The clock must fit the physical plant.

Slowing the clock is not a free adjustment.

For each physical canonical pair, normalized plant coefficients satisfy αβ=t₀²/(LC). Their smallest possible common maximum is t₀/√(LC), achieved by balanced flux–charge scales. The inherited λ≤1 requirement therefore needs t₀≤min √(LC).

The timing allocation gives the other side: t₀ must be at least 400000 times the largest of output period, filter-history moment plus electronics latency, and actuator time constant. Both sides must hold on calibrated intervals.

The audited averaging-mode spacing yields a lower clock of 2048/5 seconds under this allocation. Every pair would then require LC≥4194304/25 seconds squared. A declared LC≤1 second squared envelope has no overlap. This is a restriction on the stated contract, not a universal impossibility result.

A filtered sample contains history.

The readout theorem bounds error by state speed times the absolute first age moment of the filter, plus bounded raw error times the absolute coefficient sum. Signed filter weights can have zero signed group delay and still carry nonzero history error.

Filter mode, coefficients, aperture times, startup tails, channel skew and output availability must be recorded. Typical RMS noise and an advertised bit count do not establish the required simultaneous reconstruction bound.

A revised accuracy allocation passes mathematically.

The alternative preserves the earlier conditional Gauss fault gaps and preload tolerance. It extends the valid earlier allocation; it does not replace its historical record. Actual noise, affine errors, drift, reference and driver errors must fit the remaining allowances.

The initial screen covers AD5791B, LTC2500-32 and an ADS127L21 alternative. All are candidates. Exact conditions and manufacturer links are in the audit; no part is admitted or selected for purchase.

The required identification records are explicit.

First return actual L, C, loading, canonical action and energy scales, and a nonempty clock interval. Next freeze the converter history and timing. Then provide traceable static calibration, simultaneous noise bounds, and independent actuator output and derivative records.

The actuator test must bound rₐ=(physical output derivative)/(physical pole)+output−held command on every relevant held segment, including one-sided transitions and glitches. Finite samples require a separate regularity and coverage certificate.

Release tracking, prehistory speed, range, slew and work remain required. Blinded nominal and fault runs follow only after the premises and attributable non-constructor review are recorded. No experiment has been performed.