nrf51822 clock problem on voltage drop

[Sergy]

Hi,

Im using an app_timer in my application to create a timeout for UART read (im not using SoftDevice). On my board I have USB and battery power, so at the moment when USB power disappears I need to send a message over UART about it. I think there is a voltage drop happening at that moment and app_timer handler doesnt get executed at all, so my program gets stuck in endless UART loop. The question is - if there is a voltage drop why MCU doesnt reset itself, but instead there is something happening with internal clock. And second question - how can I work around it?

Thanks!

[Asbjørn]

Do you have a plot or picture of how much the voltage drops and how low it gets? Is it something you can smooth out with a large capacitor in between the two power supply options?

[Sergy]

USB and battery go to buck-boost DC/DC which outputs a 4v and there I have a pretty big capacitor since I have GSM module connected to it - I dont see any voltage drop there at all. But then further it goes onto 3.3v LDO and to nrf51822. Didnt measure 3.3v line.

On USB+battery input i get a drop down to 1.2v.

[Asbjørn]

If the voltage of the chip changes too fast, you could create issues for the oscillators on board and you could potentially see different behavior from the oscillator. What kind of behavior change do you experience?

Regarding the voltage drop, do you see that it drops down to 1.2 V for a short time?