ASIC Hashboard Repair Deep Dive: Chip-Level Diagnostics, Failure Analysis & Rework Techniques

A hashboard is the beating heart of every ASIC miner — and when it fails, it is rarely obvious why. This guide takes you inside the hashboard: its architecture, the failure modes that kill it, the diagnostic process that isolates faults down to individual components, and the repair techniques that bring dead boards back to full hashrate. Whether you are a home miner trying to understand what went wrong or a technician building repair skills, this is the reference you have been looking for.

At D-Central Technologies, we have repaired thousands of hashboards since 2016 — from legacy Antminer S9 boards to the latest S21 generation. This deep dive distills that experience into a single, comprehensive resource. No fluff. No hand-waving. Just the technical reality of what happens when silicon meets solder at scale.

What Is a Hashboard?

An ASIC miner is built around one to four hashboards, each one a specialized printed circuit board (PCB) populated with dozens of Application-Specific Integrated Circuits — the ASIC chips that do the actual SHA-256 computation. The control board orchestrates operations, but the hashboard is where every single hash is generated. It is the production floor of a bitcoin mining operation condensed onto a board roughly the size of a large laptop screen.

Each hashboard is a self-contained hashing engine. It receives power from the PSU, clock and control signals from the control board, and returns nonce results when it finds a hash below the current target. Understanding the hashboard as a system — not just a collection of chips — is the first step toward effective diagnosis and repair.

For a broader comparison of how hashboards relate to control boards and the overall ASIC architecture, see our Control Board vs Hashboard guide.

Hashboard Architecture: Anatomy of a Hashing Engine

Despite differences between manufacturers and generations, every hashboard shares fundamental architectural elements. Understanding this architecture is essential before attempting any diagnostic or repair work.

ASIC Chip Array

The ASIC chips are the primary active components on the board. Depending on the model, a single hashboard may carry anywhere from 8 chips (on a compact design like the NerdQAxe) to 76 or more (on an Antminer S19 series board). These chips are arranged in a serial daisy chain — a critical design detail for repair. Each chip receives data from the previous chip in the chain and passes results to the next. This means a single dead chip can take down the entire chain, effectively reducing that board’s hashrate to zero even though dozens of other chips are perfectly healthy.

Common ASIC chips you will encounter:

• BM1387 — Bitmain’s 16nm chip used in the Antminer S9 series (63 chips per board, 3 boards per unit). The most widely repaired ASIC chip in history due to the S9’s enormous installed base.
• BM1397 — 7nm chip in the Antminer S17/T17 series. Tighter ball pitch and higher power density than the BM1387.
• BM1398 — 7nm chip powering the Antminer S19 series (76 chips per board). Currently one of the most common chips in active repair queues.
• BM1366 — 5nm chip in the Antminer S19 XP. Smaller die, finer pitch, more demanding rework requirements.
• BM1370 — The chip behind the Antminer S21 generation. Latest-generation BGA packages with extremely tight tolerances.
• MicroBT chips — Used in Whatsminer M30/M50/M60 series. Different packaging and board layout conventions than Bitmain designs.

Voltage Domains and Buck Converters

ASIC chips run at very low voltages — typically between 0.3V and 0.5V per chip, depending on the generation and frequency settings. Because the PSU delivers 12V (or in some designs, a higher bus voltage), the hashboard must step this down dramatically. This is handled by a series of buck converters (DC-DC step-down regulators) that divide the board into voltage domains.

Each voltage domain powers a group of ASIC chips. On an Antminer S19, for example, the 76 chips are divided across multiple voltage domains, each supplied by its own buck converter stage. These buck converters are among the hardest-working components on the board — they handle massive current loads (sometimes 20A+ per domain) in a confined space with limited cooling. When a voltage domain fails, every chip in that domain goes dark.

Key components in each voltage domain:

• Buck converter IC — The switching regulator that steps down voltage. Often a multi-phase design on newer boards.
• Inductors — Store energy during the switching cycle. Typically large ferrite-core components, easily identifiable on the board.
• Output capacitors — Filter the switched output to a stable DC voltage. Usually ceramic or polymer capacitors rated for low ESR at high frequencies.
• MOSFETs — Handle the high-current switching in the buck converter circuit. Thermal stress on these components is a common failure point.
• Current sense resistors — Allow the buck converter IC to measure and regulate output current.

Signal Chain

The signal chain is the nervous system of the hashboard. It carries clock and data signals between the control board connector and every ASIC chip in the chain. The key signals include:

• CLK (Clock) — The timing reference that synchronizes all chip operations. Usually a high-frequency differential signal.
• CO/CI (Chain Out / Chain In) — Serial communication between chips. Data flows from the first chip in the chain to the last, then returns via a separate path.
• RI/RO (Return In / Return Out) — The return path for the serial data chain. Nonce results travel back to the control board via this path.
• RST (Reset) — Resets the entire chip chain. Used during initialization and error recovery.
• BO (Busy/Output) — Status signaling between chips.

These signals pass through coupling capacitors between domains (to handle different voltage references), termination resistors at the ends of signal chains (to prevent reflections), and ESD protection diodes at the control board connector. A break anywhere in this signal chain — whether from a cracked trace, a failed coupling capacitor, or a dead chip — can render the entire chain non-functional.

Supporting Components

Beyond the core chips, buck converters, and signal chain, every hashboard includes:

• Temperature sensors — Usually NTC thermistors placed at strategic locations across the board. The control board reads these to regulate fan speed and trigger thermal shutdowns.
• LDO regulators — Low-dropout linear regulators that provide clean, stable power for logic-level circuits, PLL (phase-locked loop) stages, and I/O buffers on the ASIC chips.
• EEPROM / Configuration Memory — Stores board-specific calibration data, including voltage trim values and chip frequency tables. Corruption of this data can cause a board to behave erratically even when all hardware is functional.
• Decoupling capacitors — Hundreds of tiny ceramic capacitors placed near each ASIC chip to filter high-frequency noise from the power supply. These are critical for stable operation but rarely fail individually.
• Test pads and connectors — Present on most hashboards for factory testing. Invaluable during repair for probing voltage domains and signal paths.

Common Failure Modes

Hashboards fail for specific, identifiable reasons. Understanding the failure modes is half the diagnostic battle. Here are the most common causes of hashboard failure, ranked roughly by frequency.

1. Dead ASIC Chip (Most Common)

A single ASIC chip failure is the most frequent hashboard fault. Because chips are connected in a serial daisy chain, one dead chip breaks the communication path for every chip downstream. Symptoms include: the board powers up but reports zero or severely reduced hashrate, the kernel log shows the chain stopping at a specific chip number, or the board fails chip enumeration entirely.

Chip death can result from thermal stress (cumulative damage from operating near thermal limits), electromigration (microscopic metal migration within the chip over time), manufacturing defects that manifest after burn-in, or voltage spikes from unstable power supplies. The BM1387 chips in S9 boards are particularly susceptible to thermal degradation after years of continuous operation.

2. Voltage Domain Failure

When a buck converter stage fails — whether from a shorted MOSFET, a blown IC, or a failed output capacitor — the entire voltage domain loses power. All chips in that domain go offline. Voltage domain failures often present dramatically: the board may draw abnormally high or low current, and thermal imaging will show a distinct cold zone where the dead domain’s chips are not generating heat.

These failures can cascade. A shorted chip can pull a voltage domain low, overloading the buck converter and destroying it. Conversely, a buck converter failure that sends overvoltage to the chips can kill multiple chips simultaneously. This is why you must always check both the chips and the power delivery components when diagnosing a domain failure.

3. Solder Joint Failure (BGA Cracking)

ASIC chips are typically packaged in BGA (Ball Grid Array) format — hundreds of tiny solder balls form the electrical connection between the chip and the PCB. Thermal cycling (the repeated heating and cooling of operation) creates mechanical stress on these solder joints. Over thousands of cycles, micro-cracks develop, eventually breaking the electrical connection.

BGA solder joint failure is insidious because it can be intermittent. A board may work when cold, fail when hot, then work again after cooling — a classic thermal intermittent. This makes diagnosis challenging without thermal imaging. The Antminer S17/T17 series was notoriously susceptible to BGA cracking due to a combination of high operating temperatures and lead-free solder formulations.

4. Signal Chain Break

Any interruption in the serial communication chain will cause chips downstream of the break to become unreachable. Causes include: a damaged or corroded PCB trace, a failed coupling capacitor between voltage domains, a cracked solder joint on a signal-path component, or physical damage to the board (flex, drop, or improper handling).

Signal chain breaks often manifest as partial chain detection — for example, the kernel log shows 30 out of 76 chips detected, with the chain stopping at a consistent point. The break is typically located at or just after the last detected chip.

5. Thermal Damage

Sustained overheating degrades components progressively. Running a miner with inadequate cooling, blocked airflow, or in ambient temperatures above specifications does not cause instant failure — it accelerates aging of every component on the board. Solder joints weaken, capacitors lose capacitance, chip junctions degrade, and buck converter efficiency drops. By the time symptoms appear, the damage is often spread across multiple components, making repair more complex and costly.

Visual indicators of thermal damage include discolored PCB substrate (the green solder mask turns brownish or dark), darkened or bubbled flux residue around components, and in extreme cases, delamination of the PCB layers themselves.

6. Corrosion and Moisture Damage

Mining environments are often less than ideal — garages, basements, shipping containers. Humidity, condensation, and airborne contaminants can cause corrosion on exposed copper traces, connector pins, and component leads. Corrosion increases resistance, creates intermittent connections, and can eventually eat through traces entirely. Coastal environments with salt air are particularly destructive.

Moisture damage is sometimes visible as white or green crystalline deposits on the PCB surface. In severe cases, electrolytic corrosion will have visibly consumed copper traces or connector pins.

7. ESD Damage

Electrostatic discharge during handling can destroy ASIC chip input/output structures instantly. The damage is invisible — there are no burn marks, no discoloration. The chip simply stops communicating. ESD damage is preventable with proper anti-static procedures, but many home miners handle hashboards without grounding precautions, especially when swapping boards between units or inspecting for visible issues.

The Diagnostic Process

Effective hashboard repair follows a systematic diagnostic process. Jumping straight to chip replacement without proper diagnosis wastes time, money, and risks damaging working components. Here is the process used in D-Central’s repair lab, refined over thousands of board repairs.

For an in-depth guide on one of the most important diagnostic tools, see our Multimeter Guide for ASIC Repair.

Step 1: Visual Inspection

Before powering anything on or connecting any test equipment, examine the board under magnification. A good stereo microscope or high-quality magnifying lamp with LED illumination reveals issues that are invisible to the naked eye.

Look for:

• Burn marks or discoloration — Dark spots on the PCB near components indicate thermal events. Check the component at that location and its neighbors.
• Swollen or cracked capacitors — Electrolytic capacitors with domed tops have failed. Ceramic capacitors with visible cracks must be replaced.
• Corrosion — White, green, or blue deposits on traces, pins, or component leads.
• Physical damage — Cracked PCB, bent connector pins, missing components, scratched traces.
• Solder defects — Cold solder joints (dull, grainy appearance), solder bridges between pads, insufficient solder on BGA corners.
• Flux residue patterns — Darkened or spread flux around a chip may indicate a previous rework attempt.

Visual inspection catches roughly 20-30% of faults immediately. Even when it does not reveal the root cause, it provides context — a board with corrosion damage will be approached differently than one with thermal damage indicators.

Step 2: Multimeter Testing

With the board unpowered, a multimeter in resistance and diode mode reveals the health of voltage domains and individual chips without risk of further damage.

Voltage domain resistance checks: Measure the resistance across each voltage domain’s output. Compare readings between domains on the same board and against known-good reference values for that model. A shorted domain (near-zero ohms) indicates a failed component — usually a shorted chip or MOSFET. An open domain (infinite resistance) suggests a failed buck converter or blown fuse/trace.

Diode mode on ASIC chips: Placing the multimeter in diode mode and measuring across specific pins on each ASIC chip reveals whether the chip’s internal structures are intact. Each chip model has characteristic diode-mode readings — for example, a healthy BM1387 shows approximately 0.35-0.40V on its power pins in diode mode. A chip reading significantly different from its neighbors is suspect. This technique allows you to walk the chain chip-by-chip and identify the faulty one without powering the board.

Continuity checks: Verify signal chain continuity between chips, connector pins to the first/last chip in the chain, and ground connections. An unexpected open circuit points directly to a break location.

Step 3: Thermal Imaging

With the board powered in a test fixture (or temporarily installed in a working miner), a thermal camera reveals the operational state of every component. This is one of the most powerful diagnostic tools available.

Healthy board thermal signature: All ASIC chips should show roughly uniform temperature, with normal variation from airflow patterns. Buck converter components (inductors, MOSFETs) run warm but should not show extreme hot spots.

Dead chip identification: A chip that is significantly cooler than its neighbors is not hashing — it is dead or disconnected. A chip that is significantly hotter than neighbors may be failing (increased leakage current) or receiving excessive voltage.

Voltage domain mapping: Thermal imaging clearly delineates voltage domains. A dead domain shows as a cold zone — a cluster of chips all at ambient temperature while the rest of the board is hot. This immediately narrows the fault to a specific domain’s power delivery.

Hot spot detection: Individual components running abnormally hot — a specific MOSFET, a buck converter IC, a solder joint — point to imminent or active failure.

Step 4: Software Diagnostics

The miner’s own firmware provides critical diagnostic data. Kernel logs record exactly what happens during chip enumeration, frequency negotiation, and hashing operations.

Key information from software diagnostics:

• Chip count — How many chips were detected on each chain. If a chain reports fewer chips than expected, the chain is broken at the point where detection stops.
• Per-chip hashrate — Some firmware versions report individual chip hashrate. A chip hashing at a fraction of its expected rate has degraded.
• Error rates — Hardware errors (HW errors) per chip indicate failing chips. An elevated error rate on a specific chip position is a strong diagnostic signal.
• Temperature readings — On-chip temperature sensors (available on some ASIC generations) complement external thermal imaging.
• Voltage readings — Firmware-reported domain voltages can confirm buck converter health.

For a comprehensive guide to interpreting these logs and the error codes they contain, refer to our Antminer Error Code Reference and Whatsminer Error Code Reference.

Step 5: Test Fixture and Frequency Response

For complex faults that are not resolved by the above steps, dedicated test fixtures allow technicians to power individual voltage domains in isolation, inject test signals into the chain, and measure frequency response characteristics. Test fixtures are model-specific — a fixture designed for Antminer S19 boards will not work on Whatsminer boards.

These fixtures are essential for:

• Isolating which specific voltage domain has a fault when multiple domains appear problematic.
• Testing repaired chips/domains before reassembling the full board.
• Verifying signal integrity through the chain after trace repairs or chip replacements.
• Measuring the actual operating frequency of individual chips to detect degraded performance.

Repair Techniques

Once the fault is identified, the repair itself requires precision soldering skills and the right equipment. Here are the core repair techniques used in professional hashboard repair.

ASIC Chip Replacement

Replacing a failed ASIC chip is the most common hashboard repair. These are BGA (Ball Grid Array) components — there are no through-hole leads to desolder. The chip sits flat on the board with hundreds of tiny solder balls forming the connection underneath.

Removal process:

1. Apply flux around the chip to aid heat transfer and prevent oxidation.
2. Preheat the board from below using a dedicated preheater or the hot air station’s bottom heater. Target substrate temperature of approximately 150-180 degrees Celsius to reduce thermal shock.
3. Using a hot air rework station with appropriate nozzle, heat the chip from above. Target temperature varies by solder type — typically 350-380 degrees Celsius for lead-free solder (SAC305) and 300-330 degrees Celsius for leaded solder (Sn63/Pb37).
4. When the solder reaches liquidus (visible solder reflow), gently lift the chip with vacuum tweezers. Never twist or pry — this risks lifting pads from the PCB.
5. Clean the pads thoroughly with solder wick, flux, and isopropyl alcohol. Inspect under magnification for lifted or damaged pads.

BGA reballing and installation:

1. If using a donor chip (harvested from another board), the old solder must be cleaned from the chip’s BGA pads and new solder balls applied — this is reballing. A model-specific BGA stencil and jig hold the chip while solder paste or preformed balls are applied and reflowed.
2. Apply flux to the cleaned PCB pads.
3. Align the new chip using the stencil marks or visual alignment under magnification. Even a fraction of a millimeter of misalignment can cause bridges or opens.
4. Reflow the chip using the same hot air process — preheat from below, hot air from above, watch for solder reflow, and allow controlled cooldown.
5. After cooling, clean all flux residue and inspect under magnification for solder bridges, insufficient solder, or misalignment.

The entire process requires 15-30 minutes per chip for an experienced technician. Newer chips (BM1366, BM1370) have finer ball pitch and more balls, demanding greater precision and better equipment.

Buck Converter Replacement

When a buck converter IC or its associated MOSFETs fail, the entire voltage domain must be repaired. The process is similar to ASIC chip replacement for QFN-packaged ICs (hot air removal and reinstallation), but MOSFETs and inductors often have exposed thermal pads that require careful attention to ensure proper heat sinking after reinstallation.

Key considerations:

• Always check for a root cause before replacing a buck converter. If a shorted ASIC chip caused the buck converter to fail, replacing only the buck converter will result in immediate re-failure.
• Verify output capacitors in the domain — a failed capacitor can cause voltage instability that destroys the replacement IC.
• After replacement, power the domain at reduced voltage (if your test fixture allows) and verify stable operation before full-power testing.

Passive Component Replacement

Capacitors, resistors, and other passive components are simpler to replace than BGA ICs, but still require precision on hashboards due to the small component sizes (typically 0201, 0402, or 0603 packages).

• Coupling capacitors — These are in the signal chain path. A failed coupling capacitor breaks the chain just like a dead chip. Replacement requires matching the original value (typically in the nanofarad range) and voltage rating.
• Termination resistors — Must be replaced with the exact same resistance value. Incorrect termination causes signal reflections that can prevent chip communication.
• Decoupling capacitors — While individual failures are rare, physically damaged (cracked) decoupling caps must be replaced to prevent shorts.

A fine-tip soldering iron (chisel or conical tip, 0.5mm or smaller) is preferred over hot air for individual passives, as hot air risks disturbing neighboring components.

Trace Repair

Broken or corroded PCB traces can be repaired using several methods:

• Jumper wires — The most reliable method. A thin enameled wire (30-36 AWG) is soldered from pad to pad, bypassing the damaged trace. Route the jumper along the original trace path and secure with UV-cure adhesive or Kapton tape.
• Conductive epoxy — Can bridge short gaps in traces, but has higher resistance than copper and is less mechanically robust. Suitable for temporary repairs or very short trace breaks.
• Copper tape — Useful for repairing wider power traces where jumper wire gauge is insufficient for the current load.

Trace repair is skilled work — the repair must handle the signal frequencies and current loads of the original trace without introducing impedance mismatches (for signal traces) or excessive resistance (for power traces).

BGA Reflow vs. Replacement: The Decision

When an ASIC chip has cracked solder joints (thermal intermittent behavior), there are two options: reflow the existing chip or replace it entirely.

Reflow involves reheating the chip to re-melt the solder balls and (hopefully) re-establish broken connections. It is faster and preserves the original chip. However, reflow is fundamentally a temporary fix — the same thermal stresses that caused the original failure will cause it again. Additionally, each reflow cycle degrades the solder joint quality, meaning subsequent reflowed joints are weaker than the originals.

Replacement with a fresh chip and new solder balls is the definitive fix. It takes longer and requires a donor chip, but the result is a like-new solder joint. For boards that will continue operating in demanding thermal environments, replacement is almost always the better long-term choice.

D-Central’s recommendation: reflow only when you need to confirm that a chip is the fault location (as a diagnostic step) or for boards that will be retired soon. For any board returning to production duty, replace the chip.

Essential Tools and Equipment

Professional hashboard repair demands professional tools. Here is what a well-equipped repair bench requires.

Hot Air Rework Station

The centerpiece of any BGA repair setup. Look for: precise temperature control (within 5 degrees Celsius), adjustable airflow, a range of nozzle sizes to match different chip packages, and a bottom preheater (either built-in or separate).

Recommended stations: Quick 861DW (excellent value for professional work), JBC Advanced series (top-tier precision), Hakko FR-811 (reliable workhorse). Budget options like the Quick 957DW+ can work for occasional repairs but lack the thermal stability needed for high-volume work on fine-pitch BGA components.

Soldering Iron

Required for passive component replacement, jumper wires, and connector work. A temperature-controlled station with interchangeable micro tips is essential. JBC, Hakko FX-951, and Weller WE 1010 are all excellent choices. Stock tips in fine chisel (1mm and 0.5mm) and conical (0.5mm) shapes.

Digital Multimeter

Your most-used diagnostic tool. A Fluke 87V or equivalent gives you the accuracy and durability needed for daily repair work. Essential functions: DC voltage, resistance (to 0.1 ohm resolution), diode mode, and continuity beeper. A good multimeter is a non-negotiable investment — cheap meters drift, give inconsistent readings, and waste your diagnostic time. See our full Multimeter Guide for ASIC Repair for recommended techniques and models.

Thermal Camera

A thermal imaging camera transforms hashboard diagnosis from guesswork to precision. At minimum, use a smartphone-attached module like the FLIR ONE Pro or Seek Thermal CompactPRO. For bench use, a standalone thermal camera with higher resolution (like the FLIR E5-XT or similar) is faster and more detailed. The investment pays for itself in reduced diagnostic time on the first few boards.

BGA Reballing Equipment

Model-specific BGA reballing stencils, a reballing jig to hold the chip and stencil in alignment, solder balls or solder paste (matched to the correct alloy — SAC305 for lead-free boards, Sn63/Pb37 for leaded), and a small reflow setup (the hot air station can serve double duty here).

Additional Essential Equipment

• Microscope or magnifying lamp — Stereo microscope (10-40x) for inspection work. At minimum, a quality LED magnifying lamp with 5x or greater magnification.
• Flux — Rosin-based no-clean flux for general soldering; tacky flux paste for BGA work. Amtech NC-559 or MG Chemicals 8341 are popular choices.
• Solder paste — For BGA reballing and certain component replacements. Use the correct alloy for the board you are working on.
• Solder wick and desoldering pump — For cleaning pads after component removal.
• Isopropyl alcohol (99%+) — For cleaning flux residue and board contamination.
• ESD protection — Anti-static mat, wrist strap, and grounded bench. Non-negotiable when handling ASIC chips.
• Test fixture / daughter board — Model-specific board that allows powering and testing a hashboard independently of the full miner chassis.
• PCB holder — An adjustable holder (like a Stickvise or similar) to secure the hashboard during work.
• Kapton tape — Heat-resistant polyimide tape for masking areas adjacent to rework zones.

Model-Specific Repair Considerations

While the fundamental principles apply to all hashboards, each model generation brings its own challenges. Here is what to expect when working on the most common boards in today’s repair queues.

Antminer S9 (BM1387, 63 Chips, 3 Chains)

The S9 is where most repair technicians learn their craft. The BM1387 chip has a relatively generous ball pitch, making rework more forgiving. The 63-chip chain on each board means faults are common simply due to the number of components, but the chain topology is well-documented and reference resistance values are widely available.

Common S9-specific issues: thermal degradation of BM1387 chips after years of continuous operation, corroded heatsink mounting hardware causing uneven cooling, and failing buck converters due to age. Replacement BM1387 chips are still readily available and inexpensive, making S9 repair economically viable even today — though you should weigh repair cost against the board’s diminishing mining profitability. For model-specific repair details, see our Antminer S9 Repair page.

Antminer S17/T17 (BM1397)

The S17 generation is infamous in the repair community for its high failure rate. The BM1397 chip runs hot, the board layout pushes thermal limits, and the lead-free solder used in production is prone to cracking under thermal cycling. BGA reflow failure is the single most common issue on S17/T17 boards.

Repairing S17 boards requires excellent hot air control — the tighter ball pitch compared to the S9 leaves less room for error during rework. Many shops see S17 boards return for repeat failures, which is why D-Central recommends chip replacement over reflow for these boards whenever possible. See our dedicated Antminer S17 Repair page for specific guidance.

Antminer S19 Series (BM1398, 76 Chips)

The S19 series (S19, S19 Pro, S19j Pro, S19 XP) represents the current workhorse of the mining industry and consequently the highest volume of boards in repair queues. The BM1398 chip is a 7nm design with a denser BGA package than the BM1387.

S19-specific considerations: 76 chips per board means a longer chain and statistically more potential failure points; the boards use multi-phase buck converters that are more complex to diagnose and repair; heatsink attachment methods vary between S19 sub-models, affecting thermal performance and repair access. The S19 XP variant, using the BM1366 5nm chip, has even finer pitch and demands the most precise rework skills in the Antminer lineup. Visit our Antminer S19 Repair page for details.

Antminer S21 (BM1370)

The latest generation presents the greatest repair challenge. The BM1370 is fabricated on an advanced process node with extremely fine BGA pitch. Rework tolerances are tight enough that even experienced technicians report higher first-attempt failure rates compared to older models.

S21 repair demands top-tier equipment (a quality hot air station with precise nozzle sizing is non-negotiable), fresh solder paste and flux, and a very steady hand. Test fixtures for S21 boards are still limited in availability. Given the high cost of S21 boards and the difficulty of repair, this is firmly professional-only territory. See our Antminer S21 Repair and Antminer S21 Repair Guide for current information.

Whatsminer M30/M50/M60 Series

MicroBT’s Whatsminer boards differ significantly from Bitmain’s architecture. The board layouts are different, the voltage domain topology varies, and the integrated PSU design on some models means the power delivery path includes additional components not present on Antminer boards. Additionally, MicroBT uses different ASIC chip packages with their own rework profiles.

Whatsminer-specific challenges: less publicly available documentation compared to Antminer, fewer aftermarket test fixtures, and different firmware diagnostic interfaces. Technicians experienced primarily on Antminer boards should approach Whatsminer repair with additional caution and study. See our Whatsminer M30S Maintenance & Repair Guide for model-specific procedures.

When to Repair vs. When to Replace

Not every dead hashboard deserves repair. A clear-eyed cost-benefit analysis saves money and frustration. For a full breakdown of current repair pricing, see our ASIC Repair Pricing Guide.

Repair Makes Sense When:

• Single chip failure on a current-gen board — An S19 or S21 hashboard is worth hundreds of dollars. A single chip replacement at $50-$150 is clearly worthwhile.
• Known, isolated fault — A specific buck converter failure, a cracked coupling capacitor, or a single dead chip. The repair is targeted, the success rate is high (85%+ for single-chip replacements by a skilled technician), and the cost is a fraction of replacement.
• Board is still profitable to mine with — If the repaired board’s mining revenue exceeds its electricity cost with reasonable margin, repair makes financial sense.
• You have multiple units of the same model — Donor boards for chip harvesting reduce repair costs significantly. Three dead boards can often yield two working boards.

Replace (or Retire) When:

• Multiple failure modes present — A board with dead chips, corroded traces, AND failing buck converters has cumulative damage that makes reliable repair unlikely and costly.
• Repeated failure after repair — A board that comes back within weeks of repair likely has an underlying issue (substrate degradation, widespread BGA fatigue) that targeted repair cannot fix.
• Obsolete hardware — Spending $100+ to repair an S9 hashboard that generates $0.10/day at current difficulty is not a sound investment, unless you are using it as a Bitcoin Space Heater where the heat output justifies the electricity cost.
• Physical damage to PCB substrate — Cracked boards, delaminated layers, or severe warping compromise the structural integrity needed for reliable BGA connections.
• Repair cost exceeds 50-60% of replacement — At this threshold, the risk-adjusted economics favor a new board with full warranty over a repaired board with limited guarantee.

Success Rates by Failure Type

Based on D-Central’s repair data across thousands of boards:

• Single ASIC chip replacement: 85-95% success rate when properly diagnosed.
• Buck converter replacement: 80-90% success rate (lower because buck converter failure often indicates a secondary fault).
• Signal chain repair: 75-85% success rate (depends on the complexity of the break and whether trace repair is needed).
• BGA reflow (without chip replacement): 50-70% initial success, but only 30-50% remain stable after 6 months.
• Corrosion damage repair: 60-75% success rate (highly dependent on extent of corrosion).
• Multi-fault boards: 40-60% success rate, and repair cost escalates quickly.

DIY vs. Professional Repair

We believe in empowering home miners — but we also believe in honesty about where the line falls between DIY-friendly and send-it-to-a-pro. For a detailed breakdown, see our Professional ASIC Repair vs DIY guide.

What a Home Miner Can (and Should) Do

• Visual inspection — Learn to identify burn marks, corrosion, swollen caps, and physical damage. This costs nothing and takes minutes.
• Basic multimeter testing — With our Multimeter Guide for ASIC Repair, you can check voltage domains and identify obvious shorts or opens.
• Firmware diagnostics — Read kernel logs, check chip counts, monitor error rates. This is all software and requires no physical tools beyond SSH access to your miner.
• Fan and cable replacement — Swapping fans, power cables, and data cables requires no soldering skill.
• Heatsink maintenance — Cleaning dust, replacing thermal paste/pads, and ensuring proper heatsink contact can resolve thermal issues without any board-level work.
• Control board swapping — If you have a spare control board or can borrow one, swapping it in isolates whether the issue is the control board or the hashboard.
• Board-level connector cleaning — Cleaning hashboard connector pins with isopropyl alcohol and a soft brush can resolve intermittent connection issues.

What Requires Professional Equipment and Skills

• ASIC chip replacement — BGA rework demands a hot air station, preheater, proper nozzles, and practiced technique. One slip destroys the chip or lifts pads from the PCB.
• Buck converter and MOSFET replacement — High-current components with thermal pads require precise rework.
• Trace repair — Requires microscope-level precision and knowledge of signal integrity considerations.
• BGA reballing — Requires model-specific stencils, a jig, and experience to achieve consistent results.
• Multi-fault diagnosis — When multiple failure modes overlap, experience and test fixtures are needed to untangle cause from effect.

The threshold is simple: if the repair requires removing or installing a BGA component, it requires professional equipment and skills. Everything up to that point is accessible to a careful, methodical home miner willing to learn.

D-Central’s Repair Lab

At D-Central, hashboard repair is not a side service — it is a core competency we have been building since 2016. Our repair lab is equipped for every level of hashboard repair, from routine chip replacements to complex multi-fault board recovery.

Equipment

Our bench includes professional-grade hot air rework stations, precision soldering stations, FLIR thermal cameras, digital microscopes, model-specific test fixtures for all current Antminer and Whatsminer generations, a full inventory of replacement ASIC chips and passive components, and dedicated ESD-safe workstations. We invest continuously in our tooling because the quality of the repair is directly determined by the quality of the equipment.

Process

Every board that comes to us follows the full diagnostic process outlined in this guide — visual inspection, multimeter testing, thermal imaging, software diagnostics, and test fixture verification. We diagnose first, quote second, and repair only with your approval. No surprises.

After repair, every board undergoes a minimum 24-hour burn-in test on our test rigs, running at full hashrate to verify stability before it ships back to you. We do not consider a repair complete until the board has proven itself under load. For details on what to expect, read our ASIC Repair Process guide.

What We Repair

We service hashboards from all major manufacturers:

• Bitmain Antminer: S9, S17, T17, S19 (all variants), T19, S21, T21, S21 XP, S21 Pro, L3+, L7, D7, and more
• MicroBT Whatsminer: M20, M21, M30, M31, M32, M50, M53, M56, M60, M63 series
• Canaan Avalon: 1047, 1146, 1166, 1246, 1346, 1366, 1466 series
• Innosilicon: T1, T2, T2T, T3, A4+, A10 series

Explore our full ASIC Repair services or get a repair cost estimate with our ASIC Repair Cost Estimator.

Turnaround and Warranty

Standard turnaround is 5-10 business days from the time we receive your board, including diagnostics, repair, and burn-in testing. Rush service is available for production-critical situations. All repairs carry a warranty against the same fault recurring — because we fix the root cause, not just the symptom.

We serve miners across Canada and the United States. Ship your board to our facility, and we handle the rest. Contact us for a repair quote or to discuss your specific situation.

Conclusion: Repair is a Skill Worth Respecting

Hashboard repair sits at the intersection of electronics engineering, materials science, and hands-on craftsmanship. It is not glamorous work — it is painstaking, detail-oriented, and demands genuine expertise. But it is work that matters. Every successfully repaired hashboard is hashrate returned to the network, value recovered for its owner, and one less board in a landfill.

The decentralization of Bitcoin mining depends on accessible repair. When a home miner can understand what went wrong with their hashboard, make an informed decision about DIY vs professional repair, and get their machine back online — that is decentralization in action. That is the Mining Hacker ethos.

Whether you are diagnosing your first dead chain or running a repair bench of your own, we hope this guide serves as a solid foundation. And when you hit a board that needs more than your bench can handle, D-Central is here. We have been doing this since 2016. We are not going anywhere.

Frequently Asked Questions