LPS28DFWTR pressure sensor design bug?

[SSher.3]

I'd like to raise attention of ST staff regarding the LPS28DFWTR pressure sensor and what I believe is a sensor design issue (I hope I'm wrong, because it is an amazing sensor nevetheless!).

The original topic was here:

https://community.st.com/s/question/0D53W00001r8PF1SAM/what-is-the-material-the-membrane-is-made-from-on-the-lps28dfw

Just in case I copy-paste results of my quick research here:

I confirm that LPS28DFWTR is not just sensitive to sunlight, it is sensitive to ANY light! It feels like some p-n junction on sensor's die is exposed to light through transparent silicone and it's confusing readings/I2C. 

I'm attaching video here:

https://veed.io/view/a25b61dc-20d8-4442-998d-9bb3f83cf191

It's a simple setup where I measure both pressure/temp using LPS28DFWTR and convert it to altitude (you can see current altitude and temp resdings on display).

As you can see on the display, when I cover membrane and sensor to light - pressure readings are fine.

When membrane is facing bright light, readings of pressure are jumping quiet crazy, white temperature still being OK.

I used bright Cree cold led in the video, then I also tested it with IR led (850nm) and I think it got even worse.

Looks to me like a design issue, and it has to be black silicone instead of transparent one inside the sensor?

[Eleon BORLINI]

Hi @SSher.3​ ,

thank you for the notification, our suggestion is to cover the device with a gasket or mount it on the bottom layer of the board.

-Eleon

[NBake.3]

Hi @Eleon BORLINI​ - I have 2 follow up questions to this post:

1. I have noticed the same behaviour. Can you confirm if the sensor experiences a photo-electric effect that amplifies the signal current within the sensor?

Images below are of a device with two LPS33HW sensors mounted in orthogonal directions being dipped in and out of water. In the first case this is done inside under ambient light, in the second it is outside in direct sunlight.

 

2. In some cases we have observed the sensor measurement circuit appears to lock up but the digital line continues to output the last known value. Can you confirm the following: I suspect the Analog circuit has experienced some overcurrent and has shut itself down, but since the digital line remains powered it keeps reporting the last known value.

Can you confirm this behaviour?

 

[Eleon BORLINI]

Hi @NBake.3 (Community Member)​ ,

please check this thread and the related suggestions.

-Eleon