Challenges with LPS22DF: VDD_IO and VDD Interaction Causing I²C Issues

[Adi_Aljarbou]

Hi,

We are working with the LPS22DF pressure sensor to be used in our device adapting a 3.3V system and have encountered issues related to the interaction between VDD_IO, VDD, and the I²C bus.

• When VDD_IO and VDD are connected to the same 3.3V supply:
  • The I²C bus behaves as expected. (No issues here)
    
    
    
• However, when we disconnect VDD_IO and VDD, while I2C is active to be used for other sensors to communicate, the sensor seems to backfeed through VDD_IO's internal pull-up to VDD, interfering with the I²C bus (e.g., pulling lines to ~1.5–1.6V). (this is our theory at least)
  
  

• When VDD_IO is powered but VDD is not:

  • I²C communication is disrupted.
  • The sensor seems to partially power itself through internal pull-ups, potentially backfeeding into VDD.
    
    
• In addition, CS pin will sometimes influence the I2C bus if its connected to HIGH.
  
  

We have tested it using the evaluation kit and also on a breadboard. What we did was add a diode between VDD and VDD_IO to temporarily eliminate the issue. However, the I²C lines are still not stable and show a "wavy" signal.

Do you have any suggestions or recommendations? This is the first time we have encountered this kind of behavior with an ST sensor. We do not want to use power-down mode, as we are required to maintain extremely low current consumption. Using a different power source is also not an option.

[Adi_Aljarbou]

Appreciate if there is any guidance or recommendations

[Adi_Aljarbou]

I'm doubting that this is not the right channel to raise this topic?

[Federica Bossi]

Hi @Adi_Aljarbou ,

For safety during mounting and managing, all pins have two esd diodes (vs GND and vs its power supply). The two power supply lines have different strategy in terms of ESD but in any case they are intrinsically connected. This can avoid that the device is not correctly powered. The interaction between VDD and VDDIO is evident for this reason. 

Our suggestion is to avoid to have only one power supply that can create instability as described before. Preferred solution is to put the device in power down (ODR @0000).