26-02-2019, 03:23 PM
Tony's post very appropriate, but the the word fit may puzzle. It is a while since I pondered tolerance tables and it made me think. For others challenged and to whom it may not all be immediately obvious, a little expansion of the story may assist.
The makers tolerance of bearing bores and OD is self explanatory. (from NSK catalogue and Machinerys Handbook, both 1996, current tolerances slightly different)
The letter/numbers are from an established convention table and represent various tolerance ranges for shafts and for bores (capitals)
For particular applications bearing makers recommend particular fits, achieved by adopting shafts/bores in particular tolerance ranges to go with the bearings and their tolerances. Tony has given the tolerance ranges recommended for our applications .
To obtain the actual fits, compare the bearings with the shaft/bore figures. If Seven cranks were finished as per the old 1929 drawing, rear 0 to +1/4 thou, it is evident why so many bearings were found loose.
A tight fit expands the inner ring almost directly. Assembly can then be difficult with close tolerance (or even standard) roller bearings. C3 races are used where a very tight fit required.
On a different tack, Seven bearings are very generous compared with very much more powerful motor cycles, and normal failure by fatigue seems rare if not unknown. Modern bearings due better steels are rated much higher than in the 1920s/30s. I suspect two basic ball races (maybe just one of new) would suffice in the front.
The makers tolerance of bearing bores and OD is self explanatory. (from NSK catalogue and Machinerys Handbook, both 1996, current tolerances slightly different)
The letter/numbers are from an established convention table and represent various tolerance ranges for shafts and for bores (capitals)
For particular applications bearing makers recommend particular fits, achieved by adopting shafts/bores in particular tolerance ranges to go with the bearings and their tolerances. Tony has given the tolerance ranges recommended for our applications .
To obtain the actual fits, compare the bearings with the shaft/bore figures. If Seven cranks were finished as per the old 1929 drawing, rear 0 to +1/4 thou, it is evident why so many bearings were found loose.
A tight fit expands the inner ring almost directly. Assembly can then be difficult with close tolerance (or even standard) roller bearings. C3 races are used where a very tight fit required.
On a different tack, Seven bearings are very generous compared with very much more powerful motor cycles, and normal failure by fatigue seems rare if not unknown. Modern bearings due better steels are rated much higher than in the 1920s/30s. I suspect two basic ball races (maybe just one of new) would suffice in the front.