High IDLE current on nRF52832

[Otavio Borges]

Hello all,

During current measurements of our product using nrf52832QFAAB0 I've noticed a high IDLE current of ~450uA. Trying to isolate the problem we programmed our board with ble_app_beacon example from SDK 14.0.0 and Soft Device s132 5.0.0 obtaining the same IDLE currrent.

Including sd_power_dcdc_mode_set(NRF_POWER_DCDC_ENABLE) (on the end of ble_stack_init()) reduced the IDLE currrent to ~275uA. A value still quite above the one described on devzone.nordicsemi.com/.../.

We've run some tests on a board with minimal circuitry, as described on section 53.2 from the nRF52832 Product Specification v1.3, getting the same ~275uA.

Browsing the forum I stumbled on [156] GPIOTE: Some CLR tasks give unintentional behavior but I not sure if it's applicable, since ble_app_beacon does not use GPIOTE.

I've used an Oscilloscope, an instrumentation AmpOp and 1 ohm shunt resistor to read the lower currents. And have validated the minimum precision by measuring the current of a device on System OFF mode, on this mode the reading was ~2uA.

Is there any other configuration missing from the ble_app_beacon to reduce the power consumption or any other bug that could be causing this high IDLE current?

Thank you.

[a.o.]

I had a high idle current on one of my projects of around ~200uA and I believe the cause was that a pull-up had been active on a pin that was externally pulled low.

[MartinBL]

Hi. User @a.o. has a point below. The value of the pull resistors in the nRF52 is typically 13kOhms. Hence, a pull-up being pulled low, or a pull-down being pulled high will typically draw ~3V/~13kOhm = 230uA. Dependent on your supply voltage of course.

[Otavio Borges]

Checked the pulldown confĩguration, but none of the GPIOs were responsible for the high current. Turns out the FPU was being kept running. I've posted an answer detailing the issue. But still a good reminder.

[Otavio Borges]

Turns out, after SAADC successful conversion I used the FPU to calculate the voltage drop on the battery. As described in devzone.nordicsemi.com/.../ there's a bug on nrf52 that, even disabling the SAADC peripheral, the FPu is kept alive holding the CPU to go on sleep mode.

I've removed the floating-point operation and IDLE current dropped-down to around 5~10uA. Considering the others ICs on the board that value is within expected range.

Nevertheless, is there a workaround from nordic to address this issue? Or a way to manually turnoff the FPU?

Thank you all for the help.

[MartinBL]

Glad you figured it out. A workaround can be found in the documentation of PAN 87.