Okay guys, here is what I discovered at the range today:
I took a closer look at the empty cases (all of them, not just one) and what I noticed was a progression of light cratering to more severe cratering and the last two rounds showing cratering with pierced primers. So this shows a progressive increase in gas pressure, which makes sense as I progressively adjusted it upwards. So I cranked the gas block down to the lowest setting (xsn10s=score one point). Switched to a new 20 round mag (belldar=score one point), loaded it with 3 rounds instead of one to check for short stroking (ARlife4me=score points). Fired first two rounds; no failure to feed etc., but third (last shot) did not lock bolt open. Opened up GB two more clicks, and success, bolt stayed open on last round. Let my buddy shoot as I watched, and noticed brass ejects @ 3:00 o'clock - perfect!
Functioned properly there after, except for one instance of failure to lock open when just one round loaded in mag.. So I am thinking that Magpul PMAG M3 is maybe not the best choice for this 6.5 CM build. So situation is much better, but not perfect. Maybe a little experimentation with buffers as suggested might get me to perfect. On second thought, is there such a thing as a perfectly functioning M5? Thanks guys!
So in summery, I made some real rookie mistakes.
1. I Used a brand new Pmag, never tested for function. The follower did not return fully to top of lips.
2. instead of starting with 2 or 3 rounds in mag to test for short stroking first, I just put in one round to test for lock back.
3. When lock back did not happen, I assumed gas was set too low, instead of trying a tested known functioning Pmag.
4. I continued to ratchet up the gas until it was maxed out and over-gassing and probably bouncing the bolt back into battery.
5. Failed to think my way through it logically. And start over.