This loopknot is a bit cord hungry, since it requires two synthesized opposing crossing knots, adjusted in a clove configuration, while the “collar structure” (actually the returning eye leg structure, in that the collars are a consequence of the nipping component), is quite simpler, taking the form of a two opposing Us stabilizing system, as it appears in the first image of the loose knot.
The double U-shaped line appears to gear neatly with this clove based nip system, remaining perpendicular to the main axis of loading (SP, on going eye leg), while the middle strand, which is common for both Us, passes through both crossing knots, in a Carrik-like fashion, being the key component that stabilizes and adds balance and symmetry to the structure.
It appears that the structure is locked down, as the curved U-shaped strands of the WE are firmly clamped by the two opposing forces of load in this synthesized crossing knot, clove nip interlocking system, hence it is alleged to be highly jam resistant and retain its form even after high rates of strain during tensile testing.
The ring loading profile is fine, if the knot is cinched well enough, but if we exert considerable pressure on both eye legs (more than usual) at the same time, it will eventually deform. Apparently, it requires to be loaded from the SP for a rock solid, perfectly balanced symmetrical profile, which is normally the usual case.
If the returning eye leg follows the reverse line,exiting from the nipping system this time perpendicular to the SP, a similar knot is created (not illustrated), but it is less stable than the structure presented here.
So, here is the knot in the attached images along with its variant in fourth image.

