HDMI socket

Hello,
My question is about Jetson TX2 display output. I can read from the specifications that TX2 has a HDMI 2.0 display output. While I can easily see the HDMI socket on the developer kit, I cannot find it on the board. What if I want to connect TX2 directly to a monitor? Do I need an alternate socket or to mount it myself? Sorry for the silly question but I need to understand which one I need to order for my company

The module requires a carrier board. The developer kit comes with the module on one carrier board. There are some other carrier boards, you might look around here (search for “carrier”):
[url]https://devtalk.nvidia.com/default/topic/793798/embedded-systems/some-jetson-web-links/[/url]

Mostly TX1 carrier boards will work with a TX2…but verify first.

The only connector on the module is the big fat SAMTEC mezzanine connector. (And two U-FL connectors for WiFi)
The board must have some kind of carrier to receive power, as well as talk to the rest of the world (USB ports, Ethernet port, camera ports, display ports, and so forth.)
The developer kit comes with a carrier board that has a rich set of connectors, similar to what you’d see on a “real PC.” Other sellers sell carrier boards optimized for other use cases, including very-small, or six-camera-in, or rugged-deployment, and so forth.

Thanks for your answers. Now the situation is definitely more clear although I still have some doubts. I noticed TX2 is fanless, but then I see that, among recommended hardware partners, there is Coolermaster DCV-01672-N2-GP Heatsink/Fan. Do you maybe know if I’ll need a fan anyhow? We work with domotic, these boards will be used with customers and we don’t want fan to annoy with noise (sorry for the off topic question)

There is a fan on module in dev kit. DCV-01672-N2-GP contains silver aluminum heatsink and fan. If your thermal design is good enough, you can remove the fan, but usually it is recommended. BTW, there is a thermal design guide in download center for your reference.

I highly recommend reading the thermal design guidelines; they talk about the various requirements and design parameters.
For example, if your end device has a large aluminum enclosure to which you can bond the module with high-quality thermal compound, you may be able to use the enclosure itself for all the cooling you need.
How much cooling you need also depends on how hard you run the device. If you run it in “max efficiency” mode, you can get away with 8 Watts. If you run it in “max performance” mode, you may need between 12 and 15 Watts.
The specifics of the application matter, too. For example, using all of the GPU and one ARM core, may be different from using 6 ARM cores, all the video hardware encoders/decoders, but none of the GPU.