When measuring runout in a truing stand, you are measuring the amount of shift of the left and right shafts (sprocket shaft & gearshaft) from their centerline. When measuring runout in an assembled engine, you are measuring the affect that the centerline shift has on the end of the gearshaft (apples and oranges). In the case of the assembled engine, think of the gearshaft as a lever and the main bearing as the fulcrum. Since the main bearing is located much closer to the flywheel than the right end of the gearshaft, movement at the right end of the gearshaft will me considerably more than the movement at the flywheel.
Based on my understanding of H-D’s flywheel spec bulletin, it looks like the number H-D is interested in is the amount of centerline shift; 0.004” max. H-D offers the 0.012” gearshaft specification in order to provide a method to gauge centerline shift without removing the flywheels from the crankcase. It is my sense that flywheels returned to H-D with less than 0.004” of shift will not be paid for.
Again, gearshaft runout is used to gauge flywheel shift, not to measure it. In order to measure flywheel shift accurately, you must remove the crank assembly from the engine and spin it in a truing stand.
djkak
Exactly right, djkak.
Let me try and give an example that may make this easier to understand. Think of the head studs in the motor case. There is a tolerance for centerline of the head stud tapped mounting hole in the motor case - For example, let's say the tolerance for the centerline of the diameter of the head stud tapped hole at full depth as compared to the centerline of the head stud tapped hole at the face of the case is .002. Now, insert the head stud. If you could measure the centerline of the furthest end of the stud in relation to the centerline of the tapped hole at full depth, it may be .012. The result of .012 would be fine. It is simply the projection of the .002 error at the exit of the head stud tapped hole
projected 6 inches further away, at the far end of the threaded stud. The hole at full depth compared to the hole at the head face is .002.
Project this .002 error out to the end of the threaded stud and the error is .012.
The MOCO spec. of .003-.004 for crank runout in vee blocks or a truing stand
projects to a larger number when measured with the crank installed in the motor cases, when measuring at the gearshaft.
I am not commenting on whether any of these numbers are acceptable nor am I trying to defend the MOCO in any way. I am only trying to explain that the .003 spec as measured in a truing stand does not compare to the .012 spec when measuring in the motor case. As Djak said - apples and oranges.
(Reading what I just wrote does not convince me I made this easier to understand

)
Scott