Question about the state of only maintaining 5~10mA current consumption without entering the Main loop.

[SHMoon]

HI

We are producing products with nRF5232.

This time, a defect has occurred and we are reviewing it, but the analysis is not working well, so we are asking for help.

1) Visually, there is no problem with the PCB.

2) The voltages on the power pins of nRF52832 are all normal.

3) J-Link is connected, but in debugging mode, it does not enter the Main statement.

ㄴ Of course, advertising or other peripheral devices do not work.

4) The current is fixed at around 5~10mA.

5) Although there was no problem with the naked eye, I tried replacing the 32.768kHz crystal, but it did not solve the problem.

6) This is the first problem that has appeared while producing more than 100,000 units, and it is thought to be a PCB production issue rather than a design error.

~

Is there anything to consider due to the above symptoms?

[Hung Bui]

Hi SHMoon, 
Could you clarify: 

- You are observing the issue on only one board. Other boards of the same batch work as expected ? 

- Can you program the defect board ? 

- Have you tried to resolder the chip on the defected board to another board ? Did the issue follow the chip ? 

Please try to test with some simple application that doesn't use the 32kHz crystal. You can try testing by blinking the LED without using RTC . Then try do BLE and use internal RC instead of 32kHz crystal. 

[SHMoon]

HI, Hung Bui

Hung Bui said:

Please try to test with some simple application that doesn't use the 32kHz crystal. You can try testing by blinking the LED without using RTC . Then try do BLE and use internal RC instead of 32kHz crystal. 

Thank you for your reply.

.

Hung Bui said:

- You are observing the issue on only one board. Other boards of the same batch work as expected ? 

This problem only occurred on one PCB.

We have already produced more than 50,000 PCBs.

.

Therefore, I think of it as a PCB work problem rather than a design problem.

.

Hung Bui said:

- Can you program the defect board ? 

.

As mentioned in this article, program download is possible through J-Link.

Of course, debugging mode is also possible.

.

Hung Bui said:

Please try to test with some simple application that doesn't use the 32kHz crystal. You can try testing by blinking the LED without using RTC . Then try do BLE and use internal RC instead of 32kHz crystal. 

.

However, it does not enter the main() function in debugging mode.

The remaining part was a chip replacement, so I tried using a hot air blower, but the work was impossible with the hot air blower I had.

.

Hung Bui said:

- Have you tried to resolder the chip on the defected board to another board ? Did the issue follow the chip ? 

I checked it using the "blinky" example and it operates normally. (internal clock)

It was confirmed that the mian() statement is executed normally in debugging mode.

(When externally clocked, the mian() statement is not entered)

thank you

[Jared]

Hi,

So if you use an application that does not use the LFXO crystal then it works?

Can you try to replace the IC from a bad board with an IC from a good board? That could tell us whether the issue is with the IC itself or the crystal,

regards

Jared