OAM Fingerprint Protocol on a Plateau: Path vs Push, Z vs ℓ with Falsifiable Testing

This note introduces a compact, lab-ready protocol to read an orbital-angular-momentum (OAM) “fingerprint” alongside a spin (helicity) test that separates geometry from recoil. The method runs on a single, stable plateau (R≈1) using two meters: a camera for path (Minkowski channel) and a micro-cantilever or optical tweezer bead for push (Abraham channel). A helicity Z-sweep (LCP↔RCP) flips the push sign with no change in path, while an ℓ=0→+1→−1 step—at fixed polarization—produces a small, even-in-ℓ shift in push with no sign flip and no path change. The protocol supplies numeric acceptance lines (flatness, hysteresis, lock ratio), explicit pass/fail falsifiers (disable-L, ring-placement symmetry, power scaling), and a simple even/odd readout E,O to make replication portable across platforms. Weak loss lowers the effective containment ratio Reff without moving the geometry, tightening the link between recoil and opposition rather than refraction. The result is a practical bridge from theory to hardware: clear knobs, small data products, and falsifiable outcomes that distinguish spin-driven sign from OAM’s even weighting.