If you had not logic analyzer on your hand, you could try to check the following step:
1.decrease SPI communication speed, 1MHz or lower.
2.check SPI bus drive capability: test 2 slave devices, 3 slave devices, ... until 6 slave devices.
Thanks for your tip. In fact, I succeeded communicating with 2 BMP388 lowering the frequency to 2,25 MHz.
However, when I Tried with 3 BMP388 the communication was ruined. I lowered the frequency till 70 kHz with no success.
I am using Adafruit BMP388 breakout board which has more components (voltage regulator, logical level shift, resistors, diods, ...) on it than the two 100 nF capacitors required by the BMP388 datasheet.
So I pursued my investigations measuring the capacitance between the GND and the other pins to verify if the bug is not due to the capacitive load presented by the Adafruit BMP388 breakout board to the microcontroller.
I am getting following values for the different pins :
I am suspecting the unrequired components (voltage regulator, ...) to present an excessive capacitance load to the microcontroller.
I have tried to find breakout boards that don't all those unrequired components with no success. Do you have one to recommend?
For Adafruit BMP388 breakout board you used, could we know what is your power supply for 6 Adafruit BMP388 breakout boards?
Suggest you supply 6 Adafruit BMP388 breakout boards with external regulated power supply, and check the current value change.
Do you mean you supply bmp388 and stm32l476rg microcontroller from samel voltage regulator? Such as LDO, etc. If LDO was used, I am not sure if the current is enough for both 6 Adafruit BMP388 breakout boards and STM32L476RG microcontroller.
What I want to say is to use the regulator that supports large current to supply power first, and then test it after ensuring that the power supply is sufficient.