I'm also figuring out how the bolt face/ejector are going to function (still). I had planned on a 180deg opposed blade ejector and claw, but the claw portion is difficult to realize with as short a bolt head as I have (unless I go to a Remington style circlip extractor, which I don't care for).
I'm actually thinking about a shell holder face at this time, since that takes up no more room than the bolt face counterbore (might be easier for me to make with my tooling options, as well). The main downside with a shell-holder setup, is that the ejector must sweep across the bolt face; typically a complexity-inducing feature.
Here's what I'm thinking;
-Bolt face slotted and undercutted for the case rim, but just above the centerline of the firing pin hole, the slot opens up (so the shell doesn't need to be kicked upward very far to be freed.
-A long, L-shaped lever will extend from the rear of the magwell (just in front of the retracted bolt) to the rear of the receiver. It will be able to pivot vertically a small amount at its rear end, and the short leg will extend upward at the front (to punt the spent casing was the lever pivots upward)
-At rear-most travel, the bolt will contact a surface on the lever's tail, driving its front upward into the case
-The lever has a return spring driving it back down (so it won't hook the bolt in its upward position)
-The lever can pivot further down than required to clear the bolt, and doing so will cause its tail to snag the bolt; however, the front end will rotate down into the magazine lips where the next round will be. Thus, it will act as a rate reducer/bolt hold open, only allowing the bolt to drop if a round is present in the feed lips (the bolt only reciprocates 1/8" past the magwell, so the rate reducer component is probably important for reliability)
-There will probably be a manual release for the bolt catch on the underside of the rear mag latch (the actual mag release is on the side where a thumb or trigger finger can get at it)
-If possible, it'd be really cool to see if sliding or pivoting the lever side to side slightly will direct the ejection path, as that would be an easy way to get that automatic ejection direction feature I've wanted
Ideally this gizmo would end up working much like the HK pivoting ejector blade, only kicking the cartridge up instead of forward, and would have to 'reset' before the bolt could return forward.
TCB
PS; You're not French, are you, Nom? (see sig)