9 contractsExplicit metric contractsCOMPLETEInspect metric contract
27 fixturesDeterministic contract fixturesCOMPLETEInspect metric contract
10 refusalsFail-closed requestsCOMPLETEInspect metric contract
SIMULATION / SA-DEV-MODEL-NORMALIZATION-069
Device model normalization workbench
Make unlike device numbers refuse comparison.
Move from planar, fin, and nanosheet geometry to Ioff normalization, gm/Id, synthetic TLM, contact-drop bookkeeping, model residuals, corner semantics, and SPICE layer separation without hiding the metric contract.
EXPLORE
Start with the observable.
Use this page when the simulation is the primary artifact. The linked lesson explains the concept; the linked lab keeps the visual replay surface available.
- Switch between planar width, fin gated perimeter, and nanosheet gated perimeter before comparing total-device and normalized Ioff.
- Inspect gm/Id only at one declared device, bias, temperature, sign, derivative rule, model family, and model version.
- Separate synthetic TLM extraction, one-pass contact drop, train and holdout residuals, named corners, Monte Carlo samples, and five SPICE evidence layers.
V137 / DEVICE CONTRACT WORKBENCH
Compare only after the denominator is visible.
Geometry, bias, extraction rule, model lineage, and evidence state travel with every number. Incompatible requests stop before ranking.
9 metric contracts27 acceptance fixtures0 network writes
GEOMETRY FIRST
Total current is not width-normalized current.
Select the declared effective-width convention before comparing planar, fin, or nanosheet teaching examples.
planar width
Denominator: entire declared device
Denominator: 1.2 um
Comparison gatePASS
Metric basis, operating condition, derivative rule, and model lineage are explicit.
RUN CARD
Conditions before conclusions.
The browser evaluates original teaching fixtures only. It does not ingest a PDK or model card, execute an arbitrary netlist, fit measurements, rank devices, establish qualification, or issue signoff.
Primary knobs
- geometry convention and dimensions
- Ioff normalization basis
- gm and Id operating point
- derivative rule
- contact resistance
- model lineage state
- corner or Monte Carlo mode
Saved outputs
- declared effective width
- total and width-normalized Ioff
- gm over Id
- comparison gate
- synthetic TLM extraction
- one-pass contact voltage drop
- train and holdout residual summaries
- variation semantics
- deterministic JSON and CSV
EXPORT
Download the public artifacts.
These exports are the supported public files for review. PDK files and restricted third-party material are not redistributed.
V137 public model and fixture
/data/semiagora-device-model-normalization-v1.json
Open file ->METHODS
How the number was made.
Methods are written for citation discipline: the extraction rule matters as much as the plotted value.
Effective width uses one declared planar, fin, or nanosheet convention; total-device and normalized currents remain separate contracts.
Keep this method attached when reusing the figure or metric.
gm/Id requires gm and Id from the same signed operating point and derivative rule; synthetic TLM uses an exact line while contact drop remains one-pass bookkeeping.
Keep this method attached when reusing the figure or metric.
Train and holdout residuals, deterministic named corners, finite Monte Carlo samples, and SPICE model, instance, numerical, metric, and validation layers remain distinct.
Keep this method attached when reusing the figure or metric.
LIMITS
What this page does not claim.
The MVP is useful because the limits are visible. These statements prevent a teaching simulation from being cited as silicon evidence.
All new values are original deterministic teaching fixtures, not measured devices, calibrated compact models, PDK coefficients, or production distributions.
Carry this boundary into any derivative note, paper draft, or slide.
Planar, fin, and nanosheet effective-width equations are declared teaching conventions and are not silently interchangeable.
Carry this boundary into any derivative note, paper draft, or slide.
Synthetic TLM and one-pass contact voltage-drop calculations do not establish current-crowding correction, transfer length, Kelvin validation, or self-consistent compact-model behavior.
Carry this boundary into any derivative note, paper draft, or slide.
Train and holdout residuals remain separate, and neither proves silicon correlation or predictive validation.
Carry this boundary into any derivative note, paper draft, or slide.
Named corners are deterministic conditions; Monte Carlo is a declared sampled distribution. No unvalidated corner-to-sigma mapping is produced.
Carry this boundary into any derivative note, paper draft, or slide.
No PDK quality claim, device or foundry ranking, qualification, disposition, tapeout, or signoff is produced.
Carry this boundary into any derivative note, paper draft, or slide.
EVIDENCE
Trace the claim back to artifacts.
The evidence route remains the catalog-level browser; this page is the simulation-level reading card.
The immutable V137 contract independently verifies 1,215 checks and 27 acceptance fixtures before implementation.
SA-DEV-MODEL-NORMALIZATION-069
Twelve link-only sources provide terminology and replacement-evidence context; no linked body, figure, table, model card, PDK file, or numeric coefficient enters the teaching model.
SA-DEV-MODEL-NORMALIZATION-069
Ten refusal states preserve geometry, normalization, condition, lineage, corner, calibration, numerical-control, ranking, and signoff boundaries.
SA-DEV-MODEL-NORMALIZATION-069
Source spine
These links anchor the public teaching model. They do not convert the replay into a qualified process recipe or signoff result.
UC Berkeley BSIM Modelsterminology, method, or replacement-evidence boundary onlyngspice Documentationterminology, method, or replacement-evidence boundary onlyDEVSIMterminology, method, or replacement-evidence boundary onlyHow To Report And Benchmark Emerging Field-Effect Transistorsterminology, method, or replacement-evidence boundary onlyMIT OCW 6.720J MOSFET Small-Signal Behavior and Current-Gain Cutoffterminology, method, or replacement-evidence boundary onlyA Device-Level Characterization Approach to Quantify the Impacts of Different Random Variation Sources in FinFET Technologyterminology, method, or replacement-evidence boundary onlyAnalysis and Decomposition of Spatial Variation in Integrated Circuit Processes and Devicesterminology, method, or replacement-evidence boundary onlyModeling the Dependency of Analog Circuit Performance Parameters on Manufacturing Process Variations With Applications in Sensitivity Analysis and Yield Predictionterminology, method, or replacement-evidence boundary onlyRandom Dopant Induced Threshold Voltage Lowering and Fluctuations in Sub-0.1 Micrometer MOSFETs: A 3-D Atomistic Simulation Studyterminology, method, or replacement-evidence boundary onlyBYU Cleanroom contact resistance and TLMTLM terminology and replacement-evidence boundary onlyVertical-slice repository snapshotmodel-to-measurement evidence-boundary context onlyVertical-slice Apache-2.0 licensemodel-to-measurement evidence-boundary context only
Public execution boundary
No server-side live ngspice execution, uploads, accounts, payments, comments, newsletter signup, or job/event submissions are active on this page.