A leakage theorem.
Independent witness obligations.
A conditional source-to-path bound charges source impurities, full-input image response, polynomial distortion and omitted residuals. It supplies explicit requirements for a source-controlled I/Q calibration fixture.
Charge the full input, including the primary.
For full-input spectral variation U=B+A+R, the remainder bound is MR + JU + Σ βdUd/V*d−1 + D.
The terms cover source impurity, image response, mixed polynomial distortion and the aggregate residual. Subtracting the linear primary does not remove its nonlinear distortion. Budgeting distortion from only the small tail would omit required terms.
Omitted orders, memory, startup, drift, additive content and reference errors must have their own spectral-TV bound. Sample RMS or a pointwise polynomial fit cannot substitute for it.
An exact proposed design fits the budgets.
| Source-controlled lines | 100 lines, indices100–199, ≤2.5 µV each |
|---|---|
| Primary / source-impurity TV | ≤1 V / ≤1 µV |
| Full-path requirements | Gain≤1.02, image≤1 ppm, aggregate quadratic/cubic coefficients≤1 ppm at1 V |
| Conditional outside-model TV | ≤6.021506252020 µV <10 µV |
| Conditional sampled RMS | ≤52.796506252020 µV <55 µV, including20 µV readout error |
These are unverified physical requirements. The source model already gives direct tail TV≤0.261021506253 mV. This engineered fixture checks calibration and consistency; it cannot infer the native dynamics of an unknown UD source.
Exact fractions and16 controls ↓Arithmetic replay ↓The remaining work is physical evidence.
The proposed two-component I/Q voltage demonstrator needs independently justified source reconstruction, impurity, gain, image, distortion, residual, readout, timing and same-run transfer witnesses. Direct baseband testing cannot certify an omitted carrier-conversion stage.
Finite sweeps can falsify a model; they do not alone prove global gain or spectral-TV bounds. Generic noise-density specifications do not satisfy this deterministic contract.
First identify a defensible source/reconstruction model and methods for the1 µV impurity and2 µV aggregate residual-TV requirements. If those cannot be supported, the design stays HOLD even with ideal-looking samples.
Twelve missing witness categories ↓Scope and next step →Prior receipt graph →Prior sampled-monitor theorem →