The nipple coming loose and needing to be done up its valid, but as it is totally inside the fluid path, no air would "leak in" and reduce prime. If there is any air there it reduces prime anyway as it goes through the filter or if it could bypass through the thread tolerances. So any air that is at that point is already in the system from somewhere else, it doesn't create additional air.
You may be right. I do like your argument, and can see where it would be valid if the filter's mounting was loose - but still rigid enough when tightened to keep all the inside parts safely within the fluid bath. If your view of the fluid path happens, the only air available would be a finite amount trapped within the system by the filter change. At most, the volume of air would then be the volume of the filter cannister minus the denser parts of the filter itself.
Is that enough air to cause a cavitation problem?? I've no idea.... I'd hope not.....
The specific concern I was trying to get to in my post was with the way the filter cannister is screwed into place onto a threaded pipe nipple that in turn is threaded into the transmission housing. With that sort of double-ended mounting system, and with no way to positively lock the upper thread, there is no way a mechanical person removing a filter canister can know how much he has just loosened the mounting pipe nipple from the transmission housing itself. I was surprised at how far mine unscrewed. And wouldn't have suspected it to be so easily undone. I just happened to touch it with my finger and felt it wobble.
If I hadn't noticed it, and just reassembled with the transmission end of the nipple loosened, the filter's face seal might not seal tightly enough to prevent air from getting in - or fluid from getting out. Either situation could contribute to a problem.
If I can find a decent hydraulic diagram of the
M59 I'll double check the fluid pathway & immersed parts. Again, this is for the
M59, I don't know what kind of filter mounting system that the B series TLBs use.
Anyway, thanks for the correction.
rScotty