@Angelo Farina hit one nail on the head with his comment about the relationship between cracking effort and ease of freeflow.
@couv hit the second and third nails in noting a case-to-case discrepancy among 109's, and the effect of breathing technique on starting freeflow.
Each 109 was a hand-welded work of art. There were clear differences from reg to reg, and over the years, assorted dents and dings in the case subtly change air flow. That is the magic in today's plastic molded regulators. The consistency in interior shape from regulator to regulator allows similarly consistent performance. Whether that shape was created by CAD design or by trial and error, when it is molded plastic, it is more consistent from item to item. And as you are all discovering with your differing results, it takes very little difference from one case to another to show marked differences in Venturi action at high flow.
Technique in initiating airflow also plays a role, in my experience. If you give the purge button a sharp whack, and initiate flow abruptly, you are much more likely to maintain freeflow than if you ease it forward. I do not know why this should be the case, but suspect it has something to do with how somewhat laminar flow is created inside the case. With abrupt opening, the instantaneous airflow may be closer to the crossover point between negative resistance to France positioning and freeflow, measured with a dynamic evaluation.
The discussion of Venturi vanes and angles all relates to how air exit via the mouthpiece is maintained. The proper angle helps maintain a somewhat laminar flow. Angles that differ from optimal create turbulence which disrupts the Venturi action creating the decreased pressure that keeps the valve open. That decrease in negative pressure (shift towards zero) allows poppet spring pressure to play a role. Once cracking effort exceeds the negative pressure induced by Venturi outflow, the valve wants to close.
Conversely, the higher the cracking effort, the more difficult it is to create enough negative pressure inside the case due to Venturi action to keep the valve open.
Now, all of this is moot unless Venturi-induced assisted valve opening is important in real-world diving. Clearly, if case geometry fault helps reduce effective cracking effort in most diving positions to near zero, it really doesn't matter near the surface.
But where I think all of this Venturi assistance might indeed matter is at depth.
@Luis H can likely help us here, but it is my prejudice that at depth (say, 100 feet/4 atm), normal peak air flows of perhaps 2 SCFM are
roughly equivalent to 8 SCFM measured at the surface. Hence, seeing what happens to valve opening resistance with dynamic flow measurements at 8 SCFM is indeed relevant.