Hi Dan,
Apologies if someone has already got around to identifying and fixing this issue - I'm still back on page 274 of this thread, trying to catch up. But I thought I'd mention this line of thought:
Nobody so far has identified whether these dropouts using the X32 and S16's with CAT5e cable are either X32, S16 or cable related. What you really need is a way of inducing the pink noise / dropout. I would suggest using a VERY long CAT5e cable to reproduce it. I think the cable and hardware are certified for 300 feet so why not try linking several and making it 500 feet or more? That's beyond spec and hence should be unstable. Therefore you will have just replicated an unstable cable run.
If it is unstable, you will be able to hear the dropouts and record them. If they sound just like your intermittent issue, then you've identified that it's a cable issue, not the X32 or S16. What's more, a highly unstable cable might be much more inclined to fault when you do the stamp, twist and bend test.
Hope that at least makes you ponder.