Switch 1No call
Array residual
Analysis A consistent with the Beta or bootstrap null, FDR-controlled, separately for Chain S and Chain C. Analysis B's bootstrap-max consistent with the time-slide null. The |τ| ≪ τ_c class is claimed only on baselines with τ_c ≫ f_d⁻¹.
Blast radius. The hypothesis as tested by that chain. The array remains a disciplined SDR network.
Switch 2No call
Chromaticity
Uncombined L1 and L5, with uncertainty. Delay class φ₁/φ₂ ≈ 1.339 may proceed, but passing is necessary, not sufficient: troposphere, multipath, and clock error are also delay-class. Ionospheric class ≈ 0.747 returns to the null. Phase-offset class = 1 is instrumental. A mixture that matches no class is not promoted. Hardware frequency dependence is bounded by the common-clock run.
Blast radius. The candidate, not the array. This switch cannot fire on a null.
Switch 3No call
Pipeline non-recovery
Open or blind injection not recovered, including the distinctive chain pattern, or off-injection not back on the null floor, or the time-slide background disagrees with the bootstrap, or the empirical false-alarm rate misses its target, or the common-clock pair or a spot check misses its predicted null.
Blast radius. The campaign, until the pipeline is repaired. No science claim.
Switch 6No call
Simultaneity convention
If a |τ| ≪ τ_c excess moves under an independent time transfer — two-way optical, common-view versus all-in-view, or a second constellation — or does not survive substitution of the common-clock calibration, it is a clock-ensemble or hardware artifact.
Blast radius. Any non-local reading of that dataset. Load-bearing on Chain C.
Switches 4 and 5 from earlier drafts are not in this compilation. Numbering keeps Switch 6 so the field book and the repository do not fork.
Prior art, so a reviewer does not have to reconstruct it: the Holometer is co-located optical strain (Chou et al., Phys. Rev. Lett. 117, 111102, 2016; arXiv:1611.08265). This program is RF carrier phase across geographic baselines. Clock-network dark-matter searches look for transients in frequency. This one looks for stationary phase coherence after geodetic subtraction. A bound here is a different channel.
Sibling stack: the node software and the existing handset shell live in DynoGator/dslv-zpdi, package labs.dynogator.dslvzpdi. This record is the Rev 3.6 campaign book beside that array, not a second simulator.
Rev 3.6 submission packet