Whatsminer Error 218 – Voltage Below 230V in High Performance Mode

Confirm the exact code via the BTMiner API, not just the Web UI. Curl port 4028 with {"cmd":"status"} from a laptop on the same subnet and parse the error_codes array. [218] alone = HP-mode input-voltage gate failed. [218, 206] together = HP gate failed AND you are also dipping under the 180V hard floor in transients - electrical situation is worse than it looks. [218, 240] together = HP gate failed plus low-voltage protection cutoffs - you are running on the edge of a dead miner. Distinguish 218 from the 206 / 240 cluster before you spend money - they look similar in the Web UI but require different fixes.

Power-cycle at the AC mains, not the chassis switch. Kill AC at the PDU or wall breaker, wait a full 60 seconds for primary-side bulk caps to discharge. Some BTMiner firmware revisions latch the 218 flag through a soft reboot. A full AC cycle resets the latch cleanly. If 218 clears on AC cycle and returns within seconds-to-minutes when the miner attempts HP mode, you have a real input-voltage condition - not a stuck flag. Move on to clamp-meter measurement.

Drop the tune profile from HP / Turbo / Boost to Normal mode via WhatsMinerTool or the Web UI. Code 218 is HP-mode-specific - the strict 230V gate only fires when the miner is asked to run high-performance. Normal-mode minimum input is the more permissive 200V floor. Demoting the tune is the immediate workaround - lose ~10% hashrate, lose Code 218. Run there while you fix the underlying electrical situation. This is the fastest 'get back to hashing' move and should be your first action if the miner is your daily-driver income.

Clamp meter on L-L at the C19 inlet, measure under boot load. Use a Fluke 323 or equivalent true-RMS clamp meter. Cold-start the miner, watch the AC voltage during the first 60 seconds of boot and during the HP-mode ramp. Target: 235-245V sustained on 240V split-phase, with no sag below 230V at any point. Sag below 230V even briefly during HP ramp is enough to latch Code 218. Record the lowest reading - that number tells you how much margin you have lost and whether the fix is a circuit upgrade, a stabilizer, or a utility call.

Check for shared loads on the circuit. Walk the panel, identify the breaker the miner is on, list every other outlet / appliance on that breaker. M50S++ / M60S+ class miners pull 14-23A at full HP tune from 240V. Sharing a 30A breaker with a clothes dryer, water heater, garage fridge, or even another miner guarantees voltage sag during peak household load and periodic Code 218 events. Move the miner to a dedicated breaker, or move the other loads off. This single step fixes a meaningful share of home-miner Code 218 reports per D-Central's repair queue intake notes.

Inspect the C19 inlet plug and outlet for heat marks. Kill AC. Pull the C19 cable. Look at both ends under bright light. Discolouration, blackening, melted plastic, or heat-discoloured pins all indicate high contact resistance - which adds an extra voltage drop under load and pulls you below the 230V HP floor. Replace the C19 cable, replace the wall outlet if it's a NEMA 6-15 / 6-20 with damaged contacts. NEMA L6-30 twist-lock with a fresh contact set and a heavy-gauge cable is the right answer for production deployment - it locks in mechanically and resists the contact-fretting that develops on standard receptacles after years of vibration.

Run a dedicated 240V / 30A circuit with #10 AWG copper, NEMA L6-30 outlet. Single-miner home deployment running anything M50S++-class or larger needs its own circuit - period. Stranded #10 AWG copper on runs up to 25 metres handles 24A sustained without meaningful drop; #8 AWG above that. NEMA L6-30 twist-lock at the wall keeps the plug from working loose under thermal cycling. CAD $250-450 for an electrician visit (Canada / US average) and you have eliminated the entire 'shared circuit voltage sag' failure mode for the life of the miner. This is not optional gear at the M5x/M6x power class.

Re-measure under boot load with the miner on the new dedicated circuit. Same clamp-meter procedure as Step 4. New target: 240V steady with less than 2V sag during HP ramp. If you now hold above 230V comfortably, Code 218 should be gone - flip back to HP mode and verify the latch does not return. If you are still below 230V on a brand-new dedicated circuit with proper wire gauge, the problem is upstream of your panel - either utility-side voltage low at your meter, or a building-side issue like a long service lateral with too many simultaneous loads.

Measure voltage at the panel main breaker, not just at the miner outlet. With the miner running HP mode (or attempting to), put the clamp meter on the panel mains L1-L2 (240V split-phase residential) or L1-L2-L3 phase pairs (208V three-phase commercial). If the panel reads below 235V at the mains bus during miner load, the voltage drop is happening upstream of your panel. That is a utility / service-entry problem - not something more #10 AWG fixes. Document the reading and call the power company.

Call the utility company for a service voltage check. North American utility delivery spec is typically 120/240V +/- 5% (so 228-252V at the meter on residential). If your meter is consistently reading 220-228V, the utility owes you a tap adjustment on the local distribution transformer. They will send a tech with a recording voltmeter, log voltage for 24-48 hours at your meter, and adjust the transformer tap up one position if you are running below spec. This is a free service - they have a regulatory obligation to deliver in-spec voltage. Cite 'Code 218 fault on industrial equipment requiring 230V minimum' for fastest dispatch.

Install a voltage stabilizer / AVR if utility tap is correct but voltage still marginal. A 5-7 kVA automatic voltage regulator (buck-boost transformer with stepped autotap) installed between the panel and the miner clamps the input within 230-245V regardless of utility variation. Furman, APC, and SOLA-HD all sell hard-wired AVRs in this range. CAD $400-1,200 depending on capacity and topology. Sized properly (1.3x miner peak draw = headroom for inrush), an AVR eliminates the entire voltage-fluctuation failure mode and adds surge protection as a bonus. Standard kit on serious home-mining deployments where utility voltage is known-flaky.

Cross-check WhatsMinerTool input-voltage telemetry against your clamp meter reading. With the miner running and your clamp meter on L-L at the inlet, open WhatsMinerTool's PSU telemetry page and read the 'PSU Input Voltage' field. Healthy P221B: telemetry reading within +/- 2V of your clamp meter. Suspect: 5V or more delta between WhatsMinerTool reading and clamp meter under the same load. A wide gap means the PSU's input-voltage sense circuit (sense divider + ADC reference) has drifted - the PSU thinks you are below 230V even though you are not. That points at PSU-internal calibration drift, not external wiring. Tier-4 territory if you have eliminated all external causes and the gap persists.

Try a known-good P221B PSU swap to isolate. Pull a healthy P221B from another miner of the same model class (P221B-to-P221B only - do NOT cross to P21 or P222, the input-sense calibration tables differ). Swap into the suspect chassis on the same circuit, boot, attempt HP mode. If Code 218 disappears with the known-good PSU, the original P221B has an input-sense or calibration fault and is your culprit. If Code 218 reappears with a known-good PSU on the same circuit, the chassis is innocent - the problem is electrical / installation.

Check firmware version and update via WhatsMinerTool if behind. Some early BTMiner firmware revisions for M5x/M6x had calibration-blob bugs in the input-voltage table that caused false Code 218 latches at edge-of-spec voltages around 228-232V. Updating to the latest stable BTMiner image via WhatsMinerTool over the network often clears these false-positives. VERIFY the firmware image matches your exact model and sub-revision before flashing - cross-flashing an M50S++ image onto an M50S+ bricks the control board because BTMiner version-checks on the CB identity EEPROM.

Force SD-card firmware recovery if a network update fails or the input-voltage threshold remains stuck. Burn the latest stable BTMiner image (model-matched) to a microSD using Etcher or Raspberry Pi Imager. Insert the SD into the control-board slot, hold the reset button per the model-specific recovery sequence in the WhatsMiner self-service guide, let recovery complete (typically 5-10 minutes with LED status), pull the SD, reboot. This re-writes the input-voltage calibration blob from clean factory values and clears any silently-corrupted calibration that survives a simple re-flash.

Investigate building-side service-entry capacity if multiple miners on the same panel all show Code 218 simultaneously during peak hours. A residential 100A or 200A service feeder running 60-80% loaded all evening drops voltage at the panel by 4-8V from no-load reading. Adding a third or fourth miner pushes you over the edge into chronic Code 218 territory. Service upgrade to 200A (if currently 100A) or 400A (if currently 200A) is a CAD $2,500-6,000 electrician + utility job depending on jurisdiction. Run the math: at $0.08/kWh power and 5% hashrate loss on demoted tune, the upgrade pays back in 12-24 months on a single M66S - faster on a multi-miner deployment.

Test against a programmable AC source (bench-only). If you have access to a programmable AC source rated >5 kW (Pacific Power, Chroma, or Itech-class), pull the chassis from production and bench-test it at exactly 230.0V, 232V, 235V, and 240V with the tune set to HP. A healthy P221B / chassis combo enters HP cleanly at 230.0V and stays there. A drift-faulted P221B will latch Code 218 at 230.0V and only clear at 235V or higher - documenting the threshold mis-calibration precisely. This is the definitive test for input-sense drift and is the test D-Central runs on every Code 218 ticket that arrives at the bench.

Stop DIY when a properly-wired dedicated 240V/30A circuit holds steady 240V at idle but Code 218 still latches under HP boot. The fault is now PSU-internal (input-sense circuit drift on the P221B) or control-board-internal (calibration blob corruption that survives SD recovery). Bench reproduction against a programmable AC source with calibrated reference voltage is the only reliable diagnostic path. D-Central's bench process pins the fault category in one afternoon vs a weekend of swapping outlets and wires at home.

Stop DIY when intermittent 218 correlates with grid switching events you cannot reproduce at home. Substation switching, transformer tap changes, neighbourhood load swings - none of these are reproducible against a wall outlet at 11 AM on a Tuesday. Bench reproduction with a programmable AC source that can replicate dips, ramps, and waveform distortion is the only way to characterize the fault. Ship the chassis with as much PDU voltage logging context as you have - the more 'where the grid was when it tripped' data we have, the faster the bench reproduces it.

Ship with full context. Pack the chassis with the PSU (we need your exact stack to reproduce the fault - cross-stack substitution will miss marginal interactions), a copy of your last btminer-api status JSON showing the 218, your WhatsMinerTool PSU telemetry log if captured, your PDU voltage log if you have one, and your service history (recent moves, PSU swaps, tune changes, electrical work). Match chassis serial to PSU serial in your ship note. Note the building electrical context: 'NEMA L6-30 dedicated 30A breaker, #10 AWG, 22m run from panel, 240V residential split-phase' or similar. Every minute of context saved is a minute of turnaround shortened. Canada-wide standard shipping; US / international welcomed.

Discuss repair-vs-replace economics. A brand-new MicroBT P221B PSU runs CAD $340-440, graded-salvage from D-Central runs ~60% of new. A bench-level input-sense calibration repair runs CAD $90-180 if the underlying components are healthy. If your P221B is 4+ years old and the input-sense circuit has drifted, repair might be false economy - the rest of the unit is on the same aging curve and will throw a different fault next year. D-Central quotes up front with photo-documented diagnostics so you choose. For control-board calibration corruption alone, a clean SD recovery is free - bench fee of CAD $80-120 covers the diagnostic time.