Setup:

Grove port wiring per docs: White → G15 (analog), Yellow → G13 (digital), Red → 5V, Black → GND.
Analog uses ADC2 (G15). Digital uses LM393 output (G13), input configured with pull‑up.
What we’ve tried:

Analog sampling: 200 samples @ 10 kHz using adc2_get_raw (fallback to analogRead), RMS and dB conversion. Readings are stable, but calibration points (33/55/80 dB) produce very similar RMS values (~0.058–0.068 V), so the regression is effectively flat.
Digital mode: often reports “pin stuck HIGH” unless the comparator pot is tuned. With pull‑up enabled, it still sticks high depending on threshold.
Disabled extra logging and added serial commands to do calibration and disable WiFi to reduce ADC2 contention.
Verified the issue is not M5Unified; Unit MIC gain is only the on‑board trimmer.

Questions:
Has anyone gotten meaningful SPL spread from the Unit MIC on G15 with WiFi running?
Is the LM393 output meant to be used on the K131 Grove port (G13), and are there known pull‑ups needed?
Any guidance on expected RMS ranges for 30–80 dB with this unit?
Happy to share code snippets if helpful.