TX2 HDMI not working on custom carrier

I designed a custom carrier and have the +5V connected directly to the HDMI connector.
Does the driver cycle the ‘5V0_HDMI_EN’ signal? Is not having this connected a problem?

Here is my log…

TX2_VPX_Carrier_dump.txt (182 KB)

It would be better to share your schematic here.

hdmi.pdf (39.7 KB)

No 5V0_HDMI_EN saw in your design, it is necessary, the 499Ω, 1% pull-downs must be disabled when Tegra is off to meet the HDMI VOFF requirement. The enable to the FET, enables the pull-downs when the HDMI interface is to be used., please check Note 3 of HDMI part in OEM DG.

Well that is very inconvenient, I guess that mod gets put on the rev-2 design… meanwhile is there a way I can modify the device tree or driver get around this so I can finish testing my prototype? I ran ‘nv-enable-hard-coded-kernel-boot-display-mode.sh’ and it will display the NVIDIA splash screen so I know the hardware path works. Occasionally the HDMI port will work when I hot plug it after boot.
I’m just looking to make this board at least a lab asset… don’t really care about meeting the HDMI VOFF requirement on this design right now.

Here is a work around… add “echo 1 > /sys/kernel/debug/tegra_hdmi/hotplug” to /rc.local.
This forces HDMI up just before the GUI runs.


We are testing jetson TX2 with l4t version 28.2.1 on our custom carrier board. We have similar setup as @wadefreeman. In our carrier, we have NOT populated pull down resistors (499R) and 5v is directly fed to hdmi connector. No regulator for “vdd_hdmi_5v0”.

Behavior is, on bootup, we can see console in hdmi, but blank display once booted completely.
on trying echo 0 >/sys/class/graphics/fb0/blank, we can see mouse cursor sometimes.

I could see many posts about this issue, but none of them has a working solution.

Could you help debug this ?

I suppose jetson is able to read edid successfully. Monitor mode is 1920x1080@60Hz.

Did you try a different monitor? I had some issues using chezzy monitor.
You should be able to dump the resolved edid info…

Resolved it guys. Thanks.
We had an extra pullup resistor on UART1_TX pin.
Which according to design guide, should not be there.
Removed it.