-
a loose cup and cone bearing allows for fairly slack manufacturing tolerances
Not really true. You still have to press the cup into the hub shell the same as you'd press the outer race of a cartridge radial bearing into the shell. There's no particular reason to believe that poor control of the size and alignment of the bore is more or less of a problem with a cup than it is with a cartridge bearing. You can make your hubs lovely or horrible by either approach.
the trade off is the higher linear friction than a radial bearing because of the axially oriented contact angle
In practice, a pair of radial cartridge bearings preloaded against one another act as a pair of angular contact bearings. The only way to get radial bearings to support purely radial loading is to let them float on the shaft (or in the housing) and control axial position with thrust bearings.
You're then just arguing about whether a general purpose industrial radial bearing forced against it's will to act as an angular contact bearing has a contact angle better optimised for the ratio of axial to radial loads than an angular contact bearing designed from scratch to be a hub bearing.The big argument in favour of cup and cone in the context of a bicycle hub is that, within the given space constraint, a cup and cone bearing can have larger balls (for lower contact pressure and longer life) than a radial cartridge bearing. To get the 9×¼" found in classical rear hubs, you'd be looking at a 6004, which at 42mm OD is going to need a pretty big negative shim to fit inside the 35mm diameter of a sprocket thread or HG spline freehub rotor 🙂
A cup and cone bearing is a type of angular contact bearing, so smiliar wheels built with these hubs instead of radial bearings hubs will be stiffer and more predictable when cornering. But the trade off is the higher linear friction than a radial bearing because of the axially oriented contact angle.
Also just saying .. a loose cup and cone bearing allows for fairly slack manufacturing tolerances ..