Questions about standby current defects during production

[pkr2258]

hi

We have developed it to the specifications below and are preparing for mass production.

We would like to inquire about a 10% standby current test defect occurring in pre-production (100EA).

1. Specifications

- nRF52832

- SDK 17.1.0

2. Self-review results

- PCBs with standby current defects continue to defect (100%)

- Normal PCB with no standby current defects continues to be normal (100%)

- If you retouch the nRF52832 IC with a heat gun, the standby current will return to normal for a while and then increase again.

- Changing peripheral devices such as crystal, LDO, etc. does not solve the problem.

- Normal current: 100uA -> defective PBA 230uA

=> Is there anything you can guess as the cause?

[Sigurd]

Hi!

Did you try to do a swap test?

Take 1 good unit and 1 bad unit.

Then take the nRF52832 from the bad unit, and place it on the good unit PCB.

Then take the nRF52832 from the good unit, and place it on the bad unit PCB.

Then check if the behavior follows the nRF52832 or the PCB.

[pkr2258]

HI

I tried replacing the nRF52832 on the defective PCB with a new chip, but the result was the same.

And as you advised, I replaced the defective and good PCBs, but the same symptoms still appear.

Whether you change a defective PCB to a new chip or replace a chip in a good PCB, it is always defective.

However, if you look at the PCB alone, you cannot tell the difference between a good product and a defective product.

Even though ICT was used, there was no difference between defective and good products.

And the differences in the PCB cannot be found with the naked eye.

Under the circumstances, it is true that the defective symptoms follow the PCB, but I cannot understand the reason.

Is there any way to get to the root cause through debugging mode or other methods?

[hmolesworth]

Pin 13 VDD appears to be unconnected, which may affect operation though not in a predictable manner. Worth setting all 32 i/o pins to outputs and pull-ups disabled and drive all 32 pins low (unless some other circuit disallows that) and measure the supply current; it should be very low; is it different?