I've been away, except for a little comic relief over an AD, but not idle. Since Bearfoil appears to be digging up reference material day and night into the wee hours, I'll post an outline of what might be useful things to look into. Turning serious for a moment:
1. I've been led astray by the term "rigid coupling", which may have first cropped up in the official communiques. At least one of these so-called splined couplings-- it would have to be the one joining the long innermost shaft from the entry fan disc (LPC) to its counterpart from its driver, the LPT--is not rigid. This in American terminology would be called a "geared coupling", and is a flexible coupling which will accommodate both axial misalignment of the 2 joined shafts, and axial offsets. The term is more often seen in a self-contained separate coupling, but the function here is the same.
The clue is in the term "crown" applied to the spline tooth. It isn't a crown of the top of the tooth in cross-section, but a crown along the length of the so-called spline. It is better to think of this spline as a gear tooth lying on something akin to a spherical surface. It's like a ball joint in action, except more like a section out of the middle of an American football --and it's geared.
2. Being a set (really two sets in this coupling) of gears, continuous lubrication is essential into these "splines".
3. The "gears" rotate back and forward through mesh as the shaft turns. The teeth can have other geometry overlaid, such as being made slightly helical (I think, and don't yet know about substantially helical, although I would think yes to this also).
4. It's possible the other splined couplings are more rigid.
5. Ball and roller bearings need continuous lubrication. They should never be turned when dry, as this will lead to rapid, although not necessarily instantaneous, failure. (Explanation later)
6. Ball thrust bearings may be available with opposing sides open on the inner and outer races, such that thrust is resisted in only one direction. This would be the major thrust. Minor transient thrust might be resisted by a small plain flat bearing elsewhere. The arrangement might faciliate assembling/disassembling the engine.
7. Inner races need to be clamped to their shafts, otherwise the shaft will turn inside the inner race. This is because smooth bearings have less friction than roller or ball bearings. The effect is less seen on outer races. This suggests hand holes; it maybe possible to preinstall some races with a press fit. Press fits can alter bearing clearances-- well, they will, so this must be designed for.
8. Tapered roller bearings also exist, for thrust. Spherical outer race outer surfaces exist (don't see any application of these here yet).
Footnote: I used to specify geared couplings. Odd it took me so long to realize that that's what was on the innermost, longest shaft. Any coupling covered by the AD mentioning a "crowned spline" is such a coupling.
Have at it... I'm going to key in "geared coupling" to the search engine just for fun, now. My remarks are mostly from my own experience, although I admit that until seeing pictures of these coupling teeth ground to powder
, I had little interest in the tooth forms of geared couplings.
OE