Barometer - we use x/2 kPa. I think this is optimistic compared with the ECU map sensor using the calibration out of the service manual, AVC-R MAP sensor and with internet weather information for the prevailing barometric pressure. I think 0.495x kPa is nearer - it brings the atmospheric log readings down by 1kPa or 1% approx. Perhaps the supply rail is a little over 5V.

On the same lines - narrowband O2 - we use 0.0195x. 0.02x certainly makes it easier to read with 0.86, 0.88, 0.90, 0.92V being a typical range of full load values. If the ADC supply rail is over 5V then it will be likely nearer 0.02x anyway.

 

What's the inherent accuracy of the sensor? The high accuracy stuff from Rosemount that they use in nuke plants is only 0.2%, but you're looking at $2500+ per sensor. I'll bet the reference accuracy of any of the pressure sensors we use is greater than 1% (possibly much higher). Any adjustment smaller than that is effectively meaningless.

 

Agree, but if we are talking about a consistent offset between two sensors on the car measuring the same thing (MAP and Baro with engine off - whether throttle open or closed) then I think it is worth correcting as this error will probably double the error to 2%. When I last looked a typical automotive MAP sensor was <1% error, and typical 10 bit uC ADC error is <1LSB.

Maybe I'm becoming obsessed by speed and accuracy for no good reason

 

Agreed that you can cal out a consistent offset. You'll have to trend it long term to make sure. Hysteresis in the sensors could have you chasing your tail.