Diagnostics · Whatsminer M63 (Hydro)

What it means: MicroBT surfaces every problem as a NUMBERED code (visible in the web UI, WhatsMinerTool, and via the API get_error_code), grouped by subsystem. You decode the number to the subsystem and apply the matching field fix. Golden triage rule: separate CONFIG faults (2000-series) from HARDWARE faults (power/temp/fan/board).

Symptoms

• A number + timestamp in get_error_code
• Reduced or zero hashrate
• Red LED flashing

Field procedure

1. Read the active code(s) from the API (get_error_code) or WhatsMinerTool.
2. Map the code to its subsystem using the families below.
3. Apply the matching field action (reboot / reseat / cooling / config / reflash / RMA).
4. Order of operations: reboot → check network/pool → reflash → reseat/RMA.

What it means: The integrated PSU's fault family. 200 = no power found (CB never saw PSU). 206/250/251/271 = input UNDERvoltage (very common). 233-238 = output over-temp/over-current. 253/254/274 = PSU fan. Hex 0x0001-0x2000 come from the PSU itself.

Symptoms

• Dead unit or board dropping under load
• Code 200 / 206 / 250
• PSU-fan codes 253/254

Field procedure

1. Code 200: check network/reflash first, then the internal PSU (detect power wiring, update firmware).
2. Input undervoltage (206/250/251/271): improve/steady the mains to ~220-240 V; try another circuit/phase.
3. Output over-temp (233-235): cool the room, improve airflow; power off 10 min then retry.
4. Over-current (236-238/268): re-torque the copper busbar screws; check the load.
5. PSU-fan (253/254/274 / 0x0800): check/replace the PSU fan; persistent → replace PSU (RMA-class).

What it means: 300/301/302 = SM0/1/2 sensor DETECTION error (check adapter/ribbon wiring). 320-322 = reading error. 350-352 = board over-temp PROTECTION (shutdown). 360 = general overheat. 600/610 = ambient too high.

Symptoms

• Code 300-series / 350-series
• Board shutting down
• Code 600 ambient high

Field procedure

1. Sensor detect/read (300-322): check the adapter board + ribbon wiring contact to that board; swap board position to see if it follows.
2. Over-temp protect (350-352/360): cool the environment, improve airflow, verify thermal contact and ambient.
3. Ambient (600/610): lower the room temp - high-performance mode wants intake <30-35 °C; miner auto-reduces frequency.

What it means: SMx = the hashboard in slot x (the last digit is the board number, 0-3). 410-412 = SMx detect EEPROM error (can't read the board's ID/EEPROM chip). 420-422 = SMx parser EEPROM error (EEPROM data corrupt). 430-432 = SMx chip-bin type error. 440-442 = SMx EEPROM chip-num error (chip count doesn't match firmware). 450-452 = SMx EEPROM transfer error. 510-512 = SMx wrong hashboard type/version fitted. 520-522 = SMx wrong chip bin/model. 530-532 = SMx not found (board not detected — usually a connection fault). 540-543 = SMx reading chip id error (board detected but its chip IDs can't be read; where a verified model/revision exposes SM3, code 543 refers to that slot; the M56S++ board count remains unverified). 550-552 = SMx has bad chips (dead ASIC chips — swap/RMA). 560-562 = SMx loss of balance (power/hashrate imbalance).

Symptoms

• Code 410-452 (EEPROM error)
• Code 510-522 (wrong hashboard type / bin)
• Code 530-532 (board not found)
• Code 540-543 (reading chip id error)
• Code 550-552 (bad chips)
• Code 560-562 (loss of balance)

Field procedure

1. EEPROM (410-452): reflash via SD/WhatsMinerTool; check the adapter cable; RMA if it persists.
2. Wrong type/bin (510-522): the board is the wrong version or chip bin for this model — fit the correct hashboard.
3. Not found (530-532): check adapter connection, re-torque the copper busbar, look for cold solder at the connector; still missing → swap CB or RMA the board.
4. Reading chip id (540-543): check the adapter board + ribbon contact and clean dust off the board; power off and swap the board to another slot — if the fault follows the board it's the board, if it stays it's the adapter/control board.
5. Bad chips (550-552): swap/RMA that hashboard - no field chip repair.
6. Loss of balance (560-562): reseat adapter, re-torque busbar power connection; check overclock params/firmware.

What it means: The CONFIG/connectivity family - usually not hardware. 2000 = no pool configured. 2010 = ALL pools disabled (check internet/DNS). 2020-2022 = poolN connect failed. 2030 = high rejection rate. 2310-2350 = hashrate too low / loss too high.

Symptoms

• Code 2010 (all pools disabled)
• Green LED solid (no internet)
• High rejects (2030)

Field procedure

1. 2000/2010: enter pools; check internet/DNS, verify pool URL:port, worker, and re-enable pools. A bad DNS is a frequent cause of 2010.
2. 2020-2022: check network, pool URL/port, credentials for that pool.
3. 2030 (high rejects): verify stable 220-240 V, network stability; back off any overclock.
4. 2310-2350 (low/lost hashrate): check input voltage, ambient, and for a down/underperforming board (see edevs).

What it means: The cooling-fan family (air units). 110/111 = fan-in / fan-out detect speed error. 120-131 = fan speed deviation (reads far off the 2000/3000+ RPM target). 140 = fan speed too HIGH - usually the miner ramping fans against a hot intake or an airflow blockage.

Symptoms

• Code 110/111 or 120-131
• A fan reading 0 or wildly off RPM
• Code 140 with a hot intake

Field procedure

1. Detect / deviation (110-131): reseat the fan connector, clear dust/debris, and swap in a known-good ORIGINAL-spec fan - wrong-spec fans throw speed codes.
2. Fan too high (140): check intake/ambient temperature and clear any airflow obstruction - the miner is commanding max RPM for a reason.

What it means: Liquid-cooled units only. 5070/5071/5072 = SM0/1/2 water/liquid velocity abnormal - the per-board coolant flow is out of range (too low or none).

Symptoms

• Code 5070/5071/5072 on a hydro or immersion unit
• Rising board temps alongside flow warnings
• One board's flow reading near zero

Field procedure

1. Confirm the pump is running and the facility loop is live (cooling ON before the miner).
2. Check for kinks, a clogged filter, or unseated quick-disconnects; verify coolant level and the rated flow (e.g. ~8-10 L/min).
3. If flow is genuinely normal at that board, replace the velocity sensor.