nRF52832 hangs when low-power mode is used

[bluera1n]

Hi,

Recently, I faced very weird problem which stops working MCU when MCU enters low-power mode.

Not all PCBA I made have the same problem but quite many PCBA shows the same symptom (around 25%).

My setting is as follows.

- nRF52832-CIAA (I have both B0 and E1 revision)

- S132 v2.0.0 Softdevice and SDK 11

- DCDC enabled, 32MHz & 32.768kHz used

Initially, the symptom was different and I passed through following workaround.

1. MCU hangs if external 32.768kHz clock source is used -> I changed LF clock source to internal RC oscillator then MCU works.

2. MCU hangs when it enters System ON low power mode

Here, the primary problem is MCU is not working if 32.768kHz XTAL source is used.

I checked XTAL oscillation of good unit and bad unit but XTAL oscillates in both units which seems no problem in XTAL.

Good units have no problem at all in using 32.768kHz XTAL source and low-power mode.

By measuring current consumption of bad unit when MCU hangs, I can see below current consumption behavior.

It looks like multiple sine waves are overlapped by 90 degree shifted but I'm not sure it's meaningful clue for the problem.

Any comment would be appreciated. Thanks!

[ketiljo]

Looks like maybe there's an issue with the 32 kHz crystal. What is the CL value of the crystal (from the datasheet) and the value of the load caps you are using?

[bluera1n]

XTAL's CL is 9pF. I'm using exactly same XTAL used in reference design (CM8V-T1A).

And load cap is 12pF in parallel.

BTW, I found new symptom which shows weird behavior in buck output voltage when MCU hangs.

But I'm not sure there is correlation between 32.768kHz and buck converter.

[ketiljo]

For a 9 pF crystal, you should use 15 pF load caps. 

Ccap = CL * 2 - 4 pF

[bluera1n]

But reference design also uses 12pF.

Is 3pF difference huge for 32.768kHz XTAL?

[ketiljo]

The reference crystal has a lower CL value.

Please test this.