LOL, I've busy and unable to add anything, but read it all. Good stuff.
The tapered roller bearings will be just fine with the original nut and tab washer. Just think about the millions of wheel bearings fit up with a course thread and cotter pin. The beauty of the steering neck and swingarm applications is that we can preload quite heavy because of the large contract area (and very little, low speed movement). This is usefull because we can snug up the steering as needed. NO inner race spacer is required, nor would it be practical in this instance.
The use of angular contact bearings in the steering neck is an acceptable mis-application of this type, again NO inner race spacer, preload to suit by FEEL. I replace these type bearings routinely as I refurbish the thrust/support bearings on ball screws used in precision machine tools. I measure and calculate the size of multiple bearings within 2 microns and custom grind the spacers and retainers to acheive the desired preload. That level of precision isn't needed, nor practical for this usage.
My original post posed the question whether the retrofitted TRB's had shown a better resistance to "the weave", based on behavior of other bikes with too little steering neck preload. My hypothesis was that, (gasp) the sealed ball bearing design was an inexpensive, easy to assemble (the inner spacer allows it to be SLAPPED TOGETHER, the nut tightened without having to worry about bearing preload, it's all set by the spacer, just the same as all modern motorcycle wheel bearings. Cheap, easy to replace and damned near idiot-proof) and replace design, however, it was flawed (GASP!) in that it provides too little to NO preload allowing the fork to flop to the lock.
My original post was gathering info, seeing if any improvements have been observed. That's all.
I truly have enjoyed all the contributors!