GPIO24 not longer functional

After startup, the GPIO suddenly stopped showing voltage (1) and now shows zero. Normally, 3.3 volts is present at the pin. It has now been discovered that 3.3V tolerance is disabled in the PINMUX. Could this cause the pin to fail, resulting in the loss of voltage? Is it mandatory to enable the 3.3V tolerance?

hello matthias.wieland,

may I double check which Jetpack release version you’re working with?
and.. please refer to 40-pin Header, which pin you’ve the issue reported.

I wanted to add following information. JETPACK Version is 6.2.1 and we don´t use the development kit but a custom board.

hello matthias.wieland,

please double check you’ve done pinmux configuration since it’s a customize board,
see-also developer guide, MB1 Configuration Changes.

BTW,
GPIO24 maps to GPIO3_PM.03, you may see-also default pinmux configuration file.
for instance, $OUT/Linux_for_Tegra/bootloader/generic/BCT/tegra234-mb1-bct-pinmux-p3701-0000-a04.dtsi

                        dp_aux_ch3_hpd_pm3 {
                                nvidia,pins = "dp_aux_ch3_hpd_pm3";
                                nvidia,function = "eth2";
                                nvidia,pull = <TEGRA_PIN_PULL_NONE>;
                                nvidia,tristate = <TEGRA_PIN_DISABLE>;
                                nvidia,enable-input = <TEGRA_PIN_ENABLE>;
                                nvidia,io-high-voltage = <TEGRA_PIN_DISABLE>;
                                nvidia,lpdr = <TEGRA_PIN_DISABLE>;

as you can see..
that’s by default configured as SFIO with.. nvidia,function = "eth2".
it is not a GPIO output configured to drive high, and 3.3 V tolerance is already disabled by default.

Our pinmux configuration file looks like this:

dp_aux_ch3_hpd_pm3 {

            nvidia,pins = "dp_aux_ch3_hpd_pm3";

            nvidia,function = "rsvd1";

            nvidia,pull = <TEGRA_PIN_PULL_NONE>;

            nvidia,tristate = <TEGRA_PIN_ENABLE>;

            nvidia,enable-input = <TEGRA_PIN_ENABLE>;

            nvidia,io-high-voltage = <TEGRA_PIN_DISABLE>;

            nvidia,lpdr = <TEGRA_PIN_DISABLE>;

Does “nvidia,io-high-voltage” need to be set to “<TEGRA_PIN_ENABLE>” if 3.3 V is applied to this pin?

hello matthias.wieland,

you may give it a try to read this pin for details..

$ sudo busybox devmem 0x02440048
$ sudo cat /sys/kernel/debug/gpio | grep 'PM.03'
$ gpioinfo | grep 'PM.03'

yes,
if an external circuit applies 3.3 V directly to GPIO24 (i.e. PM.03), you should enable its 3.3 V tolerance with.. nvidia,io-high-voltage = <TEGRA_PIN_ENABLE>;.

BTW,
you may share the connected circuit or the schematic for double check.

Hello Jerry Chang,
Thank you very much for the information. I have also attached the relevant section of the circuit diagram.

Hi JerryChang

could this circuit diagram be of any help.

Have you found a fault that is related to the problem?

Thanks for your help

hello matthias.wieland,

let me arrange hardware team to review your diagram for relevant section of the circuit diagram.

I would like to double check your issue,
could you clarify how GPIO24 is connected and used on your custom board?
for instance,
– Is it an output intended to drive high, or an input receiving an external 3.3 V signal?
– Where did you measure the voltage, and does it change from 3.3 V to 0 V during boot, or remain at 0 V since the issue first occurred?
– What circuit is connected to the pin, including any pull-up resistor or external voltage source?
– How do you configure and control/read the GPIO? Please share the commands or code and the corresponding GPIO configuration DTSI.

Hello JerryChang

Thanks four your feedback and your efforts.

Here are my answers regarding the questions above:

– Is it an output intended to drive high, or an input receiving an external 3.3 V signal?

In the current configuration, it is an input receiving the external 3.3V signal.

– Where did you measure the voltage, and does it change from 3.3 V to 0 V during boot, or remain at 0 V since the issue first occurred?

It is 0 V from the beginning.

– What circuit is connected to the pin, including any pull-up resistor or external voltage source?

Shown in the previous post, this is the whole circuit on our board.

STORAGE_RESET# is connected to the Reset Pin of two M.2 Storages

– How do you configure and control/read the GPIO? Please share the commands or code and the corresponding GPIO configuration DTSI.

We do not actively read the GPIO at the moment. It is a feature which might be used later.

Hello Jerry Chang,

I would like to kindly ask whether you have received any feedback from your hardware team regarding the circuit diagram yet.

Thanks in advance.

On Jetson AGX Orin, GPIO pads are 1.8V by default. A pad tolerates 3.3V only when “3.3V Tolerance Enable” (E_IO_HV) is set in the pinmux, and only some pins support this option.

With a 3.3V pull-up and this option disabled, the pad is continuously over-voltage, which may damage it over time. We suspect this caused the GPIO24 failure.

It is a MUST to enable 3.3V tolerance in the pinmux when the pin is externally pulled up to 3.3V.