I2C "back powering"

Hi nvidia,

I’m wondering…

In my carrierboard I have the ability to completely turn off all supply voltages to the TX1 module.

When the TX1 is turned off completely, I have a few I2C devices (mainly a battery gauge and some related circuits) together with a small co-processor which remains powered at a seperate “auxillary” 3.3V supply which if fed from a smart battery.

The devices and co-processor talks to the TX1 when it’s on using I2C where the TX1 is single I2C master.
when the TX1 is off, the co-processor switches and becomes the single master and talks to the gauge etc…

The I2C port used between TX1 and our circuitry which is always powered is I2C_GP1_xxx
This I2C port has an active pull up IC (LTC1694) to help speed up the bus.

There are no other I/O’s from the TX1 that sees power/voltage when the TX1 is turned off other than the I2C_GP1_xxx pins which are being pulled to 3.3V via the active pull-up LTC1694.

— A thing to note is that I’m not seeing any issues with this setup ! —

My worry is that I may be damaging the port pads of the TX1 as i will be “feeding” voltage into the pads while there’s NO voltage on the TX1 hence if there’s some ESD diodes from I2C_GP1_xxx to some sort of VDD inside the module these will start to bias and current should/will flow through them sourced by my LTC1694?

My question is if this is something i need to worry about or if your I2C pads can survive having voltage on them when there’s no “VDD”?

If that’s not the case I guess i have to use some sort of hot-swap I2C bridge chip to seperate the power domains?


It’s OK if TX1 power off while external pull-up of I2C is active.