awap (and NancyLynn)--If the poppet seat is loose in the poppet and the demand lever is actuated (either by breathing or purging), air pressure upstream of the poppet will force the seat away from the orifice and air will flow.
I've read every reply in this thread and I think you had two problems: a blockage of some sort restricting air flow to your primary reg, and the creeping IP in your first stage causing free flowing of your octo, which is working as advertised when it freeflows.
Let's take these problems in order. First, a Scubapro R190/R380 is so dirt simple it is virtually impossible to screw it up when you put it together. The Nyloc nut (a/k/a self-locking nut) has a nylon ring that grips the threads of the poppet and prevents it from unscrewing itself as the reg is used. That nut is replaced during every service. It should NEVER be re-used because it loses its self-locking properties once it comes off the male threads of the poppet. If that nut were installed too loosely, it is conceivable that it would be so loose that the demand lever would not move the poppet when pushed. Basically, it would just be rattling around moving nothing. If this were the case I'd suspect that it would show up immediately when you turn air on.
Also, if the demand lever were installed above the washer instead of below it, its tangs may not move the poppet. The washer must be between the demand lever and the nut.
Third, the p/n 21080121 housing insert (item #30 on the R-190 schematic) may be broken. I've had this happen to me on an Air 2 that I rebuilt. The protrusions on it can break and fall away, and the demand lever will not have a surface to push against as it tries to open the poppet. This, too, once it happens, should not "fix itself" at any time. Key word here is "should not..."; anything can happen. These protrusions would fall away from the demand lever and probably be blown out into the water during exhalation. I wouldn't think they would get inside the pressurized part of the reg.
It is impossible to mount the diaphragm on the wrong side of the demand lever, so you can forget that idea.
Second--the creeping IP is a first-stage problem and rebuilding the first stage should fix it. Be sure to tell your technician to look for FOD (foreign object debris--a military term) inside the low-pressure side of the first stage that could block a hose port and cause your primary second to lose air flow. I've never seen a first-stage creep that couldn't be fixed by rebuilding the stage. This shoud be done every time the stage is serviced. If it's not done, your LDS is shortchanging you and you need another LDS.
I'd love to see your reg and see if I could duplicate the malfunction and see what fixes it. If you REALLY want to get to know regulators, you should read the book "Regulator Savvy" which is available at Peter Built tools
Peter Built Co. . It's well worth the price and it taught me more about regulators than any factory service manual I've ever read.