The claim
A physical benchtop cell (transistor/README.md (the reference transistor), scale 1: exempt sources, thinning shutters, coincidence scintillators) reproduces the sampler statistics measured in simulation: ~26 decays per independent sample and convergence to the programmed Boltzmann law.
The test
Build 2 to 8 cells (~$350 each), program a small instance via aperture settings, count decays per effective sample from the measured autocorrelation, compare to simulator/results.md. This is the hardware half of Phase A.
Which part of the theory it touches
The roadmap (Phase A); /simulator; kill criterion A's hardware half.
The claim
A physical benchtop cell (transistor/README.md (the reference transistor), scale 1: exempt sources, thinning shutters, coincidence scintillators) reproduces the sampler statistics measured in simulation: ~26 decays per independent sample and convergence to the programmed Boltzmann law.
The test
Build 2 to 8 cells (~$350 each), program a small instance via aperture settings, count decays per effective sample from the measured autocorrelation, compare to simulator/results.md. This is the hardware half of Phase A.
Which part of the theory it touches
The roadmap (Phase A); /simulator; kill criterion A's hardware half.