Choose a platform where coherence, avalanches, or fractional memory are expected or contested. The MPFST programme asks that you log all raw acquisitions, publish the scripts that estimate (µ, γ, H), and share the gate trajectories mℓ(t), m1, m2 so that other teams can replay your run. If your apparatus can be moved or replicated off-site, plan for a traveling calibration bundle (noise source + reference sample) so the handoff retains the same gate settings.

• Sensor chain: simultaneous readings at ≥1 kHz for the driven channel, auxiliary channels, and environmental monitors.
• Analog front-end: DC-coupled, flat phase across the capture band, with a documented transfer function for de-embedding.
• Timing: disciplined clock (GPSDO or rubidium) with <10 µs jitter and PPS capture in the log for later alignment.
• Reference path: matched dummy load / dark channel to establish the gate-closed spectrum in situ.

1. Warm up the instrument until drift < 1%/hr.
2. Run a "gate closed" control with the reference sample and log at least 10 minutes of data.
3. Switch to the active sample or field condition and log ≥1e6 samples per channel.
4. Capture environmental covariates (temperature, pressure, vibration) on the same clock.
5. Archive raw binaries + JSON metadata (sample rate, gain, filter settings, instrument serials).

Provide a repeatable three-point calibration: (1) thermal noise source with known effective temperature, (2) fractional noise board that sets γ = 1.5 ± 0.02, and (3) a synthetic avalanche trace used to verify segmentation thresholds. Include STL or PCB files so other sites can fabricate the same boards.

• Publish the notebook / CLI command that estimates (µ, γ, H) and returns the mℓ trajectory.
• Include plots for PSD fits, DFA-2 slopes, avalanche rank plots with CSN fits, and Spectral Shell Monitor outputs.
• Provide BCa intervals for each tail exponent and confirm they overlap the theoretical fractional order within error bars.

Run at least one negative-control in which the gate should stay closed (shuffled labels, inverted drive, or thermalized sample). Report the flat mℓ trajectory alongside the active run. Add a structured null (IAAFT or phase-randomized surrogate) to show that avalanche tails collapse when coherence is removed.

Once your package includes raw data, metadata, calibration bundle, and notebooks, send the Zenodo/OSF link to lab@mpfst.com with a two-paragraph summary and any safety considerations. We coordinate peer swaps so another lab can install your apparatus or rerun the protocol on a mirrored rig.