BlueNRG chip becomes unresponsive

[charles23]

Posted on March 04, 2015 at 14:29

I have had two boards/chips fail in the same way - one today and one a few weeks ago. Symptoms on both boards are identical:

1) Boards had been working correctly and failed in operation. No obvious damage caused to boards (e.g. by static discharge). 2) The BlueNRG chip no longer generates an IRQ in response to a reset. 3) The scope shows that the DC-DC PSU is running continuously, and the the 1.2V and 1.8V supplies are present. 4) No oscillation on either crystal. 5) If I read the first 5 bytes of the SPI buffer the data is always 02 00 00 00 It is as though the BlueNRG chips have had some internal firmware corruption. Has anyone else seen this? What could cause it? #bluenrg-dead #bluenrg-unresponsive

[charles23]

Posted on March 06, 2015 at 18:16

I have had two more chips fail in exactly the same way - one yesterday and one today. In both cases the last successful exchange between the BlueNRG and my host processor was a EVT_DISCONN_COMPLETE event. In both cases the next attempt to send a command to the BlueNRG chip times out because the SPI buffer is never ready.

I have asked myself what is different in my code in the last few days, and the answer is that I am now using security for the first time. That probably means that the BlueNRG is frequently updating its flash memory with new crypto keying information. Would this happen just after a BLE disconnect?  My hypothesis is that there is a bug in the BlueNRG firmware that is trashing the flash memory during this operation. 1    Are you able to route this problem to someone who knows how to solve it? 2    Can you find out if there is a working application that is using Just Works security? Can I get the source code to run it on the X-NUCLEO-IDB04A1 and STM32 Nucleo?

[paulpaul9152]

Posted on April 24, 2015 at 10:31

I have exactly the same behaviour on a couple of boards, and this is a show stopper if I cannot find a fix. Have had any luck in finding a workaround/fix?

[charles23]

Posted on April 28, 2015 at 11:20

Hi Martin

I updated the BlueNRG firmware on all my boards from v6.3 to v6.4 and I have not seen the problem since. I suspect there is a bug in v6.3 that is trashing its flash. It would be good if ST could confirm this. Updating the firmware is a development project in its own right: I had to port the functionality of the Virtual COM Port project to my own boards, then connect them to the BLUENRG GUI program and use that program's firmware update capability. Not quick. Looks like DK 1.7.0 might offer an alternative to this, but I have not investigated it. Regards - Charles