A Bitaxe is a single-ASIC, open-source Bitcoin miner: one Bitmain BM1370 or BM1368 chip on an ESP32-S3 board with its own voltage regulator, fan controller and OLED. Most faults trace to four areas — power-on, the VRM/power stage, the ASIC itself, or thermals. This hub routes your symptom to the right fix and tells you what is DIY versus bench work.

D-Central has been repairing ASIC mining hardware in-house in Laval, Quebec since 2016. The Bitaxe and its cousins (NerdAxe, NerdQAxe) are some of the most repairable miners ever made — the whole design is open-source, the schematics are public, and almost every part is a standard, replaceable component. This guide is the parent page for our Bitaxe repair series. Start here, identify your symptom, then jump to the deep-dive that matches it. If a fix turns into BGA rework or short-finding, D-Central repairs Bitaxe and open-source boards in-house.

What a Bitaxe actually is (board anatomy in 60 seconds)

Unlike an Antminer — which daisy-chains 100+ ASICs across three hash boards — a Bitaxe is a single small PCB running one hashing chip. That simplicity is exactly why it is so diagnosable: there is no chain to bisect, no control board to swap, and one bad chip means a dead miner rather than a partial chain. The board breaks down into five functional blocks:

• ESP32-S3 controller — the brains. It runs the open-source AxeOS/ESP-Miner firmware, serves the web UI over Wi-Fi, talks to the ASIC over UART, and manages every peripheral over a shared 400 kHz I2C bus. This is a microcontroller, not a Linux SoC like a Bitmain control board.
• The hashing ASIC — a single Bitmain SHA-256 die. The Bitaxe Gamma runs a BM1370 (the same TSMC 5 nm silicon as the Antminer S21 Pro) at roughly 1 TH/s stock (~525 MHz, ~1.1–1.2 V core). The Bitaxe Supra runs a BM1368 (S21-class 5 nm), and the older Ultra runs a BM1366. These are no-PIC, 5 nm parts — there is no separate PIC microcontroller to fail like there was on S9/S17-era boards.
• The power stage (VRM/PMIC) — steps the board input down to the sub-1.2 V the ASIC core needs at tens of amps. The Gamma uses an integrated TI TPS546D24A PMBus buck regulator that the firmware programs directly (set-point, over-current and over-voltage limits, telemetry). Earlier Bitaxe boards used a discrete buck with a DS4432U DAC setting the feedback point and an INA260 monitoring volts/amps/watts. Voltage is regulated per power domain, not per individual core.
• Thermal management — an EMC2101 PWM fan controller (I2C address 0x4C) drives the fan, reads its tachometer, and senses the ASIC’s on-die thermal diode. The firmware enforces hard limits: it throttles the ASIC around 75 °C, treats 90 °C as the chip ceiling, and watches the regulator’s own temperature (throttle ~105 °C, hard limit 145 °C).
• OLED + connectors — a small SSD1306-class I2C OLED shows hashrate, temperature and the pool/IP, plus the USB-C/DC input, the fan header, and the antenna. These are the cheapest, most replaceable parts on the board, and a blank OLED is often a clue rather than the fault itself.

Want the full part-by-part map with test points and what each chip does? See Bitaxe Board Anatomy.

The four failure categories

Almost every Bitaxe that lands on our bench fits one of these buckets. Knowing the bucket is half the repair.

1. Won’t power on — no OLED, no fan spin, no Wi-Fi AP. The fault is upstream of the ASIC: the input jack/USB-C cable, the input protection, the 5 V/3.3 V rails, or the ESP32-S3 itself. Underpowered USB ports cause far more “dead” Bitaxes than actual hardware faults.
2. Powers on but won’t hash — OLED and Wi-Fi work, but AxeOS reports 0 chips / no ASIC found or sits at 0 GH/s. On a single-chip board this points at the ASIC, its core power, the UART link between the ESP32-S3 and the chip, or a firmware/pool misconfiguration. This is the most common real failure.
3. Overheating, throttling or VRM faults — the board hashes, then drops, restarts, or runs far under expected hashrate; or the regulator runs hot, trips an over-current/over-voltage fault, or releases the magic smoke near the power stage. Thermals and the VRM are tightly coupled, so we treat them together.
4. Intermittent / degraded — random restarts, rising rejected-share rates, hashrate that decays as the board heats up, or a high invalid-nonce count. Usually a marginal solder joint, a tired fan, dust-blocked airflow, or a voltage/frequency set too aggressively for that particular chip’s silicon lottery.

Symptom-to-page router

What you see	Most likely area	Go to
Dead board — no OLED, no fan, no Wi-Fi access point	Input power, rails, ESP32-S3	Bitaxe Won’t Power On
OLED/Wi-Fi up, but “0 ASIC” / “ASIC not found” / 0 GH/s	ASIC, core voltage, UART link	Bitaxe Not Hashing
Hashes then drops/restarts, or runs hot and throttles	Thermals + VRM	Bitaxe Overheating & VRM
VRM hot to the touch, voltage fault, smoke near regulator	Power stage (TPS546/DS4432U/INA260)	Bitaxe Overheating & VRM
Low hashrate, climbing rejects, intermittent restarts	Tuning, fan, solder joints	Bitaxe Not Hashing
“What does this chip/test point do?”	Reference	Bitaxe Board Anatomy

For model-specific fault patterns and known error states, cross-check the ASIC Fault Finder database. If your issue is really a tuning question — too much voltage for the frequency, or the wrong set-point for your cooling — the ASIC Power Profiles database lists safe operating points.

What you can fix at home vs. what needs the bench

Be honest about the toolset a job needs. A Bitaxe is forgiving, but the ASIC and VRM are fine-pitch parts under a heatsink.

DIY-fixable (no microsoldering)

• Power problems: try a known-good 5 V/3 A USB-C supply and cable before assuming a hardware fault. A flaky cable or a 2.5 W laptop port mimics a dead board.
• Reflash AxeOS / firmware via the web flasher or USB to rule out corrupted firmware after a bad update or brownout.
• Reseat the fan connector, replace a seized fan, and clear dust — a stalled fan is the number-one cause of thermal shutdowns.
• Dial voltage and frequency back to stock. Over-aggressive tuning is the most common self-inflicted “fault,” and AxeOS’s autotuner computes its targets at runtime, so a clean reset to defaults often restores stability.
• Reseat the OLED and check for an obviously cracked solder joint or bent header.

Send it to the bench

• Replacing the BM1370/BM1368 — it is a BGA part requiring a reball, hot-air rework, stencil, and a known-good reference board for verification.
• VRM/PMIC failure — diagnosing and swapping a TPS546D24A (or a discrete buck, DS4432U, or INA260) needs PMBus telemetry, a bench supply with current limiting, and short-finding gear.
• Short-finding a board that current-limits at power-on — this is thermal-camera and milliohm-meter work, not multimeter guesswork.
• ESP32-S3 or connector-level microsoldering, and any board that smoked near a power component.

Why D-Central repairs Bitaxe boards in-house

We are Bitcoin mining hackers, not a swap-and-bill shop. Because the Bitaxe is open-source, we work from the same public schematics the community publishes — full credit to skot and the Open Source Miners United (OSMU) community whose work made this platform exist. That means component-level diagnosis: we find the actual failed part, replace it, and verify the board hashes at a sane operating point before it ships back. We repair Bitaxe, NerdAxe, NerdQAxe and other open-source boards on the same bench we use for Antminer hash boards.

If a board is beyond economical repair, a fresh unit is often the better call — see the Bitaxe for a hand-built replacement, and the Bitaxe Hub for setup, tuning and firmware guides. When you are ready for a repair quote, D-Central repairs Bitaxe and open-source boards in-house — send us the symptom and we will tell you whether it is a five-minute reflash or a bench job.

Continue to the deep-dive guides

• Bitaxe Not Hashing — 0 ASIC found, 0 GH/s, low hashrate and rejects.
• Bitaxe Won’t Power On — dead board, no OLED, input and rail diagnostics.
• Bitaxe Overheating & VRM — thermal throttling, fan faults, and power-stage failures.
• Bitaxe Board Anatomy — every block, chip and test point explained.