Re: [Paparazzi-devel] Proper use of BARO_SENS

[Felix Ruess]

Hi Loic,

first of all, you are probably referring to the INS_BARO_SENS for rotorcrafts, right? So BARO_SENS in the INS section....

So this value currently depends on the scale of the raw measurements that you have in baro.absolute, so it is different depending on what baro is used (some write raw values to baro.absolute, some pressure in Pa)

> Do we want the baro_alt variation match to the altitude variation?

YES!

> So I wonder if i missed something: is it necessary to reduce BARO_SENS in order to reduce the barometer impact in altitude measurement,

> and give more weight to the accelerometers in sensors fusion?

No, definitely not... has nothing to do with the weight used in the sensor fusion. It's really just conversion of baro.absolute change to altitude change in meters.

Now for the case of Lisa/M, the BMP085 is queried for calibration data and you get the calibrated pressure in Pascal written to baro.absolute and not some random raw value scale. So this is what we want!

In that case you are right that you would end up with a INS_BARO_SENS of around 22-23, depending on your altitude:

* e.g. at about 500m above sea level you have ~11.5 Pa pressure change per meter

* converted to INS_BARO_SENS (for the fixed point position representation with INT32_POS_FRAQ = 8) that would be: 1/11.5*2^8 = 22.17

A new generic baro interface has been on the TODO list for a while and I already started drafting something up

* pressure in the baro interface is always in Pascal, let the baro specific driver worry about the correct conversion to Pascal

* have something like a baro_add_measurement function that automatically does the following for you, (so no need to do that in the actual baro specific driver again and again)

** if QFE (baro offset at ground level) is not initialized yet, average and set it

** calculate the altitude agl automatically from the absolute pressure in Pascal and apply low pass filter if enabled

This would mean that for sensors like the bmp085 you would not have to set BARO_SENS anymore, since you get the pressure in Pa and can convert that to altitude.

If you e.g. have an analog baro, you would have to specify a scale to convert the adc measurement to Pa.

Cheers, Felix

On Mon, Mar 12, 2012 at 11:53 AM, Loic Drumettaz <address@hidden> wrote:

Hello,

BARO_SENS is the barometer sensitivity and is used in the airframe configuration file to calculate the altitude from the pressure measurement. 

What is the proper way to tune it? Do we want the baro_alt variation match to the altitude variation? If I tune my BARO_SENS this way and I end up with BARO_SENS=23. I'm using a LISA M board V1.

I have checked various airframe configuration files and none of them contain a BARO_SENS higher than 16. So I wonder if i missed something: is it necessary to reduce BARO_SENS in order to reduce the barometer impact in altitude measurement, and give more weight to the accelerometers in sensors fusion?

Regards

Loïc

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel