I have a peculiar problem with a kinetis I hope the community can help me with. The problem is similar to a design based upon a coldfire controller (https://community.freescale.com/thread/365050) but have subsequently decided to switch to the kinetis. For the coldfire I used a P&E Multilink Interface and for the kinetis a Mulitlink Universal (brand new).
The pcb layout for the kinetis is for a specific design problem but for now I've only populated the pcb with the essentials to test the MCU. The MCU is a MKL02Z32VFG4 (a 16 pin QFN package). The MCU has power and ground connected to it with a decoupling cap in between. The same power supply is used to power the Multilink Universal. It is a 3.3V stable bench supply. The Multilink is connected to the SWD_CLK, Reset and SWD_DIO of the MCU. A 8.6kohm resistor is used as pull-up for the reset (also had a capacitor from reset to ground but it did not make a difference).
1) I wrote a simple program to toggle one of the pins (PTB1) (in CodeWarrior 10.4).
2) I programmed the MCU. Initially it did not want to program but reducing the debug shift frequency to 0.04MHz did the trick.
3) I then remove the debug interface, power cycle and let it run.
1) The MCU runs happily for a while but then starts to fail (can be from seconds to hours).
2) I can get an indication on when it starts to fail by monitoring the reset pin. Initially the reset pin is stable but then a square wave starts to appear (see figure 1). This however progress to become more unstable until as shown in figure 2.
3) Can not reprogram the MCU when when reconnected to programmer (in fact some of the MCUs start to smoke and are damaged permanently!!).
The above was tried on different PCBs with different designs and different power supplies, with more or less the same results.
When the MCU is flashed, is a crc done? Could the strange behaviour be as a result of the MCU being flashed incorrectly?
Help will greatly be appreciated!