Troubleshooting

Symptoms you may see while bringing up an MCP9700, their causes and the fix for each.

Start with a multimeter on VOUT: about 0.75 V at +25 °C, and compare it with ReadVoltage(). If the meter agrees with the temperature, the problem is in the ADC or the adapter; if it does not, it is in the wiring or the sensor.

Build

Symptom	Cause and fix
mcp9700_version.h not found	The header was not generated. Link hf::mcp9700, or include the settings file (Installation)
No matching call to EnsureInitialized or ReadChannelV	The adapter's methods are missing or differ. Match the contract (Porting)

Initialisation and reads

Symptom	Cause and fix
Initialize() returns false	Null adapter, v_per_c not a finite positive number, v_zero_c not finite, or EnsureInitialized() failed. Check the arguments, then the ADC set-up
Every read returns false	Check LastError(). kNotInitialized: Initialize() not called or failed. kAdcError: ReadChannelV() returns non-zero; return 0 on success only and check the channel. kOutOfRange: see below
Reads fail with kOutOfRange	ReadVoltage() shows the raw VOUT. Outside 20 mV to 1.83 V (default range): open wire, grounded input, sensor unpowered or wrong pin (0 V); input shorted to VDD; adapter returns millivolts; or the part runs outside −40 °C to +125 °C (Configuration)

Wrong temperature

Symptom	Cause and fix
Wrong by a constant factor	ADC full scale, attenuation or a divider not undone. Return volts at the pin (Hardware)
A few degrees off, steady	Sensor grade and ADC offset. Calibrate (Configuration)
Warmer than the room	The sensor follows the board, heated by nearby parts. Move it, or use TO-92 for air (Hardware)
Noisy	Supply noise or digital lines nearby. Decoupling, 200 Ω in VDD, averaging (Hardware)
Wrong right after power-up from a GPIO	Read before the turn-on time. Wait at least 800 µs, with margin (Hardware)

Still stuck

Open an issue with the MCU, the supply voltage, the ADC settings, a multimeter reading of VOUT and what the driver returned.