Update on the pillow mask. I'm finding it pretty frustrating troubleshooting the mouth leaks. No amount of tests and trials and scheming and planning during the day can simulate what goes on while sleeping. I know that this is nothing new.
I'm definitely going to give on my "no extra gear" statement. I'll do this gladly now, because this mask is SO comfortable, and I really like what I'm seeing in SH while using the pillows.
Last night, I decided to see if lowering my min pressure would help. I dropped it 3cm (to 9), and turned off EPR. I also discovered during afternoon testing yesterday that if I drop my mouth open and just let my tongue drop, it blocks off the back-flow. So I tried going to sleep this way. On my back, mouth open, tongue relaxed in the back of my mouth, blocking the back wash. I started nodding off almost immediately. But of course, as I'd nod off, my tongue would move a bit, my lips would close some, and there'd be the gentlest of hissing sound. And I'd wake up. It's the most frustrating thing. I've got no idea why this gentle hissing wakes me up. I let this go on for 30 minutes, and switched back to my F30 FFM, as I didn't want another 2-3 hour falling asleep period.
Here's those 30 minutes, so you can see the sort of minimal leaking that is waking me up. The areas in the flow rate chart where the breaths get dense show when I was asleep(so does the respiratory chart, but maybe only for me). I'll specifically point out the wakeups right before 21:24 and right after 21:42 - note the leak rate at this point. It's next to nothing, and the hiss woke me up...

- 30 minutes with pillows.PNG (60.32 KiB) Viewed 5169 times
And here's the entire night. Man I wish I'd stuck with the pillows longer. I woke up around 5:40, and it felt like my homemade mask liner had shifted, so I pulled it out from the side and went back to sleep. Looks like I should've checked the mask seal after doing this. I woke up with a headache(never happens), and I think it's related to the 45 minute cluster at the end of the night. I'm starting to believe that when the leaks are very large, the machine decreases pressure trying to get the leak stopped - regardless of what's going on at the time event-wise. Is this true?