HDC1080 in dev with strange temperature value

[TheBarz]

Hi to all !
I 've got some trouble since i done a factory reset and upload las firmware on my NodeMCU8266 and HDC1080 sensor connected.
Software displayed wrong temperature display compare with BMP280 as -39.97°C vs 22,5°C. On log is the same.
In the opposite way relative humidity value seems to be correct like 50.7%

I done a change with a new sensor spare and it's a same.

Are they somes INO buggy thing with this sensor ?

Thx to all.

[TD-er]

Can you try running at a lower I2C clock speed?
Either tick the "Force Slow I2C speed" in the task settings, or lower the 400 kHz default to 200000.
For now you need to reboot to make changed the I2C clock settings active (fix is already merged, not yet included in a nightly build)

[ThomasB]

I'll add a some more things to do:
1. Reduce sensor's wire length. For testing, keep it very short (under 8cm). After it is confirmed working you can attempt to increase wire length.
2. Confirm the SDA and SCL pullup resistors values are appropriate. 4.7K ohms is typical if wire length is reasonable.
3. Run the i2cScanner tool multiple times. Confirm that address 0x40 is reliably reported each time.

BTW, a temperature of -39.97 means the raw sensor reading is all zeros. So I think there is a i2c communication issue. All the suggestions provided so far will help reduce common I2C problems.

And if you need more help then I advise you provide more info. Such as info to your sensor module (e.g. vendor link), a clear photo of the wiring, and screenshot of the device settings.

- Thomas

[Ath]

And as the definition of 'last firmware' is somewaht ambigous, please provide the name of the .bin file as that is the most informative.

[TD-er]

And as the definition of 'last firmware' is somewaht ambigous, please provide the name of the .bin file as that is the most informative.

At least call it "latest", or else one may think I'm done with it

[TheBarz]

Thanks for your answers

In fact the sensor needs time to adjust good measure. I suppose it's like some gas sensor need to heat some internal component.
It(s take around 5/10 minutes.

Sensor gives good metric value now.

Tx again.

[TD-er]

5 - 10 minutes doesn't sound normal.
Are you sure the sensor wasn't exposed to some external factors like excessive humidity?
I have one of those on my breadboard and it immediately gives reasonable values.

If I remember correctly the sensor does use some kind of calibration, which is also stored to survive a reboot (warm reboot, not a power cycle)
So it might be this data was still present from a previous run of another task?

Can you test by performing a full power cycle to see if it again does show some strange values?
Please let me know the result.