Use of 2 HEX HERE (v1) GPS+Compass modules with my Cube Black

Hi.

I’ve decided to mount a second HEX HERE (v1) GPS+Compass module in my Tarot T810 using a Cube Black updated to CHIBIOS 3.6.10.

Several questions came up when installing and setting up the 2 GPS+Compass units:

  • 1 ¡ SOLVED !

Can i use both of the 2 compasses included in this 2 GPs modules simultaneously with no I2C address conflicts or issues?

Here are my HW IDs with only 1 HERE (GPS 1) connected:

And here are HW IDs with both HERE modules connected:

Captura_002

As you can see, there are 2 external I2C devices now but identified as a different DEVTYPE each. If the GPS&Compass modules are jut the same hardware they should be identidied as the same DEVTYPE. Should’nt they ? Could this difference in detection be an issue ?

  • 2 ¡ SOLVED !

It is safe to enable the 2 externaly mounted compasses or is better use 1 external and the second from Cube Black internal SPI compasses ?

Captura_003

Why once both compasses are calibrated the OFFSET are so different if the orientation in the frame is the same for both modules ?

  • 3

GPS lock times are quite different from one of the units to the other. The unit i’ve just added connected to GPS2 seems not to be booting in HOT mode. This 2nd unit needs several minutes to get a similar number of satellites and hdop values than the faster unit that gets a GPS lock in a few seconds. It goes worse as time that modules are powered off gets higher. Could it be caused by a dead battery in the slower module ?

  • 4 ¡ SOLVED !

I’ve been told throught answers in previous topics that NO more than 2 satellites constellations (plus SBAS) must be enabled using bitmask in GPS_GNSS_MODE due to M8N timing related known issues. It is safe to setup different GPS_GNSS_MODE values for both units ? I.E. Unit 1 using GPS+GLONASS+SBAS (GPS_GNSS_MODE = 67) and unit 2 using GPS+GALILEO+SBAS (GPS_GNSS_MODE = 71). Should it be reliable in a BLEND configuration ?

Thanks in advance.

Miguel.

I think it picks up the internal compasses. Try loading only the driver of your external device trough dedicated bitmask and connect the 2 external compasses to 2 different i2c bus to avoid same id on same bus.

Hi Corrado.

The initial datasheet from Hex manufacturing (http://www.hex.aero/wp-content/uploads/2016/07/DRS_Pixhawk-2-17th-march-2016.pdf) shows something like this on page 4:

#############################################

Pixhawk FMU Main Board
• STM32F427; flash 2MiB, RAM 256KiB.
• On-board 16KiB SPI FRAM
• MPU9250 or ICM 20xxx integrated accelerometer / gyro.
• MS5611 Baro
All sensors connected via SPI.
• Micro SD interfaces via SDIO.

Vibration Damped IMU board
• LSM303D integrated accelerometer / magnetometer.
• L3GD20 gyro.
• MPU9250 or ICM 20xxx Gyro / Accel
• MS5611 Baro
All sensors connected via SPI.

#############################################

It seems that all integrated sensors use SPI interfaces.

And if we compare info shown in

to info shown in

Captura_002

It seems that external compasses are detected as I2C connected and (to bus 0 and 1). I’m assuming that GPS1 is connected to one I2C bus and GPS2 to the other one (please correct me if i’m wrong).

First time i went to compass calibration in Mission Planner after connecting the 2nd HERE the 2d compass was already configured as “externally mounted”.

My only doubt about compasses at this point is why they are detected as different devtype (DEVTYPE_AK09916 one and DEVTYPE_HMC5883 the other) and wich compass setup is more reliable, external+external or external+internal.

Thanks again.

Hi again.

Related to compasses different DEVTYPE a consistent answer was given at https://discuss.ardupilot.org/t/can-i-use-and-configure-2-externaly-mounted-compasses-2-included-in-hex-here-v1-modules/45700

What cuold be a problem is now an advantage. That 2 different mag chips circunstance avoids hardware conflicts at all.

Thanks a lot.

Just the answers for specific GPS questions (#3 $ #4) are still missing.

regarding question #4:
You can set different GNSS mode and blending will work, however I can’t say if that’s necessarily a good idea. I know the blending can produce bad results if one GPS is giving weird readings, but I don’t know how different the fix solutions might be or whether the blending will handle it gracefully.

Thanks. I know that GPS Blending needs GPSs with very healthly readings in order to work in a reliable way.

Guys, Thanks for your answers.

Only question #3 is left and i will replace the battery in the slow HERE unit to probe if my theory is valid.