A Bitaxe that shows no OLED, no Wi-Fi access point and no web UI almost always has one of four faults: no clean 5V reaching the board, an ESP32-S3 that won’t boot a corrupt AxeOS, a buck converter that never raises the ASIC core rail, or blown input protection from a short. This guide isolates which one — and whether you can fix it at your desk or need a bench.

How a healthy Bitaxe powers up (so you know what’s broken)

The Bitaxe is open-source hardware (CERN-OHL-S) running the open AxeOS firmware (ESP-Miner, GPL-3.0) we owe to skot and the Open Source Miners United community. Knowing its boot order tells you exactly where a dead unit stalls. A normal power-up runs like this:

1. 5V arrives at the USB-C port or the DC barrel jack.
2. An onboard regulator drops that to 3.3V to feed the ESP32-S3-WROOM-1 (N16R8) module — the board’s brain.
3. The ESP32 boots AxeOS and lights the OLED splash. If the device is unconfigured, it broadcasts a Wi-Fi access point named Bitaxe_XXXX with a captive setup portal.
4. Firmware initialises the core-voltage regulator (the TPS546 buck on current boards, or a DS4432U DAC on older ones) and brings the ASIC core rail up to roughly 1.1–1.2V.
5. It pulses the ASIC reset line (GPIO), opens the UART at 115200 baud, enumerates the chip, then ramps frequency.

The key insight: the Bitaxe has two independent power islands — the 3.3V ESP32 logic and the buck-regulated ASIC core. A totally dead board (no OLED, no AP, no USB serial device) is a logic-island or input problem. A board that boots and shows the OLED but never hashes is an ASIC-island problem and belongs on our ASIC fault finder instead.

Symptom triage — read the LEDs, OLED and USB first

What you see	Most likely cause
No LED, no OLED, no Bitaxe_XXXX AP, and your computer never sees a USB serial device	No 5V reaching the board, dead ESP32, or blown input protection / short
USB serial device does enumerate, but OLED stays blank and no Wi-Fi AP appears	ESP32 is powered but AxeOS isn’t running — corrupt firmware, boot loop, or an unsupported module
OLED and Wi-Fi AP appear, but hashrate stays at 0 GH/s	Logic is fine; the fault is on the ASIC/buck side — see the sibling guide below
Web UI loads blank or half-broken after a firmware update	Corrupt www.bin — recover at http://<ip>/recovery

Cause 1 — USB-C / 5V input problems

This is the most common “dead Bitaxe” and the easiest to rule out. Work through it before opening anything:

• Cable and port. Use a known-good data-and-power USB-C cable. Charge-only cables and some quick-charge bricks will not enumerate the ESP32. Try a different host port.
• The USB-C handshake quirk. Some Bitaxe Gamma and earlier boards can’t negotiate power on certain native USB-C host ports. The documented workaround is a plain USB-A to USB-C adapter — it forces legacy 5V and often revives a board that looked dead.
• Barrel jack vs USB. Many boards gate the ASIC buck on the DC barrel-jack plug-sense line, but the ESP32 should still boot on USB power alone. If nothing comes alive on either source, the fault is upstream of the regulators.
• Input voltage window. The rail must sit near 5V. AxeOS’s own self-test rejects an input more than 10% off nominal, and the TPS546 latches off if its input sags below roughly 4.5V. Measure 5V at the input pads before chasing anything else.

Cause 2 — ESP32-S3 not booting / corrupt AxeOS

If the board enumerates as a USB serial device but the OLED never lights and no access point appears, the ESP32 is powered yet AxeOS isn’t running. The usual culprit is corrupt firmware after a bad flash, a power loss mid-update, or an incompatible module.

• Confirm the module. Stock AxeOS only runs on the ESP32-S3-WROOM-1 N16R8 (16MB flash, 8MB octal PSRAM). A swapped module without PSRAM, or with quad-SPI PSRAM, will not boot the firmware.
• Re-flash AxeOS. The fastest fix is a clean reflash over USB. Use the D-Central web flasher at d-central.tech/flash/ — it runs in Chrome over WebSerial with nothing to install — or the community esptool-js / Bitaxe web flasher.
• Match the image to the hardware series. This matters: 100 = Max (BM1397), 200 = Ultra (BM1366), 400 = Supra (BM1368), 600 = Gamma (BM1370). The BM1368 and BM1370 are both TSMC 5nm S21-class silicon — flashing a 600 image onto a 400 board will load the wrong ASIC init and the chip will never start.
• Procedure. Connect USB; if the chip won’t enter download mode, hold the BOOT button (GPIO0) while plugging in. Flash the factory image to address 0x0 at 921600 baud, wait for the console to say “Leaving…”, press RESET, and let the on-board self-test run.
• Half-bricked web UI. If the firmware boots but the page is blank after an AxeOS update, the front-end www.bin is corrupt — browse to http://<ip>/recovery and re-upload it.

Cause 3 — buck/VRM not bringing up the core rail

The TPS546 (or DS4432U on older boards) steps 5V down to the ASIC core voltage. That regulation is done per voltage domain — the chip’s hash cores are stacked in series voltage domains, not trimmed per individual chip. A buck fault has a specific signature:

• It usually still lets the ESP32 boot. You typically do get the OLED and the Wi-Fi AP, but the ASIC won’t initialise and you see 0 GH/s. So a completely dead board is rarely a simple buck fault.
• The exception is a hard short on the core rail, which drags the whole 5V input down and kills everything at once.
• Expected reading. A healthy core rail measures about 1.0–1.3V once firmware enables the buck — that is the exact window AxeOS’s self-test checks. Read 0V there, or watch the input current spike and then collapse, and you’re looking at the regulator, its enable line, or a shorted domain.
• The TPS546 self-protects against over-current, over-voltage, under-voltage and over-temperature, then latches off. A board that visibly cycles on and off is often the regulator faulting and retrying.

Cause 4 — blown input protection or a dead short

Bitaxe boards carry input protection (reverse-polarity and transient-suppression parts). A wrong PSU, a reversed barrel-jack polarity, or a solder bridge can blow it or create a direct short. Tell-tales:

• 0V everywhere downstream of the connector, or the PSU/USB port immediately current-limits and nothing boots.
• Near-zero resistance from the 5V rail to ground with power off — a dead short. Do not keep applying power; hunt the short instead. Common sources are a tombstoned or cracked input capacitor, a failed buck converter, or a physically cracked ASIC corner.

Step-by-step diagnostic with a multimeter

1. Visual first. Look for scorch marks, bulged or tilted caps, a cracked ASIC corner, or a lifted ESP32 shield.
2. Power off, measure resistance from the 5V rail to ground. A few ohms or less is a dead short — stop here; it’s a board-level fault.
3. Power on USB only. Does a serial device enumerate? Does any LED or the OLED light? Does Bitaxe_XXXX appear in your Wi-Fi list?
4. Measure the 5V input — expect ~5V, and it must be 4.5–5.5V.
5. Measure the 3.3V ESP32 rail. Missing 3.3V with good 5V points to the ESP32 regulator or an upstream short.
6. Measure the ASIC core rail — expect ~1.0–1.3V after the buck enables. 0V here while every other rail is alive isolates the buck, its enable signal, or a shorted domain.
7. Logic rails good but no AxeOS? Re-flash via the web flasher with the correct series image.

DIY vs. send it to the bench

Fix it yourself when the cure is a cable, port, or PSU swap, a USB-A adapter for the handshake quirk, a clean AxeOS reflash, the recovery page, or simply matching the right firmware image. No soldering required — these account for a large share of “dead” Bitaxes.

Send it to a bench for a dead short, blown input protection, a TPS546 that won’t start, a cracked or de-soldered BM1370/BM1368, or a damaged ESP32 module. These need a thermal camera to find the short, a microscope, hot air, and component-level rework — not something to attempt on a kitchen table.

D-Central repairs Bitaxe and open-source boards in-house. We’ve been doing component-level ASIC repair in Laval since 2016, and the same bench that rebuilds Antminer hashboards diagnoses Bitaxe power and ESP32 faults. If your board is past the reflash stage, send it to our ASIC repair service and we’ll trace the rail that died.

Keep going

If your Bitaxe boots fine but won’t hash, work through the zero-hashrate guide or our ASIC fault finder, and check tuning targets in the ASIC power profiles database. New to the platform or looking for a replacement unit? Start at the Bitaxe Hub or pick up a Bitaxe from D-Central.