fix: Bosch Twinguard 8750001213: Rename expose co2 to eco2

[burmistrzak]

This is a follow-up to a discussion started at #11204 (comment).

To avoid further confusion about capabilities, renaming co2 to eco2 seems to be an appropriate step. This should prevent ecosystems from misrepresenting sensor readings.

[Koenkk]

Thanks!

[mf19862]

I'm totally confused now. Where is the Co² data now? It is just gone. And the VOC data is now "ppm" it should be "ug/m³". This change totally messed up my setup and data

[burmistrzak]

@mf19862 Everything as intended. Let me explain really quick:

1. Devices like the Twinguard are not CO2 sensors. The BME680 inside doesn't detect CO2. It measures TVOC and derives a CO2-equivalent (eCO2) value for presentation using a proprietary algorithm.
2. According to Bosch Sensortec, the BME680 expresses relative change in TVOC in ppm. It therefore cannot measure the absolute concentration of TVOC in ug/m³.
3. eCO2 is a TVOC value, thus the change in device_class

I know it takes some time to getting used to, but at least it's correct now. 😅

[mf19862]

@burmistrzak Thank you for your detailed reply! This does clarify most of it. But there was an additional sensor "CO2" before, which is now removed. Where did this sensor come from?
The entity "CO2" did output a different value than "AQI" and "TVOC" but it was changing relatively.

[burmistrzak]

@mf19862 Well... We made it up (kinda).

Long story short: Before, we used a non-linear formula to calculate a sensible CO2 value from raw IAQ, but switched back to the default formula (likely also used by the Bosch SHC) because the sensor's raw IAQ output was not behaving as described in the BME680 data sheet.

Bosch is probably doing some sort of pre-processing in firmware before sending out the IAQ value via Zigbee.

Hope that makes sense. 🤞