I am using STM32F4 processor. The problem I am seeing is that at times the processor resets and the CSR register reports 04 in reset cause which is NRST pin reset. Problem is that we do not see any reset on the NRST Pin.

[RBharol]

Scope is configured correctly. Reading of CSR register is done way early in the code and is reliable. Register is cleared after reading. Scope is configured correctly.

When we trigger pin reset externally, we can catch it on the scope. When we trigger Watchdog reset, we can see the pulse on the pin but not when device is taking a reset on it's own.

 What could it be? 

[TDK]

Do you have a 0.1uF cap on the NRST pin? Could be quick enough that you don't see it. Although that seems far fetched.

Is the reset correlated with any program behavior?

[RBharol]

Thanks TDK,

We don't have 0.1uF Cap on the pin. Attached image shows how our circuitry looks like. We see is more often when Modem on the board is active (and hence is drawing extra current).

[TDK]

I would put a 0.1uF to ground on the NRST pin to see if that fixes it. Could be something causing noise on that wire. Since you have no capacitance and just a weak pullup, it's going to be susceptible to noise. If that's not it, I'm not sure what else could be causing it.

[RBharol]

Putting 0.1uF appears to have fixed it. I will keep monitoring and update here for everybody's benefit.