nrf52DK 32k crystal current higher than RC

[lev]

Testing on nrf52DK (QFAAB0), without softdevice, unfortunately still on SDK10 and NRF52 sleeping on RTC. Programming pins disconnected.

Measured:

• low frequency clock source being external crystal with 3.05uA,
• running on RC oscillator it shows (without calibration) 2.0 uA.

Product specification v1.3 reads as follows:

• I_LFXO Run current for 32.768 kHz crystal oscillator: typ. 0.25uA

• I_LFRC Run current for 32.768 kHz RC oscillator: typ. 0.6 uA max. 1uA

I can't understand why does running on external crystal consume more than on RC (>1uA (3.05 - 2.0) more, whereas it sould be ~0.35 (I_LFXO - I_LFRC) less instead)?

Found one errata issue thats somewhat similar, though there is no soft reset occuring: [101] CLOCK: Sleep current increases after soft reset

[Turbo J]

Try nrf_gpio_cfg_default() for all those pins - this should yield less current consumtion.

[lev]

Thank you so much Now on Xtal consumes around 2,1uA and RC 2.0uA. Coming from Atmel world where unconfigured floating inputs always drew unexpected amounts (at least tinys)... So this must be possible because of true input-disconnect...