CHIP DESIGN · COURSE

Independent Production ASIC Capstone

Own one deliberately small digital ASIC from product scope to an authorized production decision. Freeze measurable requirements, excluded claims, resources, rights, budgets, interfaces, risks, prototypes, and stop authority. Build one generated hardware–software contract, reproducible RTL and firmware, nonvacuous verification, DFT and tester-reconstructable patterns. Reconcile lawfully supplied PDK, libraries, SRAM, I/O, clocks, IP, tools, decks, physical databases, MMMC, power, IR/EM, DRC/LVS/ERC/antenna/fill, and final GDSII or OASIS readback. Complete package, PCB, test, firmware, lab, safety, characterization, and reliability readiness before wafer start. Control foundry intake, masks, fabrication, assembly, genealogy, incoming data, first-silicon bring-up, PVT characterization, correlation, errata, yield, limits, escape/overkill, reliability, containment, cost, and field monitoring. Finish with independent raw-artifact reconstruction and separate design-release and production-release certificates—or an explicit blocked state.

Before this course: Completed all nineteen prerequisite-ordered chip-design courses, including all seven Production Bridge courses. Real completion requires lawful exact PDK/library/macro/IP/tool/deck access; sufficient compute and independent expert review; foundry or shuttle acceptance and mask/wafer authority; package, PCB, components, assembly, ATE/test hardware and programs; calibrated safe laboratories; traceable fabricated silicon populations; characterization and reliability equipment/time; quality, inventory, customer/field, stop-ship, rollback/recall, and production-release authority. Without these resources, only the academic core can be completed.

COURSE FACTSStage, chapters, units, prerequisite, and outcome
Chapter 1

Product scope includes explicit exclusions and one exact use case

Objective: For “Product scope includes explicit exclusions and one exact use case,” which frozen inputs determine the result, what is the first independently observable claim, and which mutation proves the check is alive?

The candidate is a deliberately small digital controller or sensor-interface/control ASIC using qualified analog, SRAM, I/O, clock, ESD, and process-sensitive macros rather than learner-invented production-critical structures. This lesson uses the route “build the smallest observable case.” Begin with a hand-checkable instance before invoking automation: name the state that enters the step, the transformation that is permitted, the observation that must change, and the evidence that would falsify the claim. Choose a bounded digital-control ASIC whose requirements, external dependencies, cost, schedule, risks, interfaces, and excluded claims one learner can own end to end. Connect every abstraction back to the physical structure or executable evidence it represents, and state where that representation stops being reliable.

Product scope includes explicit exclusions and one exact use case has a reviewable contract: Scope is accepted only when user, environment, interfaces, function, performance, power, area, cost, volume, lifetime, security, test, package, software, quality, and every excluded claim are measurable and reviewable. Name the applicable scope, identities, units, conditions, exclusions, threshold, evidence source, owner, and change rule before using the result. Separate control-plane success from design evidence: a process can exit zero while consuming the wrong revision, skipping work, reusing stale output, suppressing a violation, or publishing an incomplete artifact. A project called “secure AI chip” has no bounded workload, threat model, memory, interface, power, package, volume, or certification boundary. The learner must identify the first divergence and repair the dependency, not merely rerun until a dashboard becomes green.

Scope is accepted only when user, environment, interfaces, function, performance, power, area, cost, volume, lifetime, security, test, package, software, quality, and every excluded claim are measurable and reviewable. This invariant is accepted only for the named candidate and declared environment; any changed input invalidates every dependent result until reconstruction proves otherwise.

Freeze the exact objects, conditions, units, and source evidence in the worked case “Rewrite the target as one low-speed authenticated configuration controller with fixed interfaces and explicit exclusions of AI, RF, safety, and side-channel claims.” First freeze the candidate and predict the expected observation without reading a generated summary.

Apply the stated physical or engineering model, showing each transformation and preserving values that fail, are missing, or remain outside the model. Then execute the smallest transformation, retaining raw standard output, standard error, exit status, generated files, and resource use.

Compare the derived observation with “Scope is accepted only when user, environment, interfaces, function, performance, power, area, cost, volume, lifetime, security, test, package, software, quality, and every excluded claim are measurable and reviewable.” and identify the first downstream decision invalidated by the failure boundary. Finally reconcile the observation with the invariant, inject the named failure, and verify that the expected consumer refuses the corrupted or stale state.

Rewrite the target as one low-speed authenticated configuration controller with fixed interfaces and explicit exclusions of AI, RF, safety, and side-channel claims. Before revealing the trace, predict the exact command or state transition, expected exit and artifact status, first checker that should react, and minimum safe recovery.

  1. Freeze the exact objects, conditions, units, and source evidence in the worked case “Rewrite the target as one low-speed authenticated configuration controller with fixed interfaces and explicit exclusions of AI, RF, safety, and side-channel claims.”
  2. Apply the stated physical or engineering model, showing each transformation and preserving values that fail, are missing, or remain outside the model.
  3. Compare the derived observation with “Scope is accepted only when user, environment, interfaces, function, performance, power, area, cost, volume, lifetime, security, test, package, software, quality, and every excluded claim are measurable and reviewable.” and identify the first downstream decision invalidated by the failure boundary.

Result: The bounded controller is auditable; the broad marketing label is rejected as unfinishable. Accept the result only after a clean second execution reproduces the decisive artifact and a targeted mutation fails at the predicted boundary.