we are basically using the TIDA 00489 hardware setup. Testing the motion detection behavior of our modules after fabrication in QM we noticed some strange behavior for a minority of our sensos (2-3 out of 100). In the following I am only describing the behavior of this "malfunctioning" sensors.
We put the sensors into a box where no motion (or IR radiation variation) can stimulate the sensors. However, the sensors in average detect about 20 motions, i.e. about 20 interrupts are triggered within 5mins. When we open the box and stimulate the sensors intensively (150 interrupts in 5mins) and close the box afterwards, those "false" detections do not occur anymore and no interrupts are triggered anymore when the senors are in the closed box. The following image illustrates this case in more detail.
Note that each point represents the accumulated number of interrupts within 5mins.
We did not change anything within the hardware parts from sensor, over comparator to the chip.
Any idea what could cause this behavior? In this thread I described my definitions for the HWI pins.