TRS Stack Qualification Procedure
Document ID: fab-001
Canonical Path: docs/fab/TRS_Qualification.md
Revision: 1.0.0
Status: CANONICAL
Type: Procedure
Normative: Yes
Updated: 2026-08-08
Maintainer: TriadicFrameworks
Table of Contents#
- Purpose and Scope
- Normative Language
- Qualification Overview and Gate Structure
- Prerequisites and Entry Conditions
- Instrument Requirements
- Test Coupon Requirements
- L1 Intent Layer Qualification
- L2 Sequencing Layer Qualification
- L3 Resolution Layer Qualification
- L4 Commit Layer Qualification
- Full-Stack Integration Test
- Acceptance Criteria by SC Class
- Disqualification, Remediation, and Re-Qualification
- Documentation and Sign-Off
- Integration with SCR Commissioning
- Glossary
- Related Documents
1. Purpose and Scope#
1.1 Purpose#
This document is the normative procedure for qualifying a Temporal Resolution Stack (TRS) installation across all four operator layers — L1 Intent, L2 Sequencing, L3 Resolution, and L4 Commit — in a production post-ASML fab environment. Successful completion of this procedure is the mandatory prerequisite for SCR zone commissioning, for any silicon lot entering temporal manufacturing operations, and for tapeout sign-off eligibility under the TRS-Aware PDK.
The TRS qualification procedure establishes that:
- Each individual TRS operator layer functions within its specified fidelity limits.
- The four layers function correctly as a composed stack.
- The stack's performance is consistent with the SC class of the substrate lot on which qualification is being performed.
- The qualification evidence is traceable, reproducible, and sufficient to support the downstream claims made by the SCR Specification and the TRS-Aware PDK Specification.
1.2 Scope#
This procedure applies to:
- New TRS installations at any TriadicFrameworks-aligned fab.
- TRS installations following equipment replacement at one or more layers.
- TRS installations following a major firmware update at L1, L2, L3, or L4.
- TRS installations following any Coherence-Loss Event (CMA Level 3 alert) as defined in
docs/post-ASML_era/The_SCR_Specification.md, §9.2.5. - Periodic re-qualification as specified in §13.3.
This procedure does not govern:
- Individual instrument calibration, which is governed by
docs/post-ASML_era/TCT_Protocol.mdanddocs/post-ASML_era/The_TGI_Metrology_Standard.md. - SCR zone commissioning, which follows this procedure and is governed by
docs/fab/SCR_Zone_Config.md. - SC class assignment, which is governed by
docs/materials/SC_Classification.md.
1.3 Position in the Commissioning Sequence#
SC_Classification.md TCT_Protocol.md
(lot SC class assignment) (measurement traceability)
│ │
└──────────────┬─────────────┘
│
▼
TRS_Qualification.md ← THIS DOCUMENT
(L1 → L2 → L3 → L4 sequential gate)
│
▼
SCR_Zone_Config.md
(zone commissioning)
│
▼
The_TRS-Aware_PDK_Specification.md
(PDK certification)
No SCR zone may be commissioned for production until the TRS qualification for that zone's substrate lot is signed off under this procedure.
2. Normative Language#
| Keyword | Meaning |
|---|---|
| MUST | Absolute requirement. Non-compliance constitutes a conformance failure. |
| MUST NOT | Absolute prohibition. Violation constitutes a conformance failure. |
| SHOULD | Recommended practice. Deviation requires documented justification. |
| SHOULD NOT | Discouraged practice. Adoption requires documented justification. |
| MAY | Permitted but not required. |
Requirement identifiers use the format R-TRSQ-NN, where TRSQ denotes the TRS
Qualification procedure and NN is a two-digit sequential number. Requirements are
normative obligations on any entity executing or reviewing this qualification procedure.
3. Qualification Overview and Gate Structure#
3.1 The Sequential Gate Model#
TRS qualification proceeds as a strictly sequential set of four layer-level gates, followed by a full-stack integration test. Each gate must be passed before the next gate is opened.
┌──────────────────────────────────────────────────────────────────┐
│ TRS QUALIFICATION GATE SEQUENCE │
│ │
│ [Prerequisites] → [L1 Gate] → [L2 Gate] → [L3 Gate] │
│ → [L4 Gate] │
│ → [Full-Stack] │
│ → [Sign-Off] │
└──────────────────────────────────────────────────────────────────┘
R-TRSQ-01 The four layer qualification gates MUST be executed in the order L1 → L2 → L3 → L4. Execution of a gate for layer L(n+1) before L(n) has passed is non-conformant regardless of the apparent readiness of L(n+1).
R-TRSQ-02 A gate failure at any layer MUST halt the qualification sequence. The qualification procedure cannot continue to later layers while a gate failure is unresolved. The failure disposition procedure of §13 applies.
3.2 What Each Gate Verifies#
| Gate | Layer | Functional Role | Primary Measurement |
|---|---|---|---|
| L1 | Intent | Temporal intent injection; conversion of logical design intent to coherence-referenced temporal signals | Intent signal fidelity (ISF); intent latency |
| L2 | Sequencing | Ordering of temporal operations; causal dependency graph traversal; scheduling of commit operations | Sequencing order integrity (SOI); inter-layer handoff latency |
| L3 | Resolution | Spatial and temporal apodization of commit signals; address spacing enforcement | Resolution uniformity (RU); apodization profile accuracy |
| L4 | Commit | Physical commitment of temporal addresses to substrate; CA timing; SLF acquisition | Commit jitter (CJ); AER post-commit; SLF stability |
3.3 Qualification Substrate Requirement#
R-TRSQ-03 TRS qualification MUST be performed on a substrate lot with a valid SC class
assignment issued under docs/materials/SC_Classification.md before qualification begins.
The SC class of the qualification lot determines which acceptance criteria table applies
(§12).
R-TRSQ-04 Qualification performed on a substrate lot without a current, valid SC class assignment is non-conformant and its results MUST NOT be used to commission an SCR zone.
R-TRSQ-05 The SC class of the qualification lot MUST be recorded in the Qualification Record (§14.2). If the SC class changes during the qualification procedure (due to re-classification), the gate currently in progress MUST be re-executed under the new class's acceptance criteria.
4. Prerequisites and Entry Conditions#
4.1 Pre-Qualification Checklist#
R-TRSQ-06 Before the L1 gate is opened, all of the following entry conditions MUST be satisfied and documented in the Qualification Record:
| # | Entry Condition | Verification Method |
|---|---|---|
| EC-01 | Substrate lot has a valid SC class assignment (SC-II or better) | SC Classification Record present and current |
| EC-02 | All TRS layer instruments (L1–L4) have current calibration | Instrument calibration log reviewed; no expired entries |
| EC-03 | Test coupons prepared per §6 | Coupon preparation log signed |
| EC-04 | CEC (Coupon Environment Controller) conditioned and within specification | CEC log confirms ±0.1 °C, EM isolation ≥ 80 dB |
| EC-05 | SCR clock reference is locked and stable | CCG lock indicator; phase noise measurement on file |
| EC-06 | Reference address pattern loaded in TAIS | Pattern checksum verified against master |
| EC-07 | Qualification personnel are certified for this procedure | Personnel qualification records current |
| EC-08 | Qualification Record template is open and session ID assigned | Session ID recorded in fab tracking system |
R-TRSQ-07 Entry condition EC-01 requires that the SC class be SC-II or better. A lot classified SC-III MUST NOT enter TRS qualification — it is ineligible for temporal manufacturing and no qualification data produced from it has normative standing.
4.2 Reference Patterns#
R-TRSQ-08 All four qualification gates use a common set of reference address patterns
drawn from the FSCP (Full-Spectrum Contrast Pattern) defined in
docs/post-ASML_era/TCT_Protocol.md, §5.2, augmented with layer-specific test vectors
defined in §7 through §10 of this procedure. The reference pattern set MUST be version-
controlled and checksummed; the checksum MUST match the master pattern record in the fab's
qualification management system before any gate begins.
4.3 Environmental Conditions#
R-TRSQ-09 All qualification gates MUST be executed with the CEC maintaining the following conditions throughout the gate:
| Parameter | Required Value |
|---|---|
| Temperature | 25.0 °C ± 0.1 °C |
| EM isolation | ≥ 80 dB at address-encoding carrier frequency |
| Vibration | Below substrate-class sensitivity threshold |
| Humidity | 40–60% RH, non-condensing |
R-TRSQ-10 If any CEC parameter exceeds its specification during a gate, the gate MUST be paused. The gate is resumed only after the CEC parameter returns within specification and is stable for ≥ 5 minutes. Pause events MUST be logged in the Qualification Record with timestamps.
5. Instrument Requirements#
5.1 Instrument Classes#
The following instrument classes are required for TRS qualification. Specific instrument instances MUST be identified by asset ID in the Qualification Record.
| Instrument Class | Abbreviation | Gates Used | Minimum Specification |
|---|---|---|---|
| Temporal Address Injection System | TAIS | L1, L2, L3, L4 | Injection accuracy ≤ 0.0002 normalized addr. units (1σ) |
| Address Readback System | ARS | L1, L2, L3, L4 | Readback resolution ≤ 1/4096 addr. range; noise floor ≤ 0.0001 (1σ) |
| Temporal Registration Microscope | TRM | L3, L4 | Spatial accuracy ≤ 0.5 nm (3σ); address resolution ≤ 1/1024 |
| Interface Structure Profiler | ISP | L3 | Roughness resolution ≤ 0.1 nm RMS; depth resolution ≤ 0.5 nm |
| Coherence Gradient Scanner | CGS | L3 | Depth resolution ≤ 1 nm; SC_norm calibrated to TCT bulk anchor |
| Commit Jitter Analyzer | CJA | L4 | Timing resolution ≤ 0.1% of T_c; phase noise floor ≤ −130 dBc/Hz at 100 Hz |
| SLF Stability Monitor | SSM | L4 | Frame acquisition time ≤ 1 coherence cycle; lock stability window ≥ 1000 cycles |
R-TRSQ-11 All instruments used in TRS qualification MUST have current calibration
traceable to the calibration chain defined in their respective governing protocols
(TAIS/ARS: docs/post-ASML_era/TCT_Protocol.md, §11; TRM/ISP/CGS:
docs/post-ASML_era/The_TGI_Metrology_Standard.md, §12). Instruments with lapsed
calibration MUST NOT be used; the gate is blocked until calibration is renewed.
R-TRSQ-12 The CJA (Commit Jitter Analyzer) and SSM (SLF Stability Monitor) are introduced in this procedure and are not separately defined in upstream metrology documents. Their minimum specifications are as stated in the table above. Fabs MUST characterize and document the calibration method for these instruments in their equipment qualification records before executing L4 gate measurements.
5.2 Instrument Cross-Registration#
R-TRSQ-13 All instruments used in a qualification gate MUST be co-registered to the
same die coordinate reference before gate measurements begin, per
docs/post-ASML_era/The_TGI_Metrology_Standard.md, §9.1 (R-INSTR-01 and R-INSTR-02).
Cross-instrument coordinate registration error MUST be ≤ 2 nm (3σ) and MUST be verified
by measurement of a qualified reference coupon at the start of each gate.
6. Test Coupon Requirements#
6.1 Coupon Specification#
R-TRSQ-14 TRS qualification coupons MUST satisfy all requirements of
docs/post-ASML_era/TCT_Protocol.md, §4, with the following additional constraints:
- SC class minimum: SC-II or better (established by prior SC class assignment per §4.1 EC-01). SC-I coupons are preferred where availability permits.
- Coupon count: A minimum of one coupon per gate; four coupons per full qualification run. Additional coupons SHOULD be prepared as spares in the event of coupon failure during a gate.
- Coupon traceability: Each coupon MUST be identified by a coupon ID that is traceable to the lot and wafer from which it was diced, and that lot MUST have a valid SC class assignment.
R-TRSQ-15 Coupons used in gate L1 MUST NOT be reused in gates L2, L3, or L4. Each gate uses a fresh coupon. This requirement preserves measurement independence across gates: residual committed state from one gate must not confound measurements in a subsequent gate.
6.2 Coupon Preparation Sequence#
Before each gate, the gate's designated coupon MUST be prepared in the following order:
- Surface inspection per
docs/post-ASML_era/TCT_Protocol.md, §4.3 (R-COUP-05, R-COUP-06). - Prior injection check: ARS scan at five random sites confirms no structured address state.
- Load into CEC; thermal stabilization ≥ 20 minutes at 25.0 °C ± 0.1 °C.
- TAIS calibration check per §5 of this procedure.
- ARS calibration check per §5 of this procedure.
- Log coupon ID, preparation timestamp, and CEC initial conditions in the Qualification Record.
R-TRSQ-16 A coupon that fails the prior injection check (step 2) MUST be discarded and replaced. A coupon showing structured address state from prior use is non-conformant for TRS qualification regardless of its SC class.
7. L1 Intent Layer Qualification#
7.1 Purpose#
The L1 Intent Layer is responsible for converting logical design intent — expressed as a directed acyclic graph (DAG) of temporal operations with causal dependencies — into coherence-referenced temporal signals delivered to L2. The L1 qualification gate verifies that this conversion is performed with sufficient fidelity: that the temporal signals output by L1 accurately represent the intent map and carry the correct address assignments, dependency declarations, and scheduling constraints.
7.2 L1 Reference Test Vectors#
The L1 gate uses a set of eight Intent Test Vectors (ITVs), each a fully specified intent map with known correct L1 output. ITVs are defined in the qualification management system and MUST be version-locked to the current qualification procedure revision.
| ITV | Description | Primary Test Dimension |
|---|---|---|
| ITV-01 | Single operation, no dependencies, single zone | Baseline injection fidelity |
| ITV-02 | Two operations, one dependency, single zone | Dependency declaration accuracy |
| ITV-03 | Eight operations, full binary DAG, single zone | DAG traversal fidelity |
| ITV-04 | 32 operations, random DAG, single zone | Scaling under complexity |
| ITV-05 | Two operations, cross-zone dependency | Inter-zone intent signal accuracy |
| ITV-06 | Maximum intent map size (N_slot_max × N_usable operations) | Capacity boundary behavior |
| ITV-07 | Cyclic graph (intentionally malformed) | Error detection: CGR-001 violation |
| ITV-08 | Boundary address assignments (τ = 0.001, τ = 0.999) | Address boundary handling |
R-TRSQ-17 All eight ITVs MUST be executed in the L1 gate. Partial execution of the ITV set is non-conformant. If any ITV fails, the full ITV set MUST be re-executed after remediation, not only the failing ITV.
7.3 L1 Measurement Protocol#
For each ITV:
- Load the ITV intent map into the L1 system under test.
- Command L1 to generate the corresponding coherence-referenced output signals.
- Capture the output signals using the TAIS operating in signal-capture mode.
- Compare captured output signals against the ITV reference output using the Intent Signal Comparator (ISC) function of the TAIS.
- Compute the Intent Signal Fidelity (ISF) metric for this ITV (§7.4).
- Record ITV result, ISF value, and any signal anomalies in the Qualification Record.
R-TRSQ-18 The maximum elapsed time from ITV load to captured signal readback MUST be ≤ 2 coherence cycles. Intent signals that require more than 2 cycles to produce indicate an L1 processing bottleneck that must be remediated before qualification can proceed.
7.4 Intent Signal Fidelity (ISF) Metric#
ISF quantifies the accuracy of the L1 output signal set relative to the ITV reference:
ISF = 1 − (1/N_ops) × Σ_i |τ_output,i − τ_reference,i|
Where:
N_ops = total operation count in the ITV intent map
τ_output,i = temporal address carried in the L1 output signal for operation i
τ_reference,i = temporal address specified in the ITV reference output for operation i
ISF ∈ [0, 1]; 1.0 = perfect fidelity
R-TRSQ-19 ISF MUST be computed per ITV and for the aggregate across all eight ITVs. Both the per-ITV and aggregate ISF values MUST be recorded in the Qualification Record.
7.5 L1 Dependency Declaration Accuracy#
In addition to ISF, the L1 gate verifies that dependency declarations in the L1 output signal set correctly reflect the DAG edges in the input intent map.
R-TRSQ-20 For each ITV containing dependency edges (ITV-02 through ITV-08), the Qualification System MUST verify that:
- Every declared edge in the ITV intent map produces a corresponding dependency signal in the L1 output.
- No spurious dependency signals appear in the L1 output that are not present in the intent map.
- The direction of each dependency signal (predecessor → successor) is correct.
Dependency declaration accuracy MUST be reported as a fraction: (correctly declared edges) / (total edges in ITV intent map). A value of 1.0 is required for L1 gate passage.
7.6 Error Detection Verification (ITV-07)#
R-TRSQ-21 ITV-07 presents a cyclic graph to the L1 system. The L1 system MUST detect the cycle and refuse to generate output signals, instead returning a defined error code. An L1 system that processes ITV-07 and produces output signals without error detection has failed this requirement and the L1 gate, regardless of all other metrics.
7.7 L1 Gate Acceptance Summary#
| Metric | Required Value |
|---|---|
| ISF per ITV | ≥ ISF_min per §12 for the substrate's SC class |
| ISF aggregate | ≥ ISF_min per §12 |
| Dependency declaration accuracy | 1.000 (all ITVs with dependencies) |
| ITV-07 error detection | PASS (error code returned, no output generated) |
| ITV-08 boundary address handling | PASS (no address wrap or clipping artifacts) |
| Maximum intent generation latency | ≤ 2 coherence cycles per ITV |
8. L2 Sequencing Layer Qualification#
8.1 Purpose#
The L2 Sequencing Layer receives the coherence-referenced output from L1 and produces an ordered schedule of commit operations, respecting all causal dependencies and assigning each operation to a specific coherence slot. The L2 qualification gate verifies that this scheduling is performed correctly: that the produced schedule is causally valid, slot assignments are consistent with the coherence budget, and scheduling latency is within specified bounds.
8.2 L2 Reference Test Vectors#
The L2 gate uses eight Sequencing Test Vectors (STVs), each a fully specified L1 output signal set with a known correct L2 schedule.
| STV | Description | Primary Test Dimension |
|---|---|---|
| STV-01 | Single operation, no dependencies | Baseline scheduling |
| STV-02 | Linear chain of N_usable operations | Maximum sequential depth |
| STV-03 | Fully parallel flat graph (N_slot_max operations, depth 1) | Maximum slot occupancy |
| STV-04 | Balanced binary DAG, 32 operations | Mixed depth-width scheduling |
| STV-05 | Cross-zone dependency with L_handoff = 3 cycles | Inter-zone scheduling latency |
| STV-06 | Deferred operation (predecessor blocked for 3 cycles) | Deferral handling up to D_max |
| STV-07 | Operations at slot boundaries (slots 1 and N_slots−2) | Boundary slot assignment |
| STV-08 | Over-subscribed slot (N_slot_max + 1 ops assigned to one slot) | Oversubscription detection |
R-TRSQ-22 All eight STVs MUST be executed. STV-06 verifies the deferred-slot procedure (SCR Specification §5.4); the SCAA MUST confirm that deferred operations are placed in the correct re-assigned slot without exceeding D_max.
8.3 L2 Measurement Protocol#
For each STV:
- Inject the STV signal set into the L2 system under test via the L1 output interface.
- Command L2 to produce a commit schedule.
- Capture the produced schedule: operation-to-slot assignment table, causal ordering verification flags, and scheduling latency.
- Compare against the STV reference schedule using the Schedule Comparator function.
- Compute Sequencing Order Integrity (SOI) and scheduling latency for this STV.
- Record all results in the Qualification Record.
R-TRSQ-23 Maximum scheduling latency — elapsed time from STV injection to commit
schedule output — MUST be ≤ T_graph per the definition in docs/post-ASML_era/ The_TRS-Aware_PDK_Specification.md, §11.1. Schedules delivered after T_graph produce
authorization timing violations at the L4 Commit Arbiter.
8.4 Sequencing Order Integrity (SOI) Metric#
SOI = 1 − (N_violations / N_dependencies)
Where:
N_violations = number of dependency edges in the STV where the successor
operation is assigned to a slot ≤ the predecessor's slot
N_dependencies = total dependency edges in the STV
SOI ∈ [0, 1]; 1.0 = no violations
R-TRSQ-24 SOI MUST be 1.000 for every STV. An SOI value below 1.000 means that at least one causal ordering constraint was violated by the schedule; any such violation is a disqualifying failure.
8.5 Oversubscription and Deferral Detection#
R-TRSQ-25 STV-08 presents an oversubscribed slot to L2. L2 MUST detect the oversubscription and either re-assign the excess operation to the next available slot or return a rejection error. An L2 system that produces a schedule assigning N_slot_max + 1 operations to a single slot without detection has failed this requirement.
R-TRSQ-26 STV-06 requires that L2 correctly handle an operation whose predecessor remains unresolved for 3 cycles. L2 MUST defer the dependent operation for each cycle the predecessor is unresolved and schedule it in the first available slot after resolution. Deferral count MUST be accurately tracked and MUST match the expected value in the STV reference.
8.6 L2 Gate Acceptance Summary#
| Metric | Required Value |
|---|---|
| SOI | 1.000 for all STVs |
| Scheduling latency | ≤ T_graph per PDK Specification §11.1 |
| STV-08 oversubscription detection | PASS |
| STV-06 deferral tracking accuracy | Exact match to STV reference |
| Inter-zone latency (STV-05) | L_handoff cycles match CBT value ± 1 cycle |
9. L3 Resolution Layer Qualification#
9.1 Purpose#
The L3 Resolution Layer applies the apodization envelope to each scheduled commit operation and resolves the commit signal to the spatial coordinates of the target substrate site. The L3 qualification gate verifies that apodization is applied correctly, that address spacing enforcement is active, and that the resolved signals are spatially uniform across the substrate extent.
9.2 L3 Reference Test Vectors#
| RTV | Description | Primary Test Dimension |
|---|---|---|
| RTV-01 | Single-site commit, standard address | Baseline apodization profile |
| RTV-02 | Address pair at Δτ = Δτ_eff + 0.005 (just compliant) | Address spacing enforcement |
| RTV-03 | Address pair at Δτ = Δτ_eff − 0.005 (just non-compliant) | Spacing rejection |
| RTV-04 | Seven-site cluster at maximum density (TD = RWDL) | Density limit enforcement |
| RTV-05 | Apodization profile at SLL_max boundary | Side-lobe level measurement |
| RTV-06 | Site at die center; site at die corner | Spatial uniformity check |
| RTV-07 | TGI ceiling proximity (site within 5 nm of z_c) | TPR-001 proximity rule |
| RTV-08 | Full 49-site FSCP reference pattern | End-to-end resolution fidelity |
R-TRSQ-27 All eight RTVs MUST be executed. RTV-08 uses the FSCP reference pattern from the TCT Protocol to establish end-to-end resolution fidelity — the L3 output for RTV-08 MUST produce AER values consistent with the qualification substrate's SC class at each FSCP spacing level.
9.3 L3 Measurement Protocol#
For RTV-01 through RTV-07:
- Inject the RTV schedule into L3 via the L2 output interface.
- Command L3 to apply apodization and resolve commit signals.
- Capture the resolved commit signal parameters: apodization envelope shape, side-lobe levels, spatial coordinates of resolved site, and address assignment.
- For rejection RTVs (RTV-03), verify that L3 returns a rejection code and does not produce a resolved commit signal.
- Record all parameters and pass/fail determinations in the Qualification Record.
For RTV-08:
- Inject the FSCP reference pattern schedule into L3.
- Command L3 to resolve all 49 commit signals.
- Commit all resolved signals to the gate's test coupon using the TAIS in L3-driven mode.
- Read back all committed addresses using the ARS.
- Compute AER at each FSCP spacing level.
- Compare AER values against the substrate SC class acceptance thresholds in §12.
R-TRSQ-28 The apodization envelope measured in RTV-01 and RTV-05 MUST match the W_apod and n_apod parameters from the TSPS (TRS Stack Parameter Set) for the qualification substrate's SC class, within ±5% on W_apod and ±10% on n_apod. Deviations beyond these tolerances indicate L3 misconfiguration.
9.4 Resolution Uniformity (RU) Metric#
Resolution uniformity measures the consistency of the apodization envelope across the die extent, using data from RTV-06 (center vs. corner sites):
RU = 1 − |ISF_center − ISF_corner| / ISF_min
Where:
ISF_center = intent signal fidelity at the die center site
ISF_corner = intent signal fidelity at the die corner site
ISF_min = minimum ISF threshold for the SC class (§12)
RU ∈ (−∞, 1]; values below 0 indicate corner ISF below ISF_min
R-TRSQ-29 RU MUST be ≥ 0.90 for SC-I qualification and ≥ 0.80 for SC-II qualification. An RU value below these thresholds indicates that spatial non-uniformity in the L3 resolution layer would produce position-dependent fidelity variations exceeding the acceptable margin.
9.5 Side-Lobe Level Verification#
R-TRSQ-30 The maximum side-lobe level of the apodized commit signal, measured in RTV-05, MUST be ≤ SLL_max from the TSPS for the qualification substrate's SC class. A side-lobe level exceeding SLL_max risks committing spurious partial addresses at the neighboring substrate sites, producing systematic TAOE.
9.6 TGI Proximity Rule Enforcement#
R-TRSQ-31 RTV-07 presents a commit site within 5 nm of z_c. L3 MUST apply the SC_eff-derived Δτ_eff for that site rather than the SC_bulk-derived Δτ_eff, per TPR-001 in the PDK Specification. The Qualification System MUST verify that the address spacing used for RTV-07 resolution is consistent with SC_eff, not SC_bulk.
9.7 L3 Gate Acceptance Summary#
| Metric | Required Value |
|---|---|
| RTV-01: Apodization W_apod match | Within ±5% of TSPS value |
| RTV-01: Apodization n_apod match | Within ±10% of TSPS value |
| RTV-02: Address spacing compliance | PASS (commit produced) |
| RTV-03: Address spacing rejection | PASS (reject code returned, no commit) |
| RTV-05: SLL_max compliance | SLL ≤ SLL_max from TSPS |
| RTV-06: Resolution Uniformity | ≥ 0.90 (SC-I) or ≥ 0.80 (SC-II) |
| RTV-07: TGI proximity SC_eff enforcement | PASS (SC_eff Δτ_eff applied) |
| RTV-08: AER at all FSCP spacing levels | Within SC class thresholds per §12 |
10. L4 Commit Layer Qualification#
10.1 Purpose#
The L4 Commit Layer executes the physical commitment of temporal addresses to the substrate. It contains the Temporal Commit Units (TCUs), the Commit Arbiter (CA), and the SCR clock interface. The L4 qualification gate verifies that commit operations are executed with the correct timing relative to the SCR coherence clock, that the Synchronization Lock Frame (SLF) is acquired and held stably, and that post-commit address readback confirms the committed state is within the substrate's SC class fidelity limits.
10.2 L4 Reference Test Vectors#
| CTV | Description | Primary Test Dimension |
|---|---|---|
| CTV-01 | Single commit at slot 1 | Baseline commit timing |
| CTV-02 | Commit at last usable slot (N_usable) | Late-slot timing margin |
| CTV-03 | Maximum slot occupancy (N_slot_max per slot, N_usable slots) | Full coherence cycle capacity |
| CTV-04 | Commit with predecessor acknowledgment dependency | CA dependency enforcement |
| CTV-05 | Commit to cross-zone site (ZBI handoff required) | ZBI timing and latency |
| CTV-06 | SLF acquisition from cold start | SLF lock acquisition time |
| CTV-07 | SLF hold over 1000 consecutive coherence cycles | SLF long-term stability |
| CTV-08 | Forced CCG holdover event (primary reference removed) | Holdover behavior |
R-TRSQ-32 All eight CTVs MUST be executed. CTV-06 and CTV-07 together establish SLF acquisition and stability; both MUST pass before the L4 gate is considered complete.
10.3 L4 Measurement Protocol#
Commit jitter measurement (CTV-01, CTV-02, CTV-03):
- Inject the CTV schedule into L4 via the L3 output interface.
- Trigger commit operations per the schedule.
- Measure the commit timing of each operation using the CJA (Commit Jitter Analyzer), referenced to the SCR coherence clock.
- Compute commit jitter CJ for each slot:
CJ = σ(t_commit − t_slot_boundary)
Where:
t_commit = measured commit timestamp (ns)
t_slot_boundary = expected slot boundary time from the SCR clock (ns)
σ(·) = standard deviation over N_commits at this slot
CJ ∈ [0, ∞) ns
- Record CJ per slot and aggregate CJ across all CTVs in the Qualification Record.
Post-commit readback (CTV-01 through CTV-05):
After each CTV commit sequence:
- Allow the T_auth setup window to elapse after the final commit.
- Read back all committed addresses using the ARS.
- Compute the post-commit AER:
AER_commit = N_error_addresses / N_total_addresses
- Record AER_commit and compare against the SC class threshold in §12.
SLF tests (CTV-06, CTV-07, CTV-08):
CTV-06: From a cold (unacquired) SLF state, command the L4 system to acquire SLF lock. Measure time-to-lock using the SSM. Record against the acceptance limit in §10.4.
CTV-07: With SLF locked, issue 1000 consecutive coherence cycles of nominal commit load (50% slot occupancy). Monitor SLF stability with the SSM throughout. Record any lock-loss events.
CTV-08: With SLF locked, remove the CCG primary reference to trigger holdover. Monitor the frequency accuracy of the holdover oscillator using the SSM for 100 ms. Verify CCG re-locks to the primary reference upon restoration within one SLF acquisition cycle.
10.4 Commit Arbiter Dependency Enforcement#
R-TRSQ-33 CTV-04 presents a commit operation whose predecessor has not yet been acknowledged. The CA MUST withhold authorization for the dependent operation until the predecessor acknowledgment is received. The Qualification System MUST verify that the dependent operation is not committed early and that the authorization is issued within T_auth after predecessor acknowledgment.
R-TRSQ-34 If the CA issues authorization for the dependent operation before the predecessor acknowledgment is received, the L4 gate fails immediately. This failure mode indicates a CA arbitration defect that would produce causal order violations in production.
10.5 SLF Acceptance Criteria#
| Parameter | SC-I | SC-II | Measurement |
|---|---|---|---|
| SLF acquisition time (CTV-06) | ≤ 5 coherence cycles | ≤ 8 coherence cycles | SSM time-to-lock |
| SLF lock-loss events over 1000 cycles (CTV-07) | 0 | ≤ 1 (with recovery ≤ 2 cycles) | SSM lock-loss counter |
| Holdover frequency accuracy at 100 ms (CTV-08) | ≤ ±0.5 ppm | ≤ ±1.0 ppm | SSM frequency counter |
R-TRSQ-35 Any SLF lock-loss event during CTV-07 for an SC-I qualification is an immediate L4 gate failure. For SC-II qualification, one lock-loss event is tolerated only if recovery completes within 2 coherence cycles and no commit operation was in the commit window at the moment of loss.
10.6 L4 Gate Acceptance Summary#
| Metric | SC-I Requirement | SC-II Requirement |
|---|---|---|
| Commit jitter CJ | ≤ 0.5% of T_c (1σ) | ≤ 1.0% of T_c (1σ) |
| Post-commit AER | ≤ 0.010 | ≤ 0.030 |
| CA dependency enforcement (CTV-04) | PASS | PASS |
| ZBI timing accuracy (CTV-05) | L_handoff ± 1 cycle | L_handoff ± 2 cycles |
| SLF acquisition time (CTV-06) | ≤ 5 cycles | ≤ 8 cycles |
| SLF stability over 1000 cycles (CTV-07) | 0 lock-loss events | ≤ 1 lock-loss event |
| Holdover accuracy at 100 ms (CTV-08) | ≤ ±0.5 ppm | ≤ ±1.0 ppm |
11. Full-Stack Integration Test#
11.1 Purpose#
The Full-Stack Integration Test verifies that all four TRS layers function correctly as a composed system. It exercises end-to-end flows that cross all layer boundaries in a single continuous test sequence, detecting failure modes that arise from inter-layer interaction rather than individual layer defects.
11.2 Integration Test Procedure#
R-TRSQ-36 The Full-Stack Integration Test MUST use the following sequence:
-
Load the Full-Stack Reference Intent Map (FSRIM) into L1. The FSRIM is a qualification-management-controlled intent map of 512 operations with a mix of sequential chains, parallel clusters, cross-zone dependencies, and boundary-address operations. Its checksum MUST be verified before loading.
-
Command the full TRS stack (L1 → L2 → L3 → L4) to process the FSRIM from intent to committed addresses on the integration test coupon, without any intermediate operator intervention.
-
Capture all inter-layer signals at the L1/L2, L2/L3, and L3/L4 interfaces throughout the processing sequence. These captures are used to detect inter-layer communication defects.
-
Read back all 512 committed addresses using the ARS after the full sequence completes.
-
Compute the Full-Stack AER (AER_fs) from the readback results.
-
Compute the End-to-End Latency (EE_latency) — total wall-clock time from FSRIM injection to last committed address acknowledged — and compare against the expected value derived from the TSPS sequencing parameters.
R-TRSQ-37 The Full-Stack Integration Test MUST be run three times on three separate integration test coupons. The AER_fs and EE_latency results across all three runs MUST each be within the acceptance limits of §11.3. A single run outside the limit constitutes a full-stack failure.
11.3 Full-Stack Acceptance Criteria#
| Metric | SC-I | SC-II |
|---|---|---|
| Full-Stack AER (AER_fs) | ≤ 0.008 | ≤ 0.025 |
| Run-to-run AER_fs variation (3σ) | ≤ 0.002 | ≤ 0.006 |
| EE_latency | ≤ (N_folds × T_c) + 2 × T_seq_max | ≤ (N_folds × T_c) + 3 × T_seq_max |
| Inter-layer signal anomalies | 0 detected | 0 detected |
Where N_folds is the expected number of fold cycles required to process the FSRIM at the qualification substrate's SC class and N_slot_max.
R-TRSQ-38 Inter-layer signal anomalies are defined as any deviation from the expected signal timing, ordering, or content at any inter-layer interface capture point. A detected anomaly is a Full-Stack Integration failure regardless of whether the committed AER passes.
12. Acceptance Criteria by SC Class#
12.1 Consolidated Acceptance Criteria Table — SC-I#
The following acceptance criteria apply when the qualification substrate's SC class is SC-I (SC > 0.92).
| Layer | Metric | SC-I Threshold |
|---|---|---|
| L1 | ISF per ITV | ≥ 0.995 |
| L1 | ISF aggregate | ≥ 0.997 |
| L1 | Dependency declaration accuracy | 1.000 |
| L1 | ITV-07 error detection | PASS |
| L2 | SOI | 1.000 |
| L2 | Scheduling latency | ≤ T_graph |
| L2 | STV-08 oversubscription detection | PASS |
| L3 | W_apod match | ±5% of TSPS value |
| L3 | SLL compliance | ≤ SLL_max from TSPS |
| L3 | Resolution Uniformity | ≥ 0.90 |
| L3 | RTV-08 AER at S3 | ≤ 0.005 |
| L4 | Commit jitter CJ | ≤ 0.5% of T_c (1σ) |
| L4 | Post-commit AER | ≤ 0.010 |
| L4 | SLF stability (1000 cycles) | 0 lock-loss events |
| L4 | Holdover accuracy at 100 ms | ≤ ±0.5 ppm |
| Full Stack | AER_fs | ≤ 0.008 |
| Full Stack | Inter-layer anomalies | 0 |
12.2 Consolidated Acceptance Criteria Table — SC-II#
The following acceptance criteria apply when the qualification substrate's SC class is SC-II (0.75 ≤ SC ≤ 0.92).
| Layer | Metric | SC-II Threshold |
|---|---|---|
| L1 | ISF per ITV | ≥ 0.985 |
| L1 | ISF aggregate | ≥ 0.990 |
| L1 | Dependency declaration accuracy | 1.000 |
| L1 | ITV-07 error detection | PASS |
| L2 | SOI | 1.000 |
| L2 | Scheduling latency | ≤ T_graph |
| L2 | STV-08 oversubscription detection | PASS |
| L3 | W_apod match | ±5% of TSPS value |
| L3 | SLL compliance | ≤ SLL_max from TSPS |
| L3 | Resolution Uniformity | ≥ 0.80 |
| L3 | RTV-08 AER at S3 | ≤ 0.020 |
| L4 | Commit jitter CJ | ≤ 1.0% of T_c (1σ) |
| L4 | Post-commit AER | ≤ 0.030 |
| L4 | SLF stability (1000 cycles) | ≤ 1 lock-loss event (recovery ≤ 2 cycles) |
| L4 | Holdover accuracy at 100 ms | ≤ ±1.0 ppm |
| Full Stack | AER_fs | ≤ 0.025 |
| Full Stack | Inter-layer anomalies | 0 |
12.3 Dependency Declaration and Error Detection — Class-Independent#
R-TRSQ-39 The following criteria apply regardless of SC class and carry no relaxation between SC-I and SC-II:
- L1 dependency declaration accuracy: 1.000
- L1 ITV-07 error detection: PASS
- L2 SOI: 1.000
- L2 STV-08 oversubscription detection: PASS
- L4 CTV-04 CA dependency enforcement: PASS
- Full-stack inter-layer anomaly count: 0
These are structural correctness requirements. They are not material-property-dependent and cannot be relaxed by any combination of SC class, fab waiver, or exceptional circumstance.
13. Disqualification, Remediation, and Re-Qualification#
13.1 Gate Failure Disposition#
When any gate metric fails to meet its acceptance criterion:
R-TRSQ-40 The gate is marked FAILED. The Qualification Record MUST record:
- The gate identifier (L1/L2/L3/L4/Full-Stack)
- The specific metric(s) that failed
- The measured value(s)
- The acceptance criterion that was not met
- The timestamp of the failure
R-TRSQ-41 No subsequent gate may proceed. The qualification sequence halts. All substrates associated with this qualification run MUST be placed on hold pending remediation.
R-TRSQ-42 The failure MUST be escalated to the TRS Qualification Engineer within 4 hours. The escalation MUST include the Qualification Record excerpt for the failing gate.
13.2 Remediation Categories#
Failures are categorized into three remediation tracks:
| Track | Condition | Remediation |
|---|---|---|
| Track A — Configuration | Metric fails within 50% of the acceptance limit; likely parameter misconfiguration | Adjust TSPS parameters or instrument configuration; re-run failing gate only |
| Track B — Equipment | Metric fails by more than 50% of the acceptance limit; likely equipment defect | Equipment inspection and repair; re-run failing gate and all prior gates |
| Track C — Substrate | AER metrics indicate the substrate SC class does not support the TRS at any configuration | Re-classify substrate; if SC class is confirmed insufficient, qualification is terminated |
R-TRSQ-43 Track determination MUST be made by the TRS Qualification Engineer within 24 hours of escalation receipt. Track A and B remediations proceed under an approved remediation plan. Track C termination requires SCAA sign-off and is recorded as a qualification termination, not a qualification failure.
13.3 Periodic Re-Qualification#
R-TRSQ-44 Periodic re-qualification is required at the following intervals:
| Trigger | Re-Qualification Scope |
|---|---|
| 12 months since last full qualification | Full procedure (L1 through Full-Stack) |
| Any L4 equipment replacement (TCU, CA, CCG, CDN) | L4 gate + Full-Stack only |
| Any L1, L2, or L3 firmware major version increment | Affected layer gate(s) + Full-Stack |
| Coherence-Loss Event (CMA L3 alert) | Full procedure |
| SC class change for the production lot | Affected acceptance criteria — full procedure |
| Process change affecting Tier 2 stack parameters | L3 gate + L4 gate + Full-Stack |
R-TRSQ-45 Periodic re-qualification uses the same procedure as initial qualification. There is no abbreviated re-qualification path.
14. Documentation and Sign-Off#
14.1 Qualification Record Structure#
R-TRSQ-46 A Qualification Record MUST be maintained throughout the entire procedure and MUST contain, in addition to gate-specific data recorded per §7 through §11:
| Section | Required Content |
|---|---|
| Header | Session ID, date/time, fab ID, SCR zone ID, qualification personnel IDs |
| Entry Conditions | EC-01 through EC-08 check results with dates and verifier IDs |
| Substrate Record | Lot ID, wafer ID, SC class, SC value, Classification Record session ID |
| Instrument Record | Asset IDs, calibration expiry dates, calibration check results for all instruments |
| Coupon Record | Coupon IDs per gate, preparation timestamps, surface inspection results |
| Gate Results | Per-gate: all metrics, measured values, acceptance criteria, PASS/FAIL determination |
| Deviation Log | Any departures from this procedure, with justification and assessed impact |
| Remediation Log | Track classification, plan, actions taken, and outcome for any failure |
| Final Determination | Overall PASS/FAIL; date; gating condition for SCR commissioning |
14.2 Sign-Off Authority Chain#
R-TRSQ-47 The Qualification Record MUST be signed by the following authorities before it is submitted to the SCR commissioning workflow:
| Authority | Role | Sign-Off Condition |
|---|---|---|
| TRS Qualification Engineer | Lead qualification technical authority | All gates PASS; Full-Stack PASS; Deviation Log reviewed |
| Metrology Lead | Confirms instrument calibration currency and measurement quality | Instrument records reviewed and confirmed current |
| Process Engineering Lead | Confirms substrate SC class is consistent with qualification results | Substrate record cross-checked against SC Classification Record |
| TRS Qualification Authority | Final approval; issues qualification certificate | All preceding signatures present; no open deviations without disposition |
R-TRSQ-48 All signatures MUST be cryptographic (GPG or equivalent PKI-based), tied to the signer's authenticated identity in the fab's identity management system. Plain-text name entries are not valid for production qualification records.
14.3 Qualification Certificate#
R-TRSQ-49 Upon successful completion of the procedure and all sign-offs, the TRS Qualification Authority issues a TRS Qualification Certificate that MUST include:
- Qualification Record session ID
- Substrate lot ID and SC class
- SCR zone ID for which qualification is valid
- Date of qualification
- Expiry date (12 months from date of qualification per §13.3)
- TSPS parameter set version qualified against
- TRS Qualification Authority identifier and signature
R-TRSQ-50 The certificate MUST be registered in the fab's qualification management system and MUST be accessible to the SCR zone commissioning workflow before any wafer lot enters production on the qualified zone.
14.4 Retention#
R-TRSQ-51 Qualification Records and Qualification Certificates MUST be retained for a minimum of five (5) years from the date of issuance, or for the lifetime of the process node, whichever is longer. Records from failed qualifications MUST be retained under the same policy and MUST NOT be deleted or anonymized.
15. Integration with SCR Commissioning#
15.1 Qualification as the SCR Commissioning Gate#
The TRS Qualification Certificate is the primary enabling document for SCR zone
commissioning. No SCR zone may proceed to commissioning under
docs/fab/SCR_Zone_Config.md without a valid, current certificate for the substrate lot
and zone combination.
R-TRSQ-52 The SCR commissioning workflow MUST verify that a valid Qualification Certificate exists for the target zone before opening any CCG, CDN, or CA configuration interface. Access to commissioning configuration without a valid certificate is a process violation.
15.2 Parameter Handoff to SCR Commissioning#
The following parameters from the TRS Qualification Record are consumed directly by the SCR commissioning workflow:
| Parameter | Source in Qualification Record | Consumed By |
|---|---|---|
| N_slot_max | TSPS parameter set used in qualification | CBT generation (PDK Specification §6.2) |
| T_seq_min, T_seq_max | L2 scheduling latency measurements (§8.3) | CSA arc (TTF Arc Library §7.2.1) |
| T_arb | L4 commit jitter measurement, CA processing component | APA arc (TTF Arc Library §7.2.3) |
| L_handoff | CTV-05 measured inter-zone latency | ZBA arc (TTF Arc Library §7.2.2) |
| SLF acquisition time | CTV-06 measurement | SCR zone topology sizing |
| AER_fs | Full-Stack AER | CBT provisional flag determination |
R-TRSQ-53 The SCR commissioning team MUST extract these parameters from the Qualification Record directly and MUST NOT use generic or design-time estimates in their place. Parameters derived from measurement are more accurate and provide better timing margin prediction than pre-qualification estimates.
15.3 PDK Generation Dependency#
R-TRSQ-54 The TRS-Aware PDK generation workflow requires a completed Qualification Record session ID as an input (PDK Specification §11.2, R-GEN-01). The PDK generation tool MUST reject any generation request that does not reference a current, signed Qualification Record. PDKs generated without a valid qualification reference are provisional and MUST carry the PROVISIONAL flag.
15.4 Re-Qualification Impact on Active PDK Releases#
When a re-qualification event occurs (§13.3):
R-TRSQ-55 If the re-qualification produces TSPS parameter values that differ from those used to generate the current active PDK by more than the minor-revision threshold (PDK Specification §14.2), a new PDK minor version MUST be generated using the updated qualification data before new design starts are accepted on the re-qualified zone.
R-TRSQ-56 Designs signed off under the prior PDK version continue to be valid for tape-out on the re-qualified zone, provided the re-qualification TSPS changes do not cause any timing paths to violate their margins when re-evaluated with the updated parameters. If any path's slack changes by more than 10% due to re-qualification parameter updates, the affected design MUST be re-checked.
16. Glossary#
| Term | Definition |
|---|---|
| AER | Address Error Rate — fraction of committed addresses read back incorrectly; primary fidelity metric across all TRS layers |
| AER_commit | Post-commit AER measured at the L4 gate after physical commitment to substrate |
| AER_fs | Full-Stack AER measured at the conclusion of the Full-Stack Integration Test |
| APA arc | Arbiter Processing Arc — TTF arc encoding CA processing latency T_arb; derived from L4 gate measurement |
| CA | Commit Arbiter — the SCR component that authorizes TCU commit operations within coherence slots |
| CBT | Coherence Budget Table — PDK component encoding SCR zone capacity limits; populated from TRS Qualification data |
| CCG | Coherence Clock Generator — master zone timing source for the SCR |
| CJA | Commit Jitter Analyzer — instrument class used in the L4 gate to measure commit timing deviation |
| CJ | Commit Jitter — standard deviation of commit operation timing relative to the coherence slot boundary |
| CTV | Commit Test Vector — one of eight reference test cases used in the L4 qualification gate |
| D_max | Maximum deferral count — maximum times an operation may be deferred before CA escalates |
| EE_latency | End-to-End Latency — total wall-clock time from FSRIM intent injection to last commit acknowledgment |
| FSCP | Full-Spectrum Contrast Pattern — standard TCT address pattern used in L3 and L4 gates |
| FSRIM | Full-Stack Reference Intent Map — 512-operation qualification intent map used in Full-Stack Integration Test |
| ISC | Intent Signal Comparator — TAIS function that compares captured L1 output against ITV reference |
| ISF | Intent Signal Fidelity — metric quantifying L1 output accuracy relative to the ITV reference output |
| ITV | Intent Test Vector — one of eight reference test cases used in the L1 qualification gate |
| L1 | Intent Layer — TRS operator layer converting design intent to coherence-referenced temporal signals |
| L2 | Sequencing Layer — TRS operator layer scheduling commit operations into coherence slots |
| L3 | Resolution Layer — TRS operator layer applying apodization and resolving commit signals to substrate sites |
| L4 | Commit Layer — TRS operator layer executing physical temporal address commitment to substrate |
| n_apod | Apodization roll-off exponent — controls steepness of the apodization envelope edge; specified in TSPS |
| RU | Resolution Uniformity — metric quantifying spatial consistency of L3 apodization across die extent |
| RTV | Resolution Test Vector — one of eight reference test cases used in the L3 qualification gate |
| SC | Substrate Clarity — bulk measure of a substrate's temporal coherence fidelity; source for all acceptance thresholds |
| SC class | One of three qualification tiers (SC-I, SC-II, SC-III) assigned by docs/materials/SC_Classification.md |
| SLF | Synchronization Lock Frame — the coherence cycle boundary reference acquired by the L4 layer from the SCR clock |
| SLL_max | Maximum permitted side-lobe level of an apodized commit signal; specified in TSPS |
| SOI | Sequencing Order Integrity — metric quantifying causal ordering compliance in L2 produced schedules |
| SSM | SLF Stability Monitor — instrument class used in L4 gate to measure SLF acquisition and stability |
| STV | Sequencing Test Vector — one of eight reference test cases used in the L2 qualification gate |
| SCAA | SC Class Assignment Authority — qualified entity authorized to perform SC class assignments |
| T_arb | Commit Arbiter processing latency — component of SLF derivation; source of APA arc in TTF library |
| T_auth | Authorization setup window — time before slot boundary within which the CA must assert authorization |
| T_c | Coherence cycle period — fundamental time unit of SCR operation |
| T_graph | Maximum elapsed time for L2 to deliver a commit schedule before the first slot of the target cycle |
| T_seq | TRS L2 sequencing latency — per-operation scheduling overhead; source of CSA arc in TTF library |
| TCU | Temporal Commit Unit — Tier 2 tool that executes physical temporal address commitments |
| TAIS | Temporal Address Injection System — instrument used to inject reference addresses and capture L1 output signals |
| TRS | Temporal Resolution Stack — the four-layer operator system (L1–L4) governing temporal manufacturing |
| TSPS | TRS Stack Parameter Set — PDK component encoding TRS operational parameters per SC class |
| TTF | Temporal Timing Format — standardized format for temporal timing arcs in EDA timing models |
| W_apod | Apodization window width — half-width of the L3 apodization envelope; specified in TSPS |
| ZBA arc | Zone Boundary Arc — TTF arc encoding inter-zone handoff latency; value derived from CTV-05 measurement |
| ZBI | Zone Boundary Interface — SCR component managing inter-zone temporal handoff |
17. Related Documents#
| Document | Path | Relationship |
|---|---|---|
| The Temporal Manufacturing Primer | docs/post-ASML_era/The_Temporal_Manufacturing_Primer.md |
Defines TRS stack layers L1–L4 and equipment tiers; prerequisite conceptual context |
| The SCR Specification | docs/post-ASML_era/The_SCR_Specification.md |
Normative authority for SCR zone architecture; commissioning follows this qualification |
| The TGI Metrology Standard | docs/post-ASML_era/The_TGI_Metrology_Standard.md |
Source of TRM, ISP, and CGS instrument class definitions used in §5 |
| TCT Protocol | docs/post-ASML_era/TCT_Protocol.md |
Source of TAIS/ARS/CEC specifications and FSCP reference pattern used in §6 and §9 |
| The TRS-Aware PDK Specification | docs/post-ASML_era/The_TRS-Aware_PDK_Specification.md |
Consumer of Qualification Record data for PDK generation; Section §11.2 depends on this procedure |
| The Logic Folding Architecture Guide | docs/post-ASML_era/The_Logic_Folding_Architecture_Guide.md |
Consumer of qualified TRS parameters for fold architecture design; tapeout eligibility gates on this procedure |
| The Multi-Regime Semiconductor Model | docs/post-ASML_era/The_Multi-Regime_Semiconductor_Model.md |
Physical basis for SC class acceptance thresholds and regime-specific fidelity limits |
| SC Classification Standard | docs/materials/SC_Classification.md |
Normative source for SC class assignment; EC-01 entry condition depends on this standard |
| SCR Zone Configuration Guide | docs/fab/SCR_Zone_Config.md |
Downstream recipient of this procedure's Qualification Certificate; zone commissioning follows |
| TTF Reference | docs/eda/TTF_Reference.md |
Consumer of L4 timing parameters (T_arb, T_seq, L_handoff) for TTF arc generation |
This document is part of the TriadicFrameworks canonical reference set. Proposed revisions
must be submitted via pull request to the docs/fab/ directory with a linked issue
describing the change rationale. Revisions to §12 (Acceptance Criteria by SC Class) require
simultaneous updates to docs/materials/SC_Classification.md (§4) and
docs/post-ASML_era/The_TRS-Aware_PDK_Specification.md (§13.2) and a cross-impact analysis
confirming that no qualification that was previously passing would fail under the revised
criteria, or that any such retroactive impact is explicitly addressed in the revision.