In that case, I would examine the connection to the reduction unit to be sure that no pre load is present from a dimensional problem that might load the engine thrust bearings. Using a stiff racing clutch on the street will fail the rear bearing in short order. The later engines have bigger bearings than the early engines.You can machine the stationary gear, and use the later bearings in early engines. But it still cannot be loaded.

 

In any case the engine thrust bearings must not see a load of any kind. The thrust bearings in the reduction unit must control the thrust loads from the prop. The old Ross unit had no such bearings and must be modified to use on a rotary. I am not familiar with the system for driving the reduction unit, but it should be possible to record the crank end thrust measurement with the unit assembled. If not, the engine bearings will fail again in short order. The bearings are exposed to a dilemma of geometry in that they have length and travel in a circle.

So one end or the other is always sliding on its mating surface. It is kind of a light duty bearing for this application. A continuous load of any kind will fail the rollers. With the engine turning at 5,500 RPM in cruise,

a loaded bearing is life limited. Synthetic oil will help but not for long. I suspect there is a set up problem.

 

Lynn E. Hanover

 

 

 

On further review I need to mention that examining the thick spacer and
its "geared wheel" appearance..from welding to the bearing.

The appearance is uniform all the way around, so maybe it wasn't pinched
after all but failed nonetheless. Or maybe I'm just grasping at straws.
either way its clear it failed. and considering installation is a known
failure point with this part..