nrf52840 power consumption with increased voltage for peripherals

[Felixx]

We are using nrf52840 USB dongle to get PIR sensor readings and send measurements over the Bluetooth Mesh
Schematics are following:

We increased power for peripherals to 3.3v using following code:
const uint32_t REGOUT0     __attribute__ ((section(".uicr_regout0")))
                           __attribute__ ((used))
                           = UICR_REGOUT0_VOUT_3V3 | ~UICR_REGOUT0_VOUT_Msk;
                           
Issue is that this change increased power idle consumption from 50μa to 550μa.

Is there any way to increase external peripherals supply voltage without increase of idle consumption?

[Andreas]

Hi,

 

Is this with the PIR sensor connected or not, if so what is it if you disconnect the PIR sensor?

 

Best regards,

Andreas

[Felixx]

Without PIR 514 μa

[haakonsh]

Felixx said:

Sensor goes idle after 10 minutes.

What do you mean, what sensor? I thought you said you did not use the PIR for these measurements

[Felixx]

Sorry, I mean nrf52840 USB dongle. In our internal communication we call "sensor" the whole assembly so wording got mixed up.  For power consumption measurements we used the dongle disconnected from PIR sensor and batteries.

[haakonsh]

Alright, thank you for the clarification, I'll try to make some measurements tomorrow.

[Felixx]

Did you have a chance to make measurements?

[haakonsh]

Sorry for the delay.

I've measured a dongle with your FW:

The current consumption in sleep is ~37µA, which is what we expect on the dongle. 

[Felixx]

Yes, it's expected consumption. Interesting why sometimes it goes up to ~500µA and if there is a way how to eliminate this behavior? If ESD is the case, is there potential Software fix or should we play with additional elements like capacitors?

[haakonsh]

Felixx said:

Yes, it's expected consumption. Interesting why sometimes it goes up to ~500µA and if there is a way how to eliminate this behavior?

 Are any other devices attached to these dongles? 
What instrument do you use to measure the current?

Felixx said:

If ESD is the case, is there potential Software fix or should we play with additional elements like capacitors?

 No fix in SW, the only HW fixes are ESD diodes. The nRF devices have two ESD diodes on GPIO, but they are there to protect the rest of the circuit, they will burn if sufficient voltage/current is applied. Capacitors won't help as they are not cost-effective. The only way to protect your devices is a good layout and regular ESD measures in your workspace.