nRF52 not booting properly without reset

[Dajgoro Labinac]

Hi

I did a bunch of boards for a product, and in the latest version I used a BQ25121 PMIC which also has a reset output which I tied to the nRF52 input. On the previous boards the reset signal was tied directly to a button and it worked. But now when the nRF52 is powered, BLE does not work for some reason, and some other bits of code don't work either. But when I reset it manually it all works fine. The reset reason(RESETREAS) after powerup on the new board is 0, while on the old boards or after it was reset via switch its 1. When its 1 it all works fine, but not when its 0. I tried detecting the reset reason being 0 and then issue a reset via firmware, but that does not help. All the BQ25121 has is a open drain output, so it should behave much like a physical switch, but something is affecting the nRF52 and making it not boot properly.

[Nguyen Hoan Hoang]

Check if you got VDD stable, enough decoupling caps. What is your VDD level ? Any soldering issue ? Any residual flux on the pcb. Have the board cleaned to remove all flux that could stuck under the IC. Check also if SWDIO or SWDCLK got pulled down at any instance after power up?

[Dajgoro Labinac]

I added decoupling as recomanded, and followed the layout design guides, and placed the PMIC close to the MCU. I've tried soldering, resoldering most of the components, but the issue persists over multiple boards. VDD level is 3.3V, but when using the PMIC switching regulator its 2.5 which I can then adjust via I2C. But that does not seem to have any effect on the issue. After BLE is initialized, there is a 50mV spike occurring at a 45ms interval. And as I was typing this, I went to chase that 50mV spike down, and I found the culprit. I have a VCNL4040 proximity sensor, and when it is initialized, the IR led somehow creates enough interference to glitch the BLE radio. Funny thing is on the old version I had it as well. Now I have to test the other boards too which I did not butcher while experimenting.

[Nguyen Hoan Hoang]

Glad that you found the issue. One thing that comes in mind is the gnd plan. You may not have enough via or floating island some where. Second thing that comes in mind. Check also that the PMIC has enough power to drive everything at once. Make sure you don't draw more than 80% of max capacity.

[Dajgoro Labinac]

Even tho its only a 2L board, I have the most of the bottom layer as ground plane, actually I only have a few short traces which act as jumper, all of the traces are on the top layer. The load should be well below of what the PMIC can handle, so that should not be an issue. I'll try investigating how exactly is the interference manifesting, and I'll probably put some sort of filter at the sensor power supply.

Thanks for the troubleshooting help!

[Nguyen Hoan Hoang]

I think there isn't much I can do at point. Good luck with troubleshooting your PCB.