Whatsminer M30S – Fan Speed Error

Hard power-cycle at the breaker for 30 seconds (not a soft restart - clears the PWM controller latch state). Power back up and watch the front-panel LED and the WhatsminerTool error log for the first 2 minutes. If the error clears and does not return inside 15 minutes you had a transient tach glitch; log the date and move on. If it returns, continue to step 2.

Identify the exact fan that tripped by exporting error.log from WhatsminerTool, or by querying the API on TCP 4028 with `get_error_code`. You are looking for one of codes 110, 111, 120, 121, 130, 131, or 140 paired with a fan index (fan1-fan4). Without this you cannot target the fix - a 110 on fan2 vs a 140 on fan4 is the difference between an $18 fan swap and a firmware rollback.

Power off at the breaker, remove the side panel (Phillips #2), power back up, and watch each fan spool up with the side off. Any fan that stutters, wobbles audibly, rocks on its axis, or fails to reach the same visible speed as its siblings is the suspect. Also look for dust packs on the blade leading edges - a ~2mm dust layer is enough to shift RPM out of spec.

Shop-vac the front intake filters. Use compressed air on the fan blades from the outside of the chassis with the miner powered off. Do not spin fans above their rated RPM with compressed air - you can back-EMF the motor and blow the controller IC. Hold the fan hub stationary while you blow through it.

Check ambient temperature at the intake. Target: intake air <= 30 degC per MicroBT; practical <= 35 degC on fresh fans, <= 30 degC on fans past 18 months. Above those thresholds on worn fans, bearing drag spikes and you throw 130/131 on marginal units that would pass in a cool room. If ambient has crept up with the season, fix your airflow before you touch the miner.

Power off at the breaker. Locate the 6-pin fan header on the control board for the tripped fan (labeled J1-J4 or FAN1-FAN4 depending on board rev). Unclip the connector, visually inspect male and female pins for oxidation, blackening, or bent contacts. Wipe with 99% isopropyl and a lint-free swab if oxidized. Re-seat firmly, listen for the click, confirm no cable tension at an angle. M30S chassis vibration will slowly back a loose connector out over months - this is the single most common intermittent fan-error fix.

Swap the suspect fan into a known-good slot. Unplug the suspect fan, plug it into a known-good header, move the known-good fan into the suspect's header. Boot and watch the error log. If the error follows the fan to the new slot, the fan is bad (continue to step 8). If the error stays with the slot, the control board tach input for that slot is dead (go to Tier 4, step 16).

Replace the faulty fan with a compatible 12 V 12038 axial fan (120x120x38mm, 12V, 6-wire PWM+tach, ~6000 RPM). OEM (Yuanqiao/SXDOOL) or quality aftermarket (Delta AFB1212SHE). Avoid 4-wire fans - they lack the tach return wire and will throw 110/111 immediately. Mount with stock M4 screws into chassis threaded standoffs; preserve airflow direction (arrow on fan frame pointing intake-to-exhaust).

With the miner hashing, measure the 12 V fan rail at a spare fan header on the control board. Expected: 11.8-12.3 V sustained, no visible dip below 11.5 V during display refresh. A consistent sag below 11.5 V means the PSU 12 V auxiliary rail is tired - this rail also powers the control board and network, so expect intermittent UI freezes too. Swap the PSU if it sags.

Inspect the PSU (MicroBT P221B or P221B+) for discoloration, bulging, burnt-component smell, or an internal fan that spins rough. A tired PSU tends to fail on the fan rail (lower current) before it fails on the hashboard rail. If the PSU is suspect, cross-reference the Whatsminer M30S Power Supply Failure page and consider a swap before spending more time on fan-side diagnostics.

With the miner powered down and the fan unplugged, spin the fan blade slowly by hand while probing the yellow (tach) wire to black (ground) on DC V with a multimeter. Expected: pulsing between ~0 V and ~5 V as the hall sensor crosses the rotor magnet. A steady 0 V, steady 5 V, or no toggle during rotation indicates the tach line is broken internally or the hall sensor failed. This is your confirmation that the cable (not the bearing) is the fault before ordering a replacement.

If the cable is visibly flexed at the fan hub exit and the tach line tests dead, a careful solder repair is possible: open the cable jacket ~20mm back from the hub, identify the three signal wires (red 12V, black GND, yellow tach, plus PWM and second GND/tach on 6-wire), cut the yellow at a clean section, re-splice with ~26 AWG stranded, heat-shrink each joint individually, then a jacket heat-shrink over the group. Shop-work only - if unsure, replace the whole fan.

Flash the last known-good firmware via WhatsminerTool > Upgrade. Common known-goods for M30S: 20240426.16.xx family - verify your hardware revision on the Whatsminer downloads portal before flashing. M30S firmware is signed but not revision-locked for downgrade within the same major family; cross-major downgrade (e.g. 20250x -> 20240x) can brick the control board on some revisions. Do not cross-major downgrade without a JTAG recovery setup.

Replace the control board if the tach input channel is dead. Tier 4 territory on most units, but advanced shops with a hot-air rework station can replace the tach-sense passive components around the PWM controller if the fault is in the input divider (typically a 10k/10k resistor pair near the PWM IC). Desolder, replace with tested values, retest with a known-good fan. If the fault is in the PWM controller silicon itself, the control board is a swap, not a repair.

Inspect the chassis for mechanical vibration sources. A marginal fan on a chassis that vibrates heavily (often from a cracked rubber fan mount or warped chassis after a drop) trips intermittently even if mechanically fine. Replace any missing or crumbled rubber fan grommets (M3 sorbothane-style grommets, 100-packs available online). Tighten all chassis screws - loose side panels amplify fan vibration into the control board.

Stop DIY when: control board tach input is dead (confirmed via Step 7 swap test), multiple fans trip at seemingly random positions on each boot (PWM controller silicon failure), or you find burnt components or smell on the control board. These are bench-level fixes - attempting them without ESD protection and a proper hot-air station will cost you the board. Book a D-Central ASIC Repair slot and ship.

What D-Central does on the bench: full fan assembly test with a fan load simulator (isolates fan vs cable vs control board in under 10 minutes), control-board-level repair including PWM controller replacement, tach divider network re-flow, 12 V rail capacitor replacement if PSU aux is sagging at the board end, and 24-hour post-repair burn-in at nameplate power before ship-back. On M30S units > 3 years old, we bundle a fan refresh quote because the remaining fans are usually within weeks of the one that failed.

Ship safely. For domestic Canada/US, ship the whole unit in its original box where possible, or double-box with >= 5cm foam on every side. Remove hashboards only if instructed (international freight). Include a note with the error codes observed, firmware build number, and which fan position tripped - that note saves 30 minutes of bench diagnostic, which comes directly off your repair bill.