Whatsminer M60S – Fan Speed Error

Read the exact sub-code off the dashboard before you do anything else. The WhatsminerTool event log and the miner's local web UI both show the numeric code. 110/111 means a tach line is silent. 120/121 means a fan is running more than 2000 RPM slower than its paired fan — usually dust, bearing wear, or a worn-out hub. 130/131 is the same deviation ≥3000 RPM, a critical threshold that will gate the miner within seconds. 140 means the control board is commanding 100% PWM and still can't cool the hashboards — that's an environment or thermal-paste issue, not a fan issue. Log the code before touching anything.

Power-cycle with a full cold reset. Kill power at the PDU or wall outlet (not the miner's own switch — you want the bulk capacitors on the PSU fully discharged), wait 60 seconds, then re-energize. The control board re-scans every fan tach on boot and re-applies the PWM curve; a transient fault that cleared itself on thermal recovery will not re-trigger. Note which fans come back healthy and which stay faulted. If every fan reports expected RPM after cold boot and stays healthy for 30 minutes under load, you had a transient fault — log the event and move on.

Blow the dust out of every fan hub, the intake grille, and the exhaust. Use dry compressed air from a can or a shop compressor with a moisture trap — never a household vacuum cleaner (electrostatic discharge into the fan driver IC is a documented kill). Pay particular attention to the rotor shaft of the suspect fan: hair, pet dander, insulation fibre, and shop-rag lint all wrap around the shaft and drag RPM down below the PWM-commanded threshold, which triggers 120/121. Also blow out any intake filter — a clogged filter raises intake temperature and causes false 140 trips.

Re-seat the fan power/PWM/tach connector at the control-board header. With the miner unplugged, trace the cable from the suspect fan back to the 4-pin header on the control board, unplug it, inspect for bent pins or scorch marks, then re-seat firmly. A surprising fraction of 110/111 tickets clear on a simple reseat — vibration over months works connectors loose, and the tach line is the first signal to drop out. If the plug is stiff or the pins look worn, mark the connector for replacement in Tier 3 or Tier 4.

Check ambient intake temperature and airflow path. The M60S is rated for intake below roughly 30 °C; above that, the control board commands fans harder to hold hashboard temp, and any fan that can't keep up trips 120/121 or 140. Measure with a digital thermometer at the intake grille while the miner is running. Clear any recirculation from a neighbouring miner's exhaust, any dust-loaded intake filter, and any physical obstruction within 30 cm of the intake grille. This is the single most common cause of 140 false alarms.

Swap the suspect fan into a known-good slot to isolate fan-side vs control-board-side fault. Power off, unplug the suspect fan from its current header, plug it into the header of a fan that was reporting healthy. Run the miner for 5 minutes and check which fan position now faults. If the fault follows the fan (suspect fan faults in its new slot), you've confirmed a fan-side failure — replace the fan in step 8. If the fault stays with the original slot (a healthy fan plugged into the suspect header now faults), the problem is the control-board PWM driver or the header itself — escalate to Tier 4.

Measure fan supply voltage at the control-board header with the miner running. Expected: 12V ± 5% (11.4 V to 12.6 V) at the V+ pin relative to ground on the 4-pin header. A sagging rail here means the control board's fan supply is compromised, usually a fuse or buck-converter failure on the board itself — fan swaps will not fix this. If the 12 V is solid but the fan still doesn't spin or runs slow, the fault is on the fan side. Probe with the miner running, not at boot — some headers soft-start the rail.

Replace the faulting fan with a matched spec unit. The M60S chassis fans are typically 120 mm × 120 mm × 38 mm (12038) axial units at 12 V, 4-pin PWM, rated around 6000 RPM peak. Verify the exact spec against the sticker on your own fan before ordering — MicroBT has shipped multiple suppliers (Delta, NMB, Sanyo Denki) across batches, and mismatched RPM specs will trigger 120/121 immediately after replacement because the new fan's peak RPM doesn't match the paired fan. Install with the airflow arrow matching the original, connect the 4-pin, re-seat, and cold-boot.

Verify the fix under 30 minutes of sustained load. Bring the miner back up, let it ramp to full hashrate, and monitor RPM of every fan plus hashboard temps on the dashboard for at least 30 minutes. Expected: all four fans reporting RPM within ~500 RPM of their paired fan, no fault codes re-appearing, hashboard temps stable in the mid-70 °C range. If 140 is still tripping after a successful fan swap, the problem was never a fan — escalate to Tier 3 step 11 for thermal-paste territory.

Audit the other three fans while the chassis is open. If one chassis fan just died at 12 months of 24/7 service, the other three are likely close behind. Spin each free by hand with power off — any that grinds, wobbles, or feels notchy is on the clock. A proactive swap of all four while the chassis is already open costs $40–$80 and prevents a second revenue-losing outage in 3–6 months. Canadian operators running miners in dusty barns, basements, or unfiltered outdoor sheds should audit every 6 months on schedule, not on failure.

If code 140 persists after fan replacement, re-paste the hashboards. Code 140 means the cooling system is at 100% and still losing the thermal battle, which points at degraded thermal interface material between the ASIC die and heatsink rather than a fan fault. On a 12–18 month-old M60S, factory thermal paste has typically dried out. Tear down, clean old paste with 99% isopropyl and lint-free wipes, re-apply Arctic MX-6 or Kryonaut (pea-sized drop per ASIC under the heatsink, not a smear), torque the heatsink bolts back in cross-pattern, and re-test. Expect 5–10 °C drop in hashboard temp.

Factory reset via the on-board RESET button. Hold for more than 3 seconds until the LED flickers red-green, then release. This clears any corrupt firmware state that might drive the PWM curve incorrectly — rare but documented. On the M60S, watch specifically for the dual-colour flicker and release the instant you see it; over-holding can land you in a maintenance mode that requires a re-flash to exit. After reset, reconfigure pools via WhatsminerTool, run the miner for 30 minutes, and check whether the fan fault recurs.

Re-flash stock MicroBT firmware via WhatsminerTool. The M60S enforces firmware signature verification hard — do not attempt to cross-flash Vnish, Asicdip, HiveOS, or any other custom firmware for this fault. Community reports of hard-brick on downgrade or unsigned-flash attempts on the M60S are well-documented and recovery requires JTAG access most field shops don't currently offer. Do not flash DCENT_OS — it is D-Central's open-source Antminer firmware only; Whatsminer hardware is not supported. Stick with the latest signed MicroBT build; verify the build string against the official MicroBT download page before flashing.

Probe the PWM duty cycle at the control-board header with an oscilloscope if you have one on the bench. Expected: a clean ~25 kHz square wave with duty cycle varying from ~20% at idle to ~100% under full thermal load. A stuck duty cycle (always 100%, always 0%, or garbage waveform) points at a failed PWM driver on the control board itself — not the fan. This is a bench diagnostic, not a field one; most operators will skip straight to Tier 4 on a confirmed header-side fault rather than chase component-level PCB work.

Ship to D-Central when the fault is isolated to the control-board PWM driver, the fan header is physically damaged (fractured pin, cracked PCB, lifted pad), or a full chassis-fan replacement does not clear the fault. At that point you are in component-level rework territory — hot-air station, pad-level PCB repair, or control-board replacement — which is not a field repair on the M60S. D-Central's Whatsminer bench covers control-board rework, header replacement, PWM driver reflow, and per-chip thermal diagnostics specifically for the SM3 generation.

Ship to D-Central when code 140 persists after both fan replacement and thermal re-paste. That means the heat problem is upstream of both fans and TIM — typically degraded VRMs on the hashboard, a shorted inductor, or in-silicon degradation of the ASICs themselves. None of those are field-fixable; they require hashboard-level rework or a hashboard replacement, and pricing the repair versus a replacement board is the bench's call, not the operator's. Canadian shipping to Québec is straightforward; US and international welcomed.

Do not reuse a used fan from an unknown Whatsminer variant as a replacement unless the sticker spec matches exactly. A wrong-RPM fan will pass a basic power-on test and then re-trigger 120/121 as soon as the control board pairs it against its neighbour — you will have wasted a teardown and still have a broken miner. D-Central stocks matched OEM-spec 12038 replacement fans for the Whatsminer M-series and will ship a known-good unit faster than chasing Aliexpress for a bin-mismatched bargain.