Whatsminer M30S – Power Supply Failure

Power cycle at the wall breaker (not the miner's power button) for 60 seconds, then restore. This clears PSU internal state machines that can wedge after a brown-out or bad startup sequence. A surprising share of code 210 tickets on the M30S resolve here alone — cost is zero, risk is zero, and it should be the first move before any disassembly or ordering of parts.

Swap the C19-to-NEMA power cord with a known-good one and try a different outlet on a different circuit if available. Cooked cords and dead GFCIs look identical to a dead PSU from the miner side. Roughly one in ten dead-miner reports is a bad cord or upstream outlet failure. Free to rule out, and you should always rule out the cheapest thing first.

Check the breaker rating. The M30S draws 3400 W nominal — roughly 30 A inrush at 120 V, 15 A at 240 V. A standard 15 A residential 120 V breaker will trip repeatedly at startup, making the miner look dead. Move to a 20 A 240 V dedicated circuit before condemning anything in the miner itself. Breaker trips at every start attempt are the diagnostic giveaway.

Read the chassis LED. Fast red blink = hardware-critical fault; slow red blink = warning in progress; red-green alternating = firmware / recovery state. Cross-reference with the D-Central LED decoder pages linked in Related Errors. The LED tells you whether the PSU even handed the control board a fault code, or whether the control board never came alive — two very different diagnostic paths.

Multimeter on AC, probe the outlet under load while the miner attempts to start. Target: 206 V minimum sustained on a 220 V feed, 228 V minimum on a 240 V split-phase residential, 202 V minimum on 208 V commercial. Below those thresholds the PSU correctly refuses to run — the grid is the problem, not the PSU. Fix the circuit (breaker, PDU, cable run, utility) before touching the miner.

Torque the copper bars. Power off at the breaker, open the miner, locate the two copper bus bars between PSU and hashboards. Check every bolt for snugness and heat-discolouration. Tighten to hand-snug plus a measured quarter-turn. Black burn marks or green/brown oxidation mean you've been losing voltage across that joint for a while — clean with IPA and re-torque. This single step resolves a real share of M30S PSU tickets at zero parts cost.

Re-seat the PSU-to-control-board data cable. Small ribbon or JST-style connector running between the PSU module and the control board. Unplug fully, inspect pins for bending or blackening, reconnect firmly until it clicks. Many intermittent code 200 and code 201 tickets resolve here alone — the handshake line is physically fine but the connector has loosened through thermal cycling.

Swap in a known-good PSU of the correct part number. Critical: the PSU must match the miner revision. A PSU that fits mechanically but carries a different firmware handshake will throw code 8700 (PSU/miner model mismatch) or code 201. Use the exact PSU part number listed on your miner's hashrate sticker. A clean swap that boots confirms the original PSU is the fault. A clean swap that still errors points upstream (grid) or downstream (hashboards).

Export logs via WhatsminerTool (Remote Ctrl > ExportLog) and record the decimal code plus the PSU hex subcode from power.log. The hex code (0x0001-0x2000) is the actual failure inside the PSU; the decimal is MicroBT's rough category. Cross-reference with the hex decoder notes in our Mining Hacker section — the hex tells you which subsystem complained, which aims your next repair move.

Disconnect all three hashboards from the PSU (copper bars off, data cables off) and power PSU + control board alone. If the PSU and control board run clean with no errors, the fault is downstream — a shorted hashboard or bad domain is pulling the PSU into OCP. Reconnect hashboards one at a time to isolate the culprit. If the PSU still clicks or errors with no load, the PSU itself is sick — proceed to Tier 3 or Tier 4.

Replace the PSU internal fan. Many 'failing PSU' tickets are actually a seized 40 mm or 60 mm fan inside the PSU chassis. Open the PSU — discharge caps first by leaving unplugged 10 minutes and shorting the large electrolytics through a resistor before touching anything — locate the fan, match voltage (12 V typical) and connector, swap. 15-30 CAD in parts plus 20 minutes. The most satisfying Whatsminer repair because the customer expected a 400 CAD invoice.

Replace bulging or leaking electrolytic capacitors. Visual inspection with PSU open: domed tops, crusty leakage, cracked sleeves. Match voltage and capacitance exactly, upgrade to 105 C-rated low-ESR series from a reputable manufacturer, re-solder carefully. Pre-failure audible buzz from the PSU is frequently the last warning before a hard short — don't ignore it, and don't run the miner in an unattended space while that buzz is audible.

Measure PSU output under full load with a DC clamp meter. Probe copper-bar current and output voltage simultaneously while the miner is hashing. A healthy M30S PSU holds regulation under a full 3400 W draw with under 2% ripple on the output rail. Ripple above 5% or voltage sag below the rail's designed window means the PSU is tired and needs replacement or rebuild, even if no error code has fired yet.

Flash latest stock firmware via WhatsminerTool. Version 20220218 and later include fixes for code 202 false-positives on the M30S+ revision. Pull logs first, then flash, then retest. Firmware bugs do occasionally present as PSU faults — this is a valid Tier-3 diagnostic. Do not cross-flash between M30S, M30S+, and M30S++ — they carry different PSU handshake tables and a cross-flash can brick the control board.

Inspect for water or humidity damage. M30S deployments in Canadian basements, garages, and unheated outbuildings are at real risk of condensation damage during shoulder-season temperature swings. Look for green corrosion on PSU board traces, white residue around electrolytics, or a faint acrid smell. Any of these means the PSU is compromised and cannot be trusted even if it powers up — replace the unit and fix the ambient humidity before installing the replacement.

Stop DIY when: (a) the PSU clicks or trips OCP with no load attached and no obvious cap/fan fault; (b) you see burn marks on the PCB or smell anything burnt; (c) the miner has taken a water event; (d) you're not confident discharging primary-side caps safely. You are in test-fixture territory, and there is real fire risk in running a compromised high-current PSU. Book a D-Central ASIC Repair slot — don't run it and hope.

D-Central bench process: PSU goes onto a programmable DC load, stepped 0-100% with data logging, regulation and ripple verified at every setpoint, caps inspected under scope, suspect components swapped from graded inventory, 24-hour burn-in at nameplate load before the miner returns. Per-M30S repair quote covers diagnostic, parts, labour, and 30-day post-repair warranty. Turnaround typically 5-10 business days.

Ship it safely. Anti-static bag around the PSU or the whole miner, double-box with 5 cm foam on every side. Include a printed copy of your exported power.log and a one-line description of the observed symptom. Clear notes on what you've already tried save diagnostic time at the bench — and diagnostic time is billable, so clear notes save you money directly. Label the box fragile and include contact info inside and outside.