DECISIONS.md — clar200-unlock Decision Log

Project-specific decisions. For global/cross-cutting decisions, see ~/bin/DECISIONS.md.

Format: YYYY-MM-DD: <decision> — <why>

Architecture

2026-06-14: Use the drive's F3 service UART, not any SCSI path — every SCSI command (incl. READ CAPACITY) returns Hardware Error on these EMC-locked drives; F3 bypasses the SCSI layer. (See EMC_CLAR200_DRIVE_FIX_NOTES.md for the exhaustive SCSI-side failure list.)
2026-06-14: ESP32-C3 as the USB-serial bridge rather than buying a USB-TTL module — one is already attached to steel141 (/dev/ttyACM0) and its 3.3 V logic is a native match for the F3 header (many FTDI cables are 5 V = unsafe).
2026-06-14: Bridge uses the C3's native USB-CDC as host side + UART1 (GPIO4 TX / GPIO5 RX) to the drive, with a live ~N baud menu so baud can be swept without reflashing.

2026-06-14: Build flag CDCOnBoot=cdc is required so Serial maps to USB-CDC, not UART0; encoded in firmware/flash.sh FQBN.
2026-06-14: F3 format commands are discovered via ?, not copied from 7200.11 forums — Exos X18 dialect differs and wrong m0,... params can worsen a drive.

2026-06-14: Rejected further SCSI-level attempts (sg_format flag variants, firmware download, sanitize) — prior session exhausted them; all return Hardware Error because the EMC firmware refuses below the command layer.
2026-06-14: Rejected doing F3 work with the drive in the pve173 JBOD — too risky next to live tank; bench-only.
2026-06-14: Deferred the BSY isolation/tape trick — the drive boots and answers INQUIRY (not BSY-hung), so it's unlikely to be needed; contingency only.
2026-06-14: Not yet decided — whether to RMA to ServerPartDeals first vs. attempt F3 first. Leaning: attempt F3 (cable is free + already here), RMA as fallback if F3 refuses.