Skip to content

Bitcoin accepted at checkout  |  Ships from Laval, QC, Canada  |  Expert support since 2016

ASIC Repair Parts & Common-Failure Components

When an Antminer hash board fails, the fault is almost always one of a small set of component classes. This database maps every replaceable part — ASIC chip, LDO, boost stage, EEPROM, PIC, crystal and more — to its role, how it fails, the diagnostic signal that confirms it, and how hard it is to replace. Need the pin-level wiring behind these parts? See the ASIC power & connector pinout reference.

Quick answer

When an Antminer hash board dies, the fault is almost always one of a small set of component classes — a shorted LDO, a failed boost stage, a cracked ASIC BGA joint, a bad clock crystal, a corrupt EEPROM. This catalog maps each of the 18 replaceable parts on a hash board (and its control-board interface) to its role, package, where it lives, how it fails, the root cause, the exact diagnostic signal to confirm it, and how hard it is to replace — every row cited to the D-Central repair bench reference. It is the component-level companion to our repair-tools and diode-voltage references.

Use it to turn a symptom into a part: read the failure mode and diagnostic-signal columns to localize the fault, then the difficulty column to decide DIY-vs-bench. Two facts worth pinning: (1) the PIC supervisor lives ONLY on S19-generation boards — it was removed on BM1368 (S21/T21) and later, so an S21 has no PIC to blame; (2) anything graded Expert (ASIC BGA reball, PCB trace repair) is microsoldering — send it to a bench. Designators (U-numbers) vary by board revision; confirm against the model schematic. Free CSV/JSON under CC BY 4.0.

Download CSV Download JSON REST API →

ComponentCommon failure modeDiagnostic signalReplace difficulty
Mining ASIC chip (BM13xx / BM14xx) Hash board
QFN (QFN-34 on BM1397). Generations: BM1387(S9), BM1391(S15), BM1393(S17), BM1397(S17+/e), BM1398(S19/Pro/T19/j), BM1366(S19 XP/K Pro), BM1368(S21/T21), BM1370(S21 Pro/XP/+)
Used in: All Antminer hash boards (model-specific chip)
Dead/open chip (chain breaks at chip N); shorted chip (0 chips detected + domain V collapses); low nonce rate (degraded transistors); signal-forwarding fail (cold solder on CO/CLKO/NRSTO); thermal shutdown / BGA solder crack (intermittent dropout)
Cause: ESD, thermal runaway/power surge, transistor degradation, thermal cycling/vibration, poor thermal paste
Chip-enumeration count per chain (PT1); domain voltage measurement; per-chip nonce count (PT3 sweep); thermal camera cold-spot for cracked BGA; probe CO/CLKO/NRSTO at chip I/OExpert
Expert (hot-air BGA removal + 0.4mm SAC reball + precise reflow; highest-skill repair)
LDO regulator (per-domain low-dropout) Hash board
SOT-23 (or 4-pin model-specific); 0.8V (all models), 1.2V (S21/S21 XP), 1.8V (S19 series)
Used in: Every Antminer hash board; quantity scales with domain count
LDO short (domain voltage near zero); LDO open (domain voltage high/absent)
Cause: Component failure / surge; cracked solder
Multimeter on domain VDD test point: input present but output absent = LDO failure; unpowered domain-impedance check (low=short)Med-High
Medium-High (small SOT-23 hot-air rework; identify correct rail variant)
Boost converter module / boost MOSFET Hash board
Pre-built boost module PCB OR discrete boost MOSFET + inductor; 18V-class module is a common drop-in part (designators Q9/U206/U146/U202)
Used in: S19 and S21 series hash boards (boost output model-specific)
Boost circuit failure -> 0 chips, no domain voltages
Cause: Boost MOSFET failure, inductor burn
Measure boost output at output capacitor (S19 ~19V / S19 Pro ~20V C69 / S21 ~25V / S21 XP ~21V); near-zero input resistance = bus shortHigh
Medium (swap pre-built module) to High (discrete MOSFET rework)
MP2019 DC-DC buck converter Hash board
TSOT-23-8 (designators U166/U200 on S21; U146/U202 on S21 XP)
Used in: S21, S21 XP (high domains 11-12 / 12-13)
DC-DC buck failure -> high-voltage domain dead
Cause: MP2019 (or similar) component failure
Measure buck output (~2.0-2.5V) at the affected high domainHigh
High (TSOT-23-8 fine-pitch hot-air rework)
Power MOSFET (input power switch) Hash board
Multi-pin power MOS (gate=pin1, source=pin4, drain=pin8); exact part per board schematic
Used in: Antminer hash boards (power input stage)
Power MOS short -> board draws excessive current
Cause: Surge, thermal
Resistance between pins 1,4,8; short between any two pins = MOSFET failure; near-zero input-connector resistance = bus shortMed-High
Medium-High (power-package hot-air; large thermal mass)
Boost/buck power inductor Hash board
Large SMD shielded power inductor (value model-specific; not enumerated in Bible — see uncertainties)
Used in: S19/S21 boost (and S21 buck) stages
Inductor burn / open -> boost stage produces no/low output (0 chips, no domain voltages)
Cause: Overcurrent from a failed/shorted switching MOSFET; thermal stress
Visual: burnt/discolored inductor; boost output absent at output cap; bus short on input resistanceMedium
Medium (large pads, high thermal mass; needs strong hot-air/preheat)
Filter / decoupling capacitor Hash board
1uF 0402 16V X5R (decoupling); 22uF 0603 6.3V X5R (LDO filtering); 100nF 0402 (CLK coupling)
Used in: All models (large quantity per board)
Filter capacitor short -> domain voltage sags
Cause: Capacitor dielectric breakdown
Impedance measurement on the affected domain (sagging V + low impedance)Medium
Medium (small 0402/0603 SMD; identify failed cap among many)
Signal-chain resistor Hash board
33 ohm 0603 +/-1% 1/20W (signal chain); 10K 0402 +/-1% 1/16W (pull-up/down)
Used in: All models; concentrated at domain boundaries
CLK-resistor cold joint -> chain breaks at domain boundary; CI/BO resistor fault -> chain breaks mid-domain
Cause: Cold/cracked solder, manufacturing defect, thermal cycling
Probe CLK before/after the resistor; probe CI/BO at each chip; oscilloscope at the boundaryMedium
Medium (tiny 0402/0603; value/placement ID is the hard part)
Crystal oscillator (CLK source) Hash board
SMD passive crystal, 25 MHz, designator Y1
Used in: All models (Y1 25MHz)
Crystal failure -> 0 chips, no CLK
Cause: Cracked crystal, cold solder
Oscilloscope on CLK test point (no/abnormal 25MHz waveform)Medium
Medium (small SMD hot-air; verify oscillation after)
Level-shifter IC Hash board
Small fine-pitch logic IC (designators U1, U2, ... per model)
Used in: S19 (U1,U2) and S21 family (11-12+ units)
Level-shifter failure -> partial chain stops at a domain boundary
Cause: IC failure, voltage loss
Signal probe at shifter input vs output (signal present in, absent out)High
High (fine-pitch IC hot-air rework)
Temperature sensor IC Hash board
LM75A SOIC-8 I2C (S19, L9); TMP1075 SOT-563 I2C (newer); NCT218 (various). S19: U4/U6/U7/U8; S21: U5(inlet)/U7(outlet)
Used in: All models (2-4 sensors per board)
Temp sensor failure -> thermal protection trips falsely; I2C bus stuck -> multiple subsystems fail
Cause: Sensor IC failure, I2C stuck/SDA-SCL shorted
I2C bus scan + sensor read; oscilloscope on SDA/SCLMed-High
Medium-High (small I2C IC; SOT-563 is very small)
EEPROM (board identity memory) Hash board
SOIC-8 I2C; S19 U5, S19 Pro U10, S21 U6 (model-specific contents)
Used in: All models
EEPROM corruption -> board not recognized; EEPROM mismatch -> mixed boards won't hash
Cause: Write failure, I2C bus fault, wrong/cloned data
EEPROM reader/programmer (CH341A); code-editor comparison against a known-good boardMedium
Medium hardware swap; data side requires correct dump + reflash (CH341A / hash-board code editor)
PIC microcontroller (S19-generation) Hash board
PIC16F1704 SOIC-14; S19 U3, S19 Pro U6
Used in: S19 series only (NOT S21/S21 Pro/XP/+ — PIC removed in BM1368)
PIC failure -> 0 chips, no 3.3V output
Cause: Programming/firmware corruption, ESD
Measure U3/U6 pin 2 for 3.3V (absent = PIC fault); chain dead at power-upMedium
Medium hardware (SOIC-14) + reprogram with PICkit3 and correct firmware image
18-pin signal cable & board connector Hash board <-> Control board interface
18-pin flat ribbon; board-side connector (pinout: 1 GND, 2 VCC_3V3, 3 RST, 4 SDA, 5 SCL, 7 TX/CI, 8 RX/RO)
Used in: All S19/S21 boards
Connector damage / cable fault -> intermittent or no communication with the board
Cause: Wear, corrosion, bent pins
Inspect connector pins; signal-probe across the cable; swap with a known-good cableLow
Low (swap ribbon) / Medium (reflow or replace board connector)
BGA solder balls (reballing consumable) Hash board (consumable for chip replacement)
0.4mm lead-free Sn/Ag/Cu (SAC) solder balls
Used in: BGA reballing on all models
N/A (consumable) — wrong size/alloy or poor reflow -> bridged/open BGA joints, intermittent dropout
Cause: Process error during reball
Post-repair: thermal camera + chip enumeration to confirm all joints madeExpert
Expert-process (part of the ASIC chip replacement workflow)
Thermal interface material Hash board (consumable)
Fujipoly SPG-30B (or equivalent high-conductivity TIM)
Used in: All models
Dried/insufficient TIM -> thermal shutdown, chip intermittently drops out
Cause: Age, heat-cure, prior poor application
Thermal camera hotspot + elevated chip temp read; visual on disassemblyLow
Low (clean + reapply during routine maintenance)
Heatsink & retention clips Hash board (mechanical)
Aluminum heatsink + spring clip (model-specific)
Used in: All models
Heatsink detachment -> thermal shutdown
Cause: Dried thermal paste, clip failure/loss
Visual + thermal camera (detached/hot region)Low
Low-Medium (mechanical; re-clip + re-paste). CAUTION: never touch black meter probe to the heatsink (short risk)
PCB trace / via (board substrate) Hash board (substrate)
Multilayer PCB copper (jumper-wire repair where damaged)
Used in: All models
Trace damage -> signal chain break or domain open; corrosion from liquid damage
Cause: Physical damage, corrosion, delamination
Visual inspection + continuity test; via integrity check under magnificationExpert
Expert (microsoldering / jumper-wire trace repair)

Source: D-Central Mining Bible (HASHBOARD_DIAGNOSTICS) — every row carries a section/line citation in the CSV/JSON. Companion to the repair tools database, the diode & voltage reference and the ASIC chip reference. Need it done? See ASIC repair services.