Gents,
I believe we would be correct in assuming that Stefanl cited the CMM application to illustrate the ruggedness of silicon nitride in order to contrast it with the much more fragile ruby material, and he was correct in doing so, if a higher degree of precision over the longterm is our primary concern. Si3N4 would be a material of choice in such case, but ruby would be its sonic equal provided it has no wear, a condition that cannot be certain. In addition to the reasons why it is often selected these days for CMM styli, the Si3N4 ball is also capable of withstanding heat far above what we will ever see in a turntable, but the mere fact that silicon nitride bearings wear better than ruby bearings (which are not ruby at all, but red colored synthetic sapphire) makes it all the more attractive. A piece of grit in your bearing well would most likely spell disaster for a ruby ball, but a silicon nitride one would remain unscathed. Also, Si3N4 has a much better friction coefficient than ruby, and does not tend to develop flats spots when run on hard surfaces, yet given time ruby balls do indeed wear on such surfaces. In fact, it is a commonplace characteristic with them, that factor alone would be an important consideration in a configuration where a captive ball runs on a metal thrust plate. Also, ruby balls in turntables have been known to shatter, regardless of the low speed and heat involved where I nave never heard of such an occurrence with silicon nitrides.
These are all valid reasons, in my opinion, (and the opinions of most who have seriously studied such materials) to opt for Si3N4 over steel or ruby sapphire.
Regards,
Win