Pixhawk 2.4.8 Barometer/Height drift

ArduCopter Copter 3.5
1

Hi,
as it states in the Headline I’ve got some problems with the Baro of my Pixhawk.
While some normal flights I noticed that the Copter is unable to hold a specific altiutde. It’s has some fluctuation up and down.
So I’ve looked in some logs and Yes I think this Behaviour of the Baro isn’t Normal:

BarometerAlt.jpg1920×494 86.7 KB

BarometerDesiredAlt.jpg1920×495 70.5 KB

For Comparison some Flight Log from the same Copter but with an APM 2.5.2:

APMAlt.jpg1920×387 55.4 KB

APMDesiredAlt.jpg1920×388 56.6 KB

So what could be the cause? I’ve read somewhere that Vibrations or/and the Wind flow inside the Case is an Issue, but look what happens as I Spin the Motors without Propellers lifting it 1-2m in the air. It’s better but still bad:

BaroInside.jpg1920×463 68.8 KB

I’m running Arducopter 3.5.2 on a 2.4.8 Pixhawk. At the time were the logs origined there was literally no Wind.

2

I had the same problem. found that the piece of foam shrunk covering my Barometer .
opened my pixhawk and replaced it.

3

Beside good quality foam and wind insulation, a clean dc power supply to the FC is an advantage.
Cost is low and advantage numerous.
LDO dc voltage regulator for small required power such as FC, Rx, GPS/Compass do marvel :slight_smile:

Henri

4

Added an LDO regulator for the same reason. I have an 10 millivolt ripple which is excellent.

5

10 millivolt ripple is not too bad.
It is certainly better than the 3dr switcher module which is in the hundred of millivolts at 1A :disappointed:

I do have an output voltage noise of 4.17 μV RMS (10 Hz, 100 kHz) using TI LDO.
In short 2398 times less than yours :slight_smile:

Ripple rejection: 82 dB (100 Hz) ≥ 55 dB (10 Hz, 10 MHz)

The LDO, I am using, has not a lot of output current, Just 1A @ 5Volts, but for the PixXXX receiver and GPS/Compass it is enough.

Cheers.
Henri.

6

Share an data flash log if you dont mind. What is the part number for your ldo?

7

Here is the LDOs I am using:

TPS7A.JPG704×539 109 KB

Of course you can find the full PDF from TI site. :slight_smile:

By the way removing noise from the source is good, having virtually none by using a separate small 2S battery for the electronics and opto isolated ESC is even better. (No common ground between sensitive electronics and very noisy switching ESC).

Cheers.

8

Thank you for sharing your experience. Wil definitely try this part.

9

Stil curious to see what your data flash log looks like on your ldo setup.

10

I used MC33269 , but the one you are using blows mine out of the water.