I2C splitter & NEO-M8N GPS + HMC5983 compass problem

I have problems with compass values when the HMC5983 compass is connected to I2C bus via I2C splitter.
Values are off range , Z over 560 .

If I connect directly the compass to I2C socket the compass values are much less and are ok for Mission Planner.

Since I need to connect at least the led How cand I solve this issue ?
Is there any setting in Mission Planner about the I2C bus that can help ?

Thanks !

Can you list all the devices connected to your I2C bus? It may be a pull-up resistor issue.

Only the Pixhawk led is connected.

I thought it was the I2C splitter but also without the led I’m unable to have a working compass.

I tried two Drotek NEO-M8N Gps, one is the XL the other the mushroom type .
The second was brand new but gives all sort of errors, i.e. the red heading line in mission planner continues to turn even if the drone is still on ground.
All sort of EKF errors related to compass .
I’m using 3.4.5 firmware version b ut also with 3.3.3 the XL gives off range values , Z over 560 .

I tried also with some cheap chinese HMC5983 compass but no way either .

I really do not know what to do , it is 3 days I’m facing this problem , on two brand new Pixhawk from Mrobotics.

I will try with a Dopix I have never use to try to understand where is the problem .

I just bought two Drotek external compass .

Found out what is the problem.

It is the I2C bus on Mrobotics Pixhawks !

I have install a Zubax Gps that has onboard compass CAN BUS and now works perfectly .

Good to know, are you sure it is the I2C bus though?

Pretty sure , all I2C compass were faulty with the CAN bus compass it is perfect.