DEC-096 - Stack-Blanking Board Corrected: ENC_ACTIVE_N/ACTUATE_REQUEST_N Termination Removed; ACTUATE_REQUEST Bridging Added

Field	Value
Decision ID	DEC-096
Status	Decided
Date	2026-09-02
Author	Izzyonstage & GitHub Copilot
Amends	-
Related	DEC-090, DEC-091, DEC-093, DEC-094, DEC-095

Context

An ESD/connector audit (DEC-095) surfaced that the Stack-Blanking Board's design had not been updated to match the current ACTUATE_REQUEST signal model (DEC-093) or the ENC_ACTIVE_N scoping confirmed during the J3 pinout work (DEC-090):

1. R1 (ENC_ACTIVE_N pull-up) is stale. ENC_ACTIVE_N is confirmed HID-local-only (Cypher Board ↔ Cypher-Input ↔ Cypher-Output only) and never appears on the STA/REF chain at all — there is nothing for this resistor to terminate.
2. R5 (ACTUATE_REQUEST_N pull-up) is based on an obsolete model. The Stack-Blanking Board's design predates DEC-093, which established that ACTUATE_REQUEST is not a signal that dead-ends at the chain terminus — it is a full round-trip signal that originates and terminates at the Cypher Board's own CPLD, reflecting twice at this board (STA-chain → REF-chain, and later REF-chain → STA-chain in reverse). A dead-end pull-up is functionally wrong for this signal.

Decision

1. R1 (ENC_ACTIVE_N) and R5 (ACTUATE_REQUEST_N) are removed entirely from the Stack-Blanking Board — no replacement termination is required for either.
2. The remaining three termination resistors are renumbered to close the gap: R2→R1 (TCK pull-down), R3→R2 (TMS pull-up), R4→R3 (CPLD_RESET_N pull-up). These three signals (TCK, TMS, CPLD_RESET_N) are the only ones that genuinely dead-end at this board.
3. Two new bridging paths are added, following the same passive J1↔J2 trace-routing pattern already used for SIG-BLOCK-A/B, C/D, and E/F:
   • ACTUATE_REQUEST_OUT_N (arriving on J1, from the last mini-stack's Stack-Input, the STA-chain forward-pass terminus) is bridged to ACTUATE_REQUEST_REF_IN_N (departing via J2 into the REF-chain) — the first turnaround in the DEC-093 round trip (step 3).
   • ACTUATE_REQUEST_REF_OUT_N (arriving on J2, from Stack-Output, the REF-chain second-forward pass terminus) is bridged to ACTUATE_REQUEST_IN_N (departing via J1 into the STA-chain in reverse) — the second turnaround (step 6).
4. Cross-references to the connector-defining boards are updated to reflect DEC-094 (Stack-Input owns J1's template, Stack-Output owns J2's template — not Cypher), and stale merge-cypher-board-j3j6-pinouts references are removed.

Rationale

• Bridging (not terminating) ACTUATE_REQUEST at this board is the only design consistent with DEC-093's round-trip model — the signal must physically continue past this board twice, not stop here. This exactly mirrors how the same board already correctly bridges (rather than terminates) every other passthrough signal in the system (ENC data, TTD).
• Removing the two resistors that no longer correspond to any real signal at this board (rather than leaving them stubbed/NC) keeps the BOM and schematic honest about what's actually connected, avoiding a future reviewer wondering what R1/R5 are for.
• Renumbering the remaining resistors closes the gap left by the removals, consistent with this project's established renumbering practice (e.g. DEC-055, the Rotor RefDes gap-removal precedent).

Impact

• Stack-Blanking/Design_Spec.md — Overview responsibility table, FR-SBLK-05/06/07 (06/07 repurposed for the new bridging FRs), DR-SBLK-02/03/05/07, mermaid diagram, §3 Signal Routing and Termination (both tables), §4 Interconnects, BOM all updated. Resistor count reduced from 5 to 3.
• Stack-Blanking/Board_Layout.md — J1/J2 sections and §3 Signal Bridge Summary updated to match.
• merge-actuate-request-routing todo — Stack-Blanking Board scope now resolved.