RE: bearing loading
The Co ( basic dynamic load) is the starting point for most bearing life calculations.
It often is described as the "load" which the pampered bearing will support for 1,000,000 revolutions before spalling starts to appear ( the end of L10 statistical life ).
A machine running 24/7 at a mere 100 rpm will complete 1,000,000 revolutions in about a week.
A wheel bearing would accumulate 1,000,000 revs in less than 2000 miles.
The life is generally figured related inversely to the load^3. Double the load, the life is reduced to 1/8. Halve the load, life improves 8X.
For a few decades FAG and SKF have subscribed to a theory, based on testing, that bearing life especially at lower loads far exceeds calculated life, and in fact can have "infinite life" if the load is below the "fatigue limit" (and cleanliness and lubrication are excellent.)
In ISO 281:2007 reportedly there is a concept of a fatigue limit within its bearing life calculation method.
Some feel the concept of bearing endurance or fatigue limit is in error.
I would first aim for an infinite bearing life, using my time instead to design excellent sealing and ensuring lubrication life rather than calculating the somewhat nebulous although frequently obligatory L10 bearing life. The FAG spindle calculating program I enjoyed using for several years red flagged any bearing choices that would result in a "contact stress" that exceeded their prescribed fatigue limit.
As you found, For some machines the shaft sizes are sometimes first dictated by stiffness/deflection requirements. Machine tools are an example.
SKF says the fatigue load limit for the 22207E bearing is 9.3 kN.
FAG says generically the fatigue limit is ~ Co/8.
They list a fatigue load limit of 9.4 Kn here -
4100 lbs ~ 18 kN, well above the "endurance limit."
I guess I would not consider that 22207 bearing grossly oversized.
An FAG 2207 self aligning ball bearing (same envelope as you 22207) has a Co of 35 kN and a fatigue limit of just .56 Kn
