nRF5340 IO pin injection current

[dyoung.mpd]

Hi,

I have an nRF5340 design that uses the ADC to monitor an external voltage.  In some operating scenarios the external voltage may be present prior to VDD at the nRF5340.  The monitored voltage is divided down using two 100K resistors before it is connected to the nRF5340 (refer to attached image).  The monitored voltage has a maximum value of 4.2V at the input to the resistor divider.

Is there a maximum current that can be safely injected into an IO pin of the nRF5340 and does it vary by IO pin type?

Are there other potential faults that could occur if current is injected to an IO pin prior to application of VDD such as a corrupted boot sequence or lockup of the SOC?

Thank you for your support.

Sincerely,

Doug Young

[hmolesworth]

This circuit will cause problems, in particular with a subsequent reset not functioning correctly due to phantom power (aka back-drive) holding up internal VDD with just a few 100s of nA. Have a look at this and similar posts.

causes-vdd-to-rise-in-deep-sleep-mode

[dyoung.mpd]

hmolesworth,

Thank you for the link, however in this case I am not concerned that leakage from +VGENIN_MON will cause the VDD net to rise beyond its rated voltage because it will not be > 2.25V due to resistor divider formed by R12/R13.  Also if +VGENIN is present VDD will be greater than 2.25V within a few milliseconds thus shutting down the leakage path.

You did mention a reset not functioning properly due to leakage.  I did not come across posts relating input leakage to reset functionality, do you have further posts or insight to share about this effect?

[hmolesworth]

Try breaking the supply from the battery momentarily such that VDD falls reducing current which in turn reduces the rate of voltage drop. At some point the voltage divider from battery supply becomes higher than VDD as the regulator has turned off due to low input voltage. Phantom power now stops the reset working as designed unless the slow discharge (currrent is now few nA)drops below some magic value, but if the battery is replaced or the switch closed before this happens the nRF5x can get stuck and never recover until the battery is removed. If you never see this issue then maybe no worries.

weird-issue-observed-with-nrf52832

[dyoung.mpd]

Thank you for the further explanation hmolesworth.  I think that the conditions of my application are sufficiently different that the few nA hold-up will not occur but I will dig into it further.