Could AP allow for non linear current sensors?

I first filed a report in the pixhawk family board because this happens to be on a pixfalcon and some testing shows that it’s unable to read a voltage below 0.1V on the current sensing wire of the power module connector.
Because my Amps per volt is around 21.4, basically it sees no current at all until I’m using about 2.5A :frowning:
That post is there: Pixfalcon doesn’t read any amps until about 3A and then power module is not linear beyond that

But then, a current to volt table shows me that even beyond 3A, there is an issue with power modules not giving linear values for current.
At 3A, amps per volt is about 25, and it slowly goes down to 21 as I reach 10A+, so even putting the pixfalcon sensor issue aside, measuring straight off the power board shows that AP needs some kind of non linear table to get proper amps from at least some sensors since their amps per volt is not linear.

I have 3 power modules, one from 3DR and 2 clones, they all look like this
https://www.aliexpress.com/store/product/CRIUS-APM-Power-Module-30V-90A-with-BEC-3A-Ouput-Support-30V-for-APM-2-6/405418_1962418918.html
I’ve also seen that with other power modules.

I’ve definitely see the non linear option with sensors plugged into a pixhawk too.

For now, I set 21 Amps per volt, and about 11.5V per volt with that sensor, and all those sensors send data via power module port in pixhawk or pixfalcon

The table below shows the pixfalcon seems to have a flaw where it cannot reliably read the voltage from the power module when it’s too small (0.037V for 0.45A, 0.12V for 3A), but beyond that the sensor is also not linear:

A was measured with an amp meter off the battery
V was measured on the power module between Gnd and I
V A A/V
0.037 0.48 13.0
0.040 0.96 24.0
0.059 1.50 25.4
0.068 1.75 25.7
0.086 2.15 25.0
0.105 2.60 24.8
0.120 3.00 25.0
0.200 4.50 22.5
0.370 8.00 21.6
0.530 11.0 20.8
0.670 14.0 20.9
0.800 16.6 20.8

If you’re talking about hall effect sensors, like the ones supplied by mauch, they are supported by mission planner. you can check the documentation here- http://ardupilot.org/copter/docs/common-mauch-power-modules.html

@Nihal you’re right that there are better sensors, but I do believe a bit of software improvement could allow getting much better values on the many 3DR power modules that were sold with pixhawks (let’s not worry too much about every single last clone since I’m sure some of those are beyond repair and use despite software kludges).
Either way I opened an RFE here: https://github.com/ArduPilot/ardupilot/issues/6213

Hello Marc, I experienced a similar issue…

Last Sunday flying with my sky-walker I noticed that voltage drop to 14 volts and flight current was between 1 A to 3 Amps too low for sky-walker (4Cells) so I decided to land. End of story, I checked lipo in land with a multimeter and mission planner, both measure the same value when the motor is not running but when motor is full load there is a difference between power module and multimeter of about 1Volt PM=14V Multimeter=15V. Too much!!!

I fligth several times this plane, this is the first time I noticed this issue, it is like Power Module is not well calibrated under load.

ONLY difference from the past is that I recently upgraded PIX firmware from 3.5.3 to last version 3.7.1. Old firmware was running OK.

Is there anyone experiencing this issue?