Thanks for the photos!
It seems like there was a "lifting" force applied to the connector, probably due to a cable pulling it from upwards. Perhaps the lowerleg KCeptor was placed too high, or the cable is just not long enough.
zaanen A real closer look it looks like the PCB board contact point is loosing it strength and the welt is still in place.
Yes, this is something that normally happens when mechanical stress is applied to a PCB component such as a connector. The cooper layer to which the connector is soldered to gets strapped, and you cannot re-solder it in place.
Luckily you can still use the other connector on the KC, let me know it is still working correctly at an electronical level.
zaanen When calibration and spacial walk is there i think i put the feet sensor not with a strap but tight up in the shoelaces.
Just make sure you are appling no mechanical strain to the PCB of the sensor. Even small amounts of torsion on the PCB may lead to incorrect measurements. That's why the socket is designed with that particular shape: to absorb torsion coming from the elastic band.
Take a look at this post for a more comprehensive explanation: https://forum.chordata.cc/d/104-effects-of-pcb-torsion-on-magnetometer-measurements
