NRF52810 draws ~4mA when advertising

[kristjanb]

Hey. 

We designed custom PCB board (basically an humidity sensor) with Nordic IC. It works fine, but when I did current measurements I noticed that in advertising mode, uC draws ~5mA with spikes up to 16mA. LOGs are disabled, there is no UART periphery active. NRF_SDH_CLOCK_LF_SRC is set to NRF_CLOCK_LF_SRC_XTAL.

If I build image with BLE functionality disabled or set all GPIOs to input, the baseline current is still at ~5mA. 

Without debugger I observe similar situation, but current draw is  ~4mA. 

To make sure it's not other ICs on the board, I took a board without Nordic and current draw was ~10uA. 

Schematics:

Am I missing something? I did tests where all GPIOs were set as inputs, tried running it without SD enabled, but still the same results. 

Thank you!

[Vidar Berg]

Hello,

Could you please try program the "Hello World" sample with CONFIG_SERIAL set to "=n" to see if you are able to measure a lower base current then?

Best regards,

Vidar

[kristjanb]

Hey Vidar, I forgot to mention that I'm doing this on NRF5 SDK 17.1.0.
I flashed blinky (/examples/peripheral/blinky/pca10040e/blank/armgcc) with NRF_LOG_ENABLED 0 using NRF PROGRAMMER.


And here is measurement with erased flash


Thanks!

[Vidar Berg]

Hi,

In that case, at the very beginning of main() before configuring any pins, etc, please add the code below to put the chip into System OFF mode (deep sleep). Also, after flashing this FW, perform a power cycle before starting the measurement to ensure the chip is not in debug interface mode. You should measure sub 1 ua going into the nRF in this mode.

NRF_POWER->SYSTEMOFF = 1;
while(1);

Best regards,

Vidar

[kristjanb]

I modified blinky example with the two lines you sent and removed everything else.

1425.nrf52810_xxaa.hex

[Vidar Berg]

Apart from the nRF52810, do you have anything else on the 3v3 rail that could draw this current, or are you only measuring current going into the nRF5?

EDIT: looks like you have a SPI flash. Have you checked the DS for this part to see what it draws when powered on?

[kristjanb]

That's the thing. We have other things on 3v3 rail, but when I did measurement without nRF IC, average current draw was in normal ranges.

kristjanb said:

I took a board without Nordic and current draw was ~10uA.

Is there any obvious mistake on the nRF IC periphery? Perhaps something with clocks,..

[Vidar Berg]

A normal functional chip can't draw this much current in sleep. The only way I can think of to get current draw in this range is if you had configured some GPIOs to drive an external load before entering system OFF (GPIO states are retained in system off), which should not be the case here since your test app is entering sleep before doing any app initialisation. I don't have a good explanation for why it is only drawing 10 ua without the nrf though. Could it be that the flash(?) IC was in sleep mode earlier?

[kristjanb]

To be sure I just removed flash IC and I am still observing the same issue. I'll remove the nRF IC from this specific board to see what happens. I'll post back in few hours/tomorrow. Thank you for your help. Making sure the IC is actually in system OFF and GPIOs states are reset was very helpful.

[kristjanb]

Vidar, it turns out it's a 555 timer that is causing issues. I "removed" it by lifting Vcc pin, but it seems it's passive periphery was the culprit. Sorry for this, it has really nothing to do with the Nordic. 
Thank you again for your time!

[Vidar Berg]

Thank you for the update. I'm just glad to hear that you were able to find the issue.