ZY12PDN PD Trigger Set Wrong Voltage Damaged Bitaxe

Lock every PD trigger in your workshop to fixed-5V mode the day you receive it. Hold the ZY12PDN button while plugging the trigger into a powered USB-C charger — the LED will rapid-flash through colours. Release the button when the LED turns red. Press once more to land on red (5V). Hold the button for ~1 second to save fixed-5V mode. Verify with a multimeter at the output: 5.0 V plus or minus 0.1 V, with no sweep through other voltages on power-cycle. Procedure documented at alexwhittemore.com/notes-on-usb-pd-triggers-and-zy12pdn-instructions/.

Tape over the button after locking. Strip of electrical tape, cap of a Sharpie cut to fit, dab of hot glue over the dome — anything that prevents an accidental press. A child, cleaner, bumped tool, or transit vibration can each cycle the trigger to the next voltage and turn your 5V-locked module into a 9V time bomb. Tape is free and saves silicon.

Multimeter the trigger output BEFORE every single Bitaxe connection. Set meter to DC, 30V range. Probe centre pin to outer ring of the trigger barrel output. Confirm 5.0 V plus or minus 0.1 V. Only then connect the Bitaxe. This 10-second discipline is the most important habit in this entire page. If you implement nothing else, implement this — it catches firmware corruption, factory-reset events, and accidental button presses before they reach the board.

Label the trigger module with Sharpie on the case: `5V FIXED — Bitaxe`. If you operate multiple triggers (one for 9V LED strip, one for 12V 3D-printer accessory, one for 5V Bitaxe), label each with the locked voltage AND the device class. Mixing them up is the failure mode this page exists to prevent. Recovering the program-mode button sequence six months later from memory is a frustration nobody needs.

Buy a fixed-voltage trigger or a sealed 5V/6A brick instead of a programmable trigger if you can. A fixed-5V-only PD trigger (no button, no other modes — STM32F0 firmware hard-coded) eliminates the wrong-mode failure entirely. A D-Central Bitaxe 5V/6A brick is even safer — no PD logic at all, just 5V from AC. The ZY12PDN flexibility is great for the toolkit; for the bench under your hashing rig, less flexibility = less risk.

Before applying any further power post-incident: cold-diagnostic the board. Multimeter on 200 ohm or diode mode, board fully unpowered, capacitors discharged. Probe to ground from: (a) +5V input pad on barrel jack — healthy is greater than 10 kohm, less than 100 ohms is input-side short; (b) TPS546 output cap (large ceramic next to BM1366/1368/1370) — healthy greater than 1 kohm, under 5 ohms is VRM output stage fused and ASIC almost certainly dead; (c) ESP32-S3 module 3V3 pin — healthy greater than 5 kohm, dead short is LDO fused. Document each reading.

If cold diagnostic reads clean across all three rails: bench-supply boot test. Set bench PSU to 5.00 V, current limit 100 mA. Connect to the barrel input via probes or a temporary pigtail. Enable output. Healthy boot: brief spike to ~300 mA (PSU folds back), then settles to 50-80 mA idle while ESP32 brings up WiFi. Pinned at 100 mA = something is shorted that the cold meter missed. Power off and re-probe. Step the limit up to 2A then 6A while watching for instability or thermal hot spots.

If only the input pad is shorted and VCORE is clean: hot-air the burnt input components off. Almost always one or two ceramic input caps near the TPS546 plus the TVS diode if fitted. Replace with same-spec parts: 10 uF / 25V X7R (0805 or 0603) for input ceramics, SMBJ6.0A (or per-board-rev TVS spec) for input clamp. Re-test cold resistance. Greater than 10 kohm at the input pad to ground means you are in business; reapply Tier 1 prevention discipline going forward.

Replace the TPS546 if its package is visibly damaged but VCORE is still clean to ground. This is QFN-class rework with small pitch — hot air, flux, careful placement, pin-1 orientation verified under microscope before reflow. Fresh TPS546D24A from Digi-Key is roughly 15 CAD. After replacement, cold-test all three rails before applying power. Step current up gradually as in step 7.

Document and label the rebuild. Sharpie or label-maker on the underside of the board: `REPAIRED <DATE> — input ceramics + TVS` (or whatever combination you replaced). Future-you, future-buyer, or D-Central if it comes to us later all need to know this board has rework history. Hides nothing, helps everything. Mining Hacker honesty in workshop record-keeping.

If VCORE is shorted to ground in cold diagnostic: the BM1366 / BM1368 / BM1370 is almost certainly dead. A new ASIC chip from a salvage source costs 30-80 CAD depending on family and supply. Removal requires hot air at ~310 C, careful chip lift, pad clean-up with desoldering braid, fresh paste reflow on a known-good replacement. ASIC removal also frequently damages adjacent passives — budget for replacing surrounding caps and resistors in the same operation. Without a 2000+ CAD rework station this is not an iron-and-flux job; book Tier 4.

Salvage the rest of the board for parts before recycling. The EMC2101 fan controller, the ESP32-S3 module if its 3V3 LDO survived, the barrel jack, the I2C pull-ups, the fan connector, the OLED ribbon connector, the bare PCB pads — all useful for repairing other Bitaxes that died from different failure modes. Mining Hackers do not throw silicon away when it can become donor parts. Anti-static bag, label, into the spares drawer.

Build a permanent PD-trigger sanity-check jig: PD trigger, then inline DC voltmeter (0.1V resolution), then barrel pigtail. Plug the trigger into a charger BEFORE every connection to a Bitaxe. The voltmeter shows you what voltage the trigger is on, in real time, before you commit a 200 CAD board to the connection. Total cost: about 15 CAD and 20 minutes of bench time. Pays for itself the first time it catches a wrong-mode trigger.

Add an inline crowbar / TVS on the bench-side 5V output. A SMBJ6.0A or SMCJ5.0A TVS diode soldered across the trigger barrel-output wires (before the Bitaxe connector) clamps any wrong-voltage event hard enough to either trip the charger over-current protection or self-destruct as a sacrificial fuse. Costs pennies. Protects every Bitaxe downstream from any future trigger misconfiguration. Mining Hacker layered protection.

Roll custom firmware on the ZY12PDN's STM32F0 via SWD. Several open-source projects on GitHub (search `ZY12PDN custom firmware`) provide ready-to-flash binaries that hard-code 5V output regardless of button presses, factory-reset state, or power cycles. Overkill for a single board. Mandatory for a fleet — when one button bounce can vaporize a BM1370, removing the bounce entirely is cheap insurance. Document the firmware revision on the trigger label.

Stop DIY and book D-Central Bitaxe repair when: VCORE-to-ground is shorted and you do not have hot-air rework gear, the PCB shows trace damage or solder-mask discoloration beyond a small area, multiple voltage domains are dead simultaneously, or you are not equipped for QFN / fine-pitch ASIC rework. Our bench triage cold-meters every rail, documents the damage, and quotes a salvage-or-replace number based on what is actually dead. We tell you honestly when a board is past the salvage line. Book at https://d-central.tech/services/asic-repair/. D-Central also stocks every Bitaxe variant pre-bundled with the matching fixed 5V or 12V brick at https://d-central.tech/product-category/bitaxe/.