart races (google "carros de rolamentos") in which the wheels are built simply from ball bearings. As I was building my own, I began wondering how I could easily mount my large bearings onto my kart, and it occurred to me that an easy way to do it would be to get some smaller bearings inside the ones previously mentioned and mount the resulting pair onto the frame, like this:

More importantly, I was under the impression that this would improve the performance of the kart in terms of speed. I am not completely sure though, hence why I am asking this question.

• On one hand, I find that my older kart reaches some sort of terminal velocity when the inner and outer races of the bearings can only move so fast in relation to each other. So, the inner bearing would sort of take some load out of the larger one by rotating in the same direction, allowing the outer-outer race to rotate faster in relation to the inner-inner race (or, in the worst case scenario, I would get no gain but also no loss, unless the smaller bearing spins backwards).

• On the other hand, I asked this to a student of mechanical engineering and he said he doubted I would get any gain, because there would be more overall friction, which does make some sense, but does not convince me much.

Also:

• We can assume that they will be well coupled together (I can file them if needed and will probably weld them afterwards).

• I can't run any practical tests at the moment because I haven't bought the smaller bearings yet (I don't know if it's worth it)

• From my research, this kind of configuration has only been mentioned in some videos about fidget spinners, and sure enough, it seems to rotate fairly well.

Model airplane motors run at 50000 RPM with ball bearings; I don't think that your limit is due to speed.

Basically, the assembly you picture will have the friction of, and be as rugged as, just using the inner bearing. The outer will just be going along for the ride (and adding weight and expense to your assembly).

• That sounds logical. Thinking about it, if we assume an ideal inner bearing, it's as if the outer bearing's inner race was left to move freely, so it will just move at the same angular speed of the outer race as if it were stuck. Thanks for your reply. Sep 29 '19 at 21:48

You will increase the friction due to two sets of balls and lubricants with the races.

Also, this will cause more play in the system...

If it worked and gave an advantage then it would be used by the F1 teams...

• My reasoning was that even if there was more friction, the gain would be much larger than the loss. In this video, I don't see much friction (youtube.com/watch?v=wvpv_NI_N04&t=2m19s). Other possible issue could be the loss of friction against the floor, because of the "slipping" effect caused by the inner bearing. Sep 29 '19 at 19:13
• How do you “see” friction? Sep 29 '19 at 19:49
• I probably used the term too loosely. In the first two sentences, I meant whichever dissipative forces that slow down the bearing. In the last sentence I meant the adherence of the outer race to the floor, but looking back at it doesn't make much sense. Also, sorry I forgot to thank you for replying. Sep 29 '19 at 21:27