Device metric
Cell stability and read condition
Static SNM and read-transient runs expose bounded cell conditions with explicit model limits.
sram-cell-conditionCurated evidence path
SRAM cell to array-read evidence
Trace reproduced cell and bitline artifacts into a sense-decision question, then stop before array Vmin, coverage, yield, or qualification.
Review question
Which model and measurement handoffs are required to move from a cell replay to an array claim?Permanent path link
- Nodes
- 4
- Bridge edges
- 2
- Stop edges
- 1
- Measured causal edges
- 0
Evidence classes
Read the label before the color.
- ProxyA proxy is not a physical prediction, calibration result, measurement, or causal proof.
- DerivedA derived result inherits every input and model assumption and is not measured evidence.
- ReproducedReproduction establishes the stated run result, not measured-silicon agreement or transfer to another context.
- MeasuredA measurement supports only its declared grain and conditions; it does not automatically prove cross-domain causality.
Ordered evidence chain
Follow each handoff. Stop at the stop.
Node order and edge fields are fixed by the V123 contract. No page interaction generates, ranks, or extends this chain.
- Reproduced evidence
Evidence bridge
cell read current and load to bitline differential
- Edge ID
- sram-cell-to-bitline
- From node
- sram-cell-condition
- To node
- sram-bitline-development
- Unit or category
- voltage / time category
- Evidence class
- Reproduced
- Status
- Bridge
- Assumption
- The public runs use a generic Level-1 cell or one pre-layout SKY130 condition with explicit lumped loads and thresholds.
- Bridge condition
- Use only under the displayed netlist, model, supply, temperature, load, initial condition, and observation threshold.
- Limitation
- No extracted array, local variation, distributed column RC, half-select, supply noise, sense kickback, or silicon measurement is included.
- Measured replacement slot
- Versioned netlist and model authority with extracted column RC, cell distribution, PVT, mismatch, noise, waveform measurement, and uncertainty.
- Deep link
- Open this exact edge
Circuit behavior
Bitline development
Reproduced transient and lumped-load artifacts show differential development under stated conditions.
sram-bitline-developmentReproduced evidenceEvidence bridge
bitline differential to sense decision
- Edge ID
- sram-bitline-to-sense
- From node
- sram-bitline-development
- To node
- sram-sense-decision
- Unit or category
- voltage margin / decision-time category
- Evidence class
- Reproduced
- Status
- Bridge
- Assumption
- The bitline run and sense-decision replay use separate simplified models and do not form a characterized sense path.
- Bridge condition
- Keep threshold, load, timing reference, sense topology, offset, supply, and temperature conditions explicit.
- Limitation
- No transistor-level latch correlation, kickback, metastability tail, local mismatch, extracted routing, clock skew, or measured decision distribution is present.
- Measured replacement slot
- Matched bitline and sense-amplifier test structure with extracted parasitics, offset/mismatch distribution, noise, timing, PVT, waveform measurements, and uncertainty.
- Deep link
- Open this exact edge
System / yield proxy
Sense-decision margin
The sense artifact adds a synthetic offset and attenuation decision proxy, not a characterized array.
sram-sense-decisionProxy evidenceEvidence stop
sense margin to array Vmin or yield
- Edge ID
- sram-sense-to-array-stop
- From node
- sram-sense-decision
- To node
- sram-array-stop
- Unit or category
- unsupported array / yield category
- Evidence class
- Proxy
- Status
- Stop
- Assumption
- The sense-decision values are synthetic teaching parameters and no measured array genealogy or coverage model is present.
- Bridge condition
- Stop before compiler timing, array Vmin, test coverage, outgoing quality, yield, reliability, or qualification claims.
- Limitation
- A cell or sense replay cannot establish an array distribution, failure mechanism, shipped-product risk, or production screen.
- Measured replacement slot
- Extracted macro and test-chip evidence across PVT, mismatch, noise, assist, array topology, BIST coverage, measured Vmin and yield genealogy, uncertainty, and owner review.
- Deep link
- Open this exact edge
Evidence stop
Stop before array Vmin or yield
No baseline artifact joins extracted array RC, mismatch, sense amplifier, test coverage, and measured silicon outcomes.
sram-array-stop
Path boundary
This path does not prove causality.
The navigator presents curated associations, assumptions, evidence classes, and replacement needs. It does not prove physical or statistical causality.
A stop is a product requirement, not missing UI. The page must not fill a stopped bridge with extrapolation, AI text, a recipe, an equipment setting, or an operating recommendation.
No public path authorizes process changes, equipment control, wafer disposition, qualification, signoff, product ranking, or production release.
Return to all causal chains