Posted by Clint Dixon [108.162.216.235] on Friday, January 18, 2019 at 19:29:36 :
In Reply to: oops posted by Kaegi [108.162.246.8] on Friday, January 18, 2019 at 17:48:58 :
No problem. ;^)
I have always assumed that standard races would surely be held to very tight manufacturing tolerances as far as overall length and relationship of the ramp (cone contact surface) to each end. But, I know that most bearing applications have some sort of adjustment, whether it be ramp (threads) or shims, to establish preload. So overall length of standard races may not be considered as critical as the OD, ramp angle, finish, etc. of the race.
I would not be surprised if the double races for our application would have even tighter tolerances on the overall length, and especially on the locations of the ramps in respect to the ends, and of course the concentricity of the ramps to each other.
The races we make are double races and many of the tolerances are very tight. Runout, concentricity, and roundness are always a priority. We measure datum targets on diameters of the ramps at specific depths. But overall lengths are pretty wide open as far as tolerances go. As I mentioned, these are double races for articulation purposes, not rotational purposes like pinion bearings.
If standard races are not held to as tight of tolerances as double races, then making spacers to fit between them would be a challenge and one would probably only be successful if shims were introduced into the assembly like Greg mentioned. This opens a new can of worms. Now would we not only be trying to hold spacers to microns of thickness, but also shims to the same standards.
The basic problem is that the double bearing race (whether it is a manufactured double race or two single races assembled together to serve as a double) needs to be held to an overall length to control placement of the pinion and maintain proper tooth contact with the ring gear. Our third members were not designed to accept shims to position the pinion along its axis, relying on the double bearing assembly instead. Preload of the bearing cones to the race is not an issue.
Just thinking out loud here and talking my way through it. Assuming available single races have tight enough tolerances on their lengths, manufacturing spacers would basically involve dealing only with tight tolerances on thickness as well as parallelism and surface finish. Manufacturing actual double races would add circularity, runout, concentricity, perpendicularity, taper, and ramp angle to the mix, and, probably some other things I am forgetting.
Time to call a meeting. ;^)
Junior