LIS3MDL going dead after few usage

[RMish.1]

Hello everyone,

I need help regarding LIS3MDL magnetic sensor aka compass issue is i am using this sensor

on my custom made gps pcb which is then used by my quadcopter the issue is that it works fine at the very start but after few usage when i connect my drone multiple times it just go dead my quadcopter start to say bad compass health and as soon as i replace the LIS3MDL

it again goes back to normal and this is repetitive i have 15 drones and it's happening to all of them

before this i had used HMC5883 compass never faced any problem but LIS3MDL

seems to having some issue i am also attaching the schematic of my compass

PLEASE ANY HELP OR ANY INSIGHT WILL BE HELPFUL I AM STUCK HERE FOR MONTHS NOW

[Eleon BORLINI]

Hi @RMish.1​ ,

but does the LIS3MDL suddenly get stuck or does it begin to output random numbers?

In the first case, does the device recover after a soft / hard restart?

In the second case you should compensate by hand, since differently from the LIS2MDL, this device doesn't have unfortunately the magnetometer offset cancellation procedure.

A more question: which is the failure rate for this issue?

-Eleon

[RMish.1]

replying to first question it always remains stuck neither it able to recover by any method we have to replace it in order to make it work

second case never happened

failure rate is like out of 15 it happens on all the drones but it's not like it will fail instantly it will fail after few times of power cycling again and again

[Eleon BORLINI]

Hi @RMish.1​ ,

I'm wondering if it could be a soldering issue, or something related to the SMT process. Or maybe, the board is receiving some shocks and the soldering joints are not robust enough...

When you rework your board with a new LIS3MDL, do you use the same process as the previous one? And does the "reworked" device never fail after?

I have also another question related to the hardware connections: are Vdd and VddIO tied together in your schematic (I mean, are they the same +3.3V net, so that both lines share the decoupling capacitors)?

-Eleon