Just to get this clear - you're saying that the "engage" and "disengage" mechanisms on the clutch are separate and a (apparent) fault in the engaging mechanism, as evidenced by a delay in rotor start in no way indicates the clutch's ability to disengage in the event of torque loss? I'm more than happy to accept that this may be the case - just asking.
Just as an aside, on a fixed wing aeroplane about 2 years ago that I was maintaining I had a horrendous starting problem - basically it would start when completely cold or very hot, but if the engine had run anything from about 5 minutes to 12 hours previously it wouldn't start at-all. I had to take expert advice and found that the clutch mechanism (actually a torsional shock absorber, it had no controls but was designed to jump in overload) in the propeller drive was at certain oil viscosities setting up a torsional oscillation in antiphase to the starting cycle. I learned about engineering from that !
G