Skip to main content
Crypto Miners Direct

Maintenance guide

ASIC Miner Maintenance

Set up a maintenance programme for ASIC miners covering what to monitor, how often to inspect and clean, firmware updates, spares and records.

Published October 1, 2026 · Reviewed by James Emerson

ASIC miner maintenance is mostly watching a few numbers and keeping the environment right. Monitor hashrate, temperatures, fan speeds and error rates continuously, inspect and clean on a fixed schedule, update firmware cautiously, and keep spares for the parts that wear. Many faults show in the data before a machine stops. The intervals below are starting points to adapt to your site; a dusty container needs more attention than a filtered room.

Key points

  • Compare each machine's 24-hour average hashrate with its rated figure and with what the pool reports. A sustained gap is an early sign of trouble.
  • Clean on a schedule. Bitmain recommends about monthly for Antminers; MicroBT's operation guide says every three to six months.
  • Roll out firmware to a small batch first. Bitmain suggests 10 to 30 units for at least 24 hours before the rest.
  • Power off and disconnect before touching anything inside a machine, and discharge static first.
  • Opening a machine can affect warranty. Read the manufacturer's terms and the manual for your exact model before you do.

Safety and warranty first

Disconnect AC power and wait for the fans to stop before any work beyond looking. Bitmain's guidance is explicit: unplug the miner when installing or maintaining it, wear anti-static protection, and do not touch the control board and fan connectors, because they connect directly to chip pins. Never plug or unplug power connectors while the power supply is live.

Warranty rules differ by manufacturer. Bitmain publishes procedures for customers to replace fans and control boards, and states that replacing a fan does not affect warranty. MicroBT's warranty policy excludes damage caused by disassembly, modification or maintenance without its authorisation, and by unofficial accessories. The manual and warranty terms for your exact model are the authority. Electrical supply work belongs to a qualified electrician.

What to monitor

Signal Where to read it What to look for
Hashrate, real-time and average Miner web interface or management tool Average within the manufacturer's tolerance of rated. Bitmain publishes ±3% for models such as the Antminer S21 XP and treats hashrate below 90% of normal as a fault condition on its indicator lights
Per-board hashrate and chip count Miner status page All boards present, similar hashrate on each, full chip count
Board and chip temperatures Miner status page Stable values, similar across boards, well below the model's limit
Fan speeds Miner status page All fans reporting, similar speeds. One fan much slower than the others, or an implausibly high reading, indicates a failing fan
Hardware error rate Miner status page or log Low and steady. A rising trend on one board points to that board, its cable or its cooling
Accepted, rejected and stale shares Pool dashboard Pool-side 24-hour hashrate close to the machine's own figure; rejects low
Intake and exhaust air, or coolant supply and return Site sensors Inside the published range for the model, with margin
Uptime and reboots Management tool Unexplained restarts

Notes on reading these:

  • Hashrate fluctuates minute to minute. Judge on averages of several hours. After a restart, WhatsMiner units run a frequency search during which displayed hashrate is below normal; MicroBT gives about 40 minutes for it.
  • Temperature limits are model-specific. Bitmain publishes PCB maximums of 75 to 85 °C for S19-generation models, depending on the model; newer models have their own values. MicroBT states that when hashboard outlet temperature exceeds 80 °C with fans above 7,000 rpm, an air-cooled WhatsMiner reduces frequency, and recommends ambient below 30 °C.
  • A pool measures hashrate from the shares you submit, so it is an estimate that needs hours to settle. A persistent shortfall against the machine's own reading suggests network problems, stale shares or a hardware error rate the machine is not showing clearly.
  • For network health, Bitmain's test is a ping to the pool with no packet loss and an average delay under 100 ms.

Maintenance schedule

Treat these as recommendations and tighten them where dust, heat or humidity are high.

Interval Task
Continuous Automated alerts for offline machines, low hashrate, high temperature, fan faults
Daily Review alerts. Walk the site: listen for noisy fans, look for blocked intakes, check for hot spots, smells and leaks
Weekly Compare pool-side and machine-reported 24-hour hashrate. Review error-rate and temperature trends. Inspect intake filters
Monthly Inspect machines for dust and clean as needed. Check that power cords and connectors are seated and not discoloured. Review the worst-performing units
Every 3 to 6 months Thorough dust clean of each machine. Check firmware releases. Count spares. Replace or wash filters
Seasonal Re-balance airflow for the coming season. In cold climates, plan how machines will be kept above their minimum start temperature
Annually Electrical inspection of distribution, terminations and grounding by a qualified electrician. Review records and retire or reassign uneconomic units

Liquid-cooled sites add coolant work. MicroBT's hydro guide calls for pH, conductivity and appearance checks every two months, a fuller analysis every six months, and recommends replacing coolant after a year. Bitmain's guidance for its hydro containers is weekly water-quality checks in the first month, monthly for the next six months, then quarterly or half-yearly. Add leak inspection, filter cleaning and pump checks to the daily and monthly rows.

Cleaning

Dust is the most common cause of avoidable overheating. Bitmain's method for Antminers is compressed air with the machine shut down; the full procedure and its cautions are in How to Clean an Antminer. MicroBT asks for dust to be cleaned from hashboards, the control board, the adapter board and cables every three to six months, and states it accepts no warranty responsibility for serious corrosion, dust or moisture.

Fit intake filtration that suits your dust load and check that it does not starve the machines of air. In seasons with airborne seeds or fluff, Bitmain recommends a fine dust screen on the intake and regular checks through the inlet fan.

Firmware updates

  • Use only firmware for your exact model from the manufacturer. Bitmain warns to confirm the firmware matches the miner model before upgrading.
  • Read the release notes and update for a reason: a fix you need or a security patch.
  • Pilot first. Bitmain's advice is to upgrade 10 to 30 units and run them for at least 24 hours before the rest.
  • Decide whether to keep settings. On Antminers, an upgrade without the keep-settings option resets pool and network configuration, and the machine's IP address may change.
  • Do not interrupt power during an upgrade. Bitmain notes that if a power failure during an upgrade leaves a machine unable to upgrade or re-flash, it generally needs repair.
  • Record the version on each machine before and after.

Spare parts to hold

Hold spares in proportion to fleet size and to how long replacements take to reach you.

  • Cooling fans of the exact size and rating your models use. These are the most frequent replacement.
  • Power supplies of the model your machines require. PSU families are not interchangeable.
  • Control boards matching the miner family.
  • Signal cables and power cables, which are cheap and useful for fault isolation.
  • Network cables and a spare switch.
  • For liquid cooling: quick connectors, hoses, seals, filter elements and coolant.

Be careful with spare hashboards. Bitmain defines a machine containing hashboards or control boards that are not its original parts as a mixed-board machine, which voids its warranty. For machines under warranty, a faulty board is a repair case, not a swap.

Record keeping

Keep one record per machine, keyed by serial number:

  • Model, hashboard and control board serial numbers, purchase and warranty dates.
  • Rack position and power circuit.
  • Firmware version history.
  • Baseline hashrate, temperatures and fan speeds when new.
  • Every fault, with the log extract, and every part replaced.
  • Cleaning dates.

Records turn single readings into trends. A board whose temperature has risen 8 °C since its baseline at the same intake temperature needs cleaning or repair, even if it is still inside its limit.

Environment control

Keep conditions inside the manufacturer's published range for your model. As examples of what manufacturers ask for: Bitmain's general requirements for Antminers are 0 to 40 °C operating temperature, 10 to 90% relative humidity without condensation, and suspended dust at or below 0.2 mg/m³, with a recommended working range of 5 to 35 °C and humidity below 65%. MicroBT's air-cooling guide gives ambient humidity of 20 to 80% and dust below 0.5 mg/m³. Individual model specifications can be wider or narrower, and they govern.

Practical controls: separate hot exhaust from intake air, filter the intake, prevent condensation after shutdowns, keep machines grounded, and give power supplies load margin. Bitmain recommends loading a PSU to no more than 90% of its rating.

When to stop and use a repair service

Stop and open a repair ticket with the manufacturer, or use a qualified repair centre, when:

  • A hashboard is missing or short of chips after cables and power supply have been ruled out.
  • You see or smell burning, or find corrosion, scorched connectors or liquid on a board.
  • A power supply is suspected. Do not open a PSU.
  • The machine is under warranty and the fix would mean exchanging boards.
  • A fault returns after the basic checks in Hashboard Troubleshooting.

Sources