# Representative A/B operator target gates — 2026-08-11

Status: PASS for the two installed SD-card specimens on the frozen candidate
Type: evidence
Updated: 2026-08-11
Next Action: retain this result with the candidate record; track export latency separately
Campaign Reason: Historical record from campaign 009-complete-clean-contract-bound-a-b-qualification; retained independently after campaign closure and requires no separate owner-queue execution.
Campaign: standalone
Parent: work/evidence/representative-ab-regression-20260811.json
Project Map: work/maps/map-smi-firmware-stabilization-and-qualification.md

## Disposition

Controllers A and B passed the remaining normal-production logging/chart-read
and configuration save/restore gates. The complete representative A/B result
for these two installed cards is therefore `QUALIFICATION_PASS_TARGET_AB`.
This disposition does not close the broader multi-sample retail-card,
controlled-document, ECO-release, or manufacturing-rollout gates.

The frozen Release SREC SHA-256 was
`37FF5C0C6F48576396C8EF3807AAE9C8F1DDAB742A72C47A4C1055454DA906F0`;
the Release ELF SHA-256 was
`88E9654FF2BFA4CFF1C3801978558A95CBD94421043550206112CA2DA0C4B1D2`.

## SD-card identity

Identity was captured at the run baseline and again after the configuration
exercise. No card was removed, inserted, or replaced, and both final CIDs
matched their baselines.

| Field | Controller A | Controller B |
| --- | --- | --- |
| Controller assembly | 1751136 | 1751219 |
| J-Link serial | 69664082 | 69660123 |
| Protocol card status | 6 — detection ready, raw HPT, driver OK | 6 — detection ready, raw HPT, driver OK |
| CID-valid status | 1 — valid | 1 — valid |
| Representative-sector-read status | 0 — success | 0 — success |
| Manufacturer name / MID | Unknown/associated; MID `0x05` | Unknown/associated; MID `0x05` |
| OEM/application ID | raw `0x000C`; non-printable display `??` | raw `0x000C`; non-printable display `??` |
| Product name | `asdfg` | `asdfg` |
| Product revision | 2.2 | 2.2 |
| Serial number | 7305 | 7298 |
| Manufacturing date | 2026-03, valid | 2026-03, valid |
| Complete raw 128-bit CID | `05000C61736466672200001C8901A34F` | `05000C61736466672200001C8201A379` |

Raw final identity captures:

- Controller A: `work/evidence/hpt-log-store-circular-mirror/controller-a/generated/qualification-continue-20260811/final-identity/sd-card-identity.json`
- Controller B: `work/evidence/hpt-log-store-circular-mirror/controller-b/generated/qualification-continue-20260811/final-identity/sd-card-identity.json`

## Normal production logging and chart read

Logging was observed over a natural 116-second interval before opening the
normal graph screen and capturing its UI/cache state.

| Check | Controller A | Controller B |
| --- | ---: | ---: |
| Completed records | 306 → 308 | 1395 → 1397 |
| Queued records | 306 → 308 | 1395 → 1397 |
| Dropped records | 0 → 0 | 0 → 0 |
| HPT next sequence | 1,064,175 → 1,064,177 | 624,100 → 624,102 |
| Graph screen active / current / follow-newest | PASS | PASS |
| Real chart data / cache ready | PASS | PASS |
| Cache records scanned / selected | 56,160 / 56,160 | 56,160 / 56,160 |
| Bulk fallbacks | 0 | 0 |
| RTC-to-newest-record lag | 55 seconds | 57 seconds |

The stable `hpt_store_error_flags=65536` value was present before and after on
both controllers. It did not change during this gate. Authoritative HPT state
and status remained 5/0, mismatch/loss/retirement counters remained zero, and
system errors remained zero.

## Configuration save and exact restore

Each controller was exercised independently. The current setpoint was captured,
changed by +0.1 °C, committed, restored to its exact captured bits, and
committed again.

| Check | Controller A | Controller B |
| --- | ---: | ---: |
| Original → temporary setpoint | 20.80005 → 20.90005 °C | 22.00009 → 22.10009 °C |
| Generation before → restored | 232 → 234 | 178 → 180 |
| Exact setpoint bits restored | PASS | PASS |
| VEEPROM copies identical | PASS | PASS |
| RAM matches persisted configuration | PASS | PASS |
| RAM payload unchanged except generation counter | PASS; only byte offset 4 changed | PASS; only byte offset 4 changed |
| Operational / last commit succeeded | 1 / 1 | 1 / 1 |
| Findings / controls inhibited / commissioning required / save pending | 0 / 0 / 0 / 0 | 0 / 0 / 0 / 0 |
| False warning shown | No | No |

Both temporary and restore transactions encountered the known bounded first
attempt `SSP_ERR_CMD_LOCKED` condition at stage `0x1100`, recovered on attempt
2 within the qualified three-attempt policy, completed with zero findings, and
kept the operator warning silent. This matches the previously accepted
behavior in
`work/incidents/incident-controller-b-false-configuration-invalid-warning.md`;
it is not a qualification failure.

Raw configuration exercise captures:

- Controller A: `work/evidence/hpt-log-store-circular-mirror/controller-a/generated/qualification-continue-20260811/config-exercise/config-save-exercise.json`
- Controller B: `work/evidence/hpt-log-store-circular-mirror/controller-b/generated/qualification-continue-20260811/config-exercise/config-save-exercise.json`

## Final target health

Both controllers ended with HPT media state/status 5/0, lane mask 7, zero
mismatch, zero loss, zero retired segments, zero system errors, valid chart
data, and continuing normal logging. The export scan-latency concern remains a
separate accepted/deferred performance issue in
`work/tickets/ticket-eliminate-full-history-scan-latency-for-bounded-hpt-customer-exports.md`;
it does not invalidate the correct export content or either installed card's
qualification result.
