I have a STM32F302C8T6 Design where about 20% of the devices (out of a of 40 devices) exhibit a rather strange behaviour:
1. On some Devices I could measure a Short circuit between VDD and GND which is internal to the device, the short cicuit was there before any power was supplied to the device. Replacing the microcontroler made the problem go away. The Device was X-rayed before the replacement, and no other shorts where found
2. On some Devices I could measure a Short circuit between VDD and GND after suppling power for about 3 msec. The short persisted for about 5minutes (measured with a multimeter) and than gradualy whent away. This process can be repeated. I noticed the following:
a) powering the micro with less than 1.9V resulted in a power consumtion of less than 1mA
b) powering the micro with more than 1.9V resulted in a power consumtion of more than 50mA (The short ciruit detection, would turn of the device) BUT: the hight current will be drawn only after about 1.5ms the 1.9V where applied
3. On other devices the J-Link Debuger could not reset the device
The power supply was checked, It ramps up very smooth, no overshoots, the maximum voltage measured (with a scope) was about 3.2V. VDD and VBAT are directly connected, VDDA is filtered with a beat (from VDD).
As far as I can tell, the other devices are OK. We could power them on/off without triggering any of the behaviour above.
For me I looks like a Latch-up effect. But why does It happen only after the internal reset block is above the POR Reset threshold? (Power on threshold is at about 1.92V, and POR reset temporization is in the range from 1.5ms to 4.5ms)
Has anyone experienced similar problems?