Hi everyone,
I’m hitting a wall with a couple of ESP32-C3 Super Mini boards and hoping someone here might have encountered this specific hardware quirk.
The Issue:
The ESP32-C3 Super Minis refuse to connect to Wi-Fi, throwing either Auth Expired or 4-Way Handshake Timeout (with output_power: 19.5dB typically Auth Expired, power 5dB 4-Way-Handshake Timeout). Crucially, an ESP32-WROOM-32D connects to the exact same network instantly with the exact same ESPHome configuration, proving the network and credentials are fine.
The Logs:
The C3 can see the networks perfectly fine (RX works):
[15:29:11][I][wifi:1456]: - 'removed ▂▄▆█ Ch: 2 -71dB P:0
[15:29:11][I][wifi:1456]: - 'removed ▂▄▆█ Ch: 2 -85dB P:0
[15:29:11][I][wifi:1456]: - 'removed ▂▄▆█ Ch: 2 -64dB P:-1
But immediately fails the handshake:
[15:29:12][W][wifi_esp32:852]: Disconnected ssid='removed' bssid=removed reason='Auth Expired'
The “Smoking Gun” - Power Consumption Data:
I hooked up a USB power meter to log the current draw.
- Boot: Normal spike to ~140mA.
- Wi-Fi Connecting: The current flatlines between 60mA and 90mA.
A healthy ESP32 transmitting a Wi-Fi handshake should show clear current spikes up to 200-400mA. The lack of TX spikes, combined with the fact that it can receive beacon frames (RSSI is good), strongly suggests the TX path is physically dead or disconnected. The router simply never hears the ESP32.
Hardware & Troubleshooting Steps Taken:
- Boards tested: Multiple ESP32-C3 Super Minis (cheap AliExpress clones, likely ESP32-C3FH4). All fail identically.
- Control test: ESP32-WROOM-32D works perfectly on the same network.
- Frameworks: Tried both
esp-idfandarduino. - Board definitions: Tried
esp32-c3-superminiandesp32-c3-devkitm-1. - Network settings: Tested on a dedicated 2.4GHz WPA2 IoT network and a standard phone hotspot. PMF/802.11r disabled.
- Antenna Mod: Soldered a 31mm quarter-wave wire directly to the antenna feed point.
- Short checks: Verified with a multimeter that the 31mm wire / antenna feed point is NOT shorted to GND or the USB shield from either side. (The ceramic antenna itself shows <1 ohm as expected).
My Hypothesis:
Since RX works but TX draws no extra current and fails to reach the router, the RF transmit path is broken.
- Theory A: Is there a hidden, undocumented RF switch GPIO (like GPIO8 or GPIO9) on these specific Super Mini clones that needs to be pulled HIGH to enable the TX path?
- Theory B: Is this a known batch defect where the internal TX trace from the ESP32-C3 QFN package to the antenna pad is broken/missing on these specific boards?
Has anyone successfully debugged the TX path on these specific C3 Super Mini clones, or is there a specific GPIO initialization required in ESPHome to wake up the RF front-end?
Any insights would be highly appreciated!