3689671345?profile=original

NVIDIA's press release states that "Jetson TX1 is the first embedded computer designed to process deep neural networks -- computer software that can learn to recognize objects or interpret information." The 3.4x2inch module includes a Tegra X1 ARM Cortex-A57 processor with 256-core NVIDIA Maxwell graphics, 4GB of LPDDR4 memory, 16GB of eMMC storage, 802.11ac WiFi and Bluetooth, and Gigabit Ethernet support.

AnandTech Article: http://www.anandtech.com/show/9779/nvidia-announces-jetson-tx1-tegra-x1-module-development-kit

The Jetson TX1 Development Kit will be available for preorder starting Nov. 12 for $599 in the United States. The kit includes the Jetson TX1 module, a carrier board (pictured below), and a 5MP camera. The stand-alone module will be available in early 2016 (for $299 in bulk).

3689671210?profile=original

The Jetson TK1 (not TX1) was released in 2014 to encourage the development of products based on the Tegra K1 processor. However, according to AnandTech, developers were using the Jetson TK1 outright as a production board, choosing to focus on peripheral and software development instead of system hardware development. With the new TX1, all of the I/O connectivity is provided on a carrier board, enabling rapid development on the credit-card sized TX1 module. After development is finished, the TX1 module can be directly deployed in products, such as drones. 

NVIDIA used a drone application to promote the Jetson TX1

https://twitter.com/NVIDIATegra/status/664238535096926208

3689671216?profile=original

E-mail me when people leave their comments –

You need to be a member of diydrones to add comments!

Join diydrones

Comments

  • @Patrick I can send you the docs, request is in your inbox.  Thanks!

  • Hello Jurgen,

    Getting back to the RPI Camera, the tradeoff is to offer experimenters with easy to get and affordable cameras to work with MIPI CSI Serial 2 lanes.The RPI camera and othe flavors like the RPI Noir and Waveshare that offers a good range  of features. The board must 100% plug and play compatible with these device and here is a link to the pinout: 

    http://www.petervis.com/Raspberry_PI/Raspberry_Pi_CSI/raspberry-pi-...

    It is a 2 line type with a clock signal generated on the camera. 2 GPIO Lines for control and LED and I2C for setup and control. I agree with you that on a designer point of view that I2C and flat printed ribbon cable are not the best option available but it has the merit of being easily available. 

    Waiting for Nvidia camera datasheet (Dustin wrote that it would be on developper site pretty soon), for the final design of the interface to ensure Hdw & Drivers compatibility.

    Best Regards

  • Alternatively download the ZIP file of the step model.

  • Hi JB,

    Cooling suppliers: in the past I have worked with Fischer Elektronik (for aluminum profiles for heat sinks) and ebmpabst (for high quality fans). But it might be a good idea to look at Chinese suppliers as well.

    Survey: thank you for this proposal - I will look into it. I have never done that.

    3D step model:

    please have a look at this step model of the 38182

    Regards, Jurgen

  • Thanks for the info Jurgen.

    I think 1.8A on the USB3 will be fine as USB3 spec is 0.9A per port.

    Re the stacking height it was merely an observation as I don't have any dimensions of the board. I trust that you have it well under control! BTW if you would like you could export a DXF file from Altium and I could draw up a 3D printable enclosure with cooling etc for the carrier and TX1. I'll have to investigate what cooling options there are as well unless you already have some suppliers.

    SD card just needed a nudge... ;-)

    BTW what do you think about doing a google form survey as previously discussed on this thread before going to prototype/production? It's completely up to you if you think it will help.

    Regards JB

  • Hi Patrick,

    I think we are ok. I have looked at the video. It shows an aluminum plate below the PCB. Nvidia calls this RF Shielding, so it is not a heat spreader. It is 2.8mm thick. The heat spreader is on top of the PCB and it has on top the heat sink with fan. So for the stackup height I just need to account for the 2.8mm RF Shielding.

    But thank you for pointing this out. It is always good to double check these things before they terribly go wrong. ;-)

    The drawing below is from the TX1 data sheet, the photo is from the video. Both seem to match closely.

    Regards, Jurgen

    3702131205?profile=original

    3702131253?profile=original

  • Jurgen
    I might be wrong, but lloking at the jetsonhack site the pictures shows a plate on bottom and thi is the text
    The Jetson TX1 Module has two aluminum plates that sandwich the actual TX1 circuit board. These are called a Thermal Transfer Plate (or TTP) The plates should not be detached from the board itself. This is because the actual process to attach the plates is a one time process. This is not like where heat sinks can be removed and replaced with some thermal paste. Reiterating, DO NOT REMOVE THE PLATES AROUND THE TX1 CIRCUIT BOARD, or else the magic smoke will get out.

    I am on the road now, ill get back for the RPI Cam

    Cheers
  • Hi Patrick,

    the thermal plate of the TX1 module is on top. So I do not have to worry about it. I just leave space for the RF shield on the bottom side of the TX1 module.

    Actually I do not think that it is a good idea to put the I2C bus on a cable. The I2C can easily be distorted and you get data transmission errors. It is much better to use the CAN bus for this purpose. I plan to connect the 2 CAN busses of the TX1 to my 30 pin motherboard connector. So these CAN busses will be connected to the other CAN devices on the modular motherboard and to the external CAN connectors that way. I could replace the 4 pin I2C connector on the right by a 4 pin CAN connector and add two solder bridges so a user could configure the carrier board to bring out the CAN interface on this 4 pin connector (for a single board setup without a motherboard plugged in). Does anybody see a need for the i2c connector?

    RPi camera: I am just concentrating on the hardware. I need to look at how the extra signals like CAM_GPIO and CAM_CLK should be connected from the TX1 to the RPi camera. Any suggestions?

    Regards, Jurgen

  • Hi JB,

    the PicoBlade contacts are rated at 1A. At 12V this is 2x12Vx1A = 24 watts. The TX1 will be 15 watts max. This leaves 9 watts for the rest. The logic on the carrier board is all very low power. So 9 watts can be assigned to USB3. At 5V this equates to 1.8A. Would this be ok?

    Stacking height: the FPC connector is 2.3mm high. The height of the the mini HDMI connectors is 3.5mm. The PicoBlade connectors are also 3.5mm. The FFC cabe is less tan 0.2mm thick. So the tallest component on the top of the carrier board at this time is 3.5mm. The RF shielding on the TX1 is 2.8mm. So the minimum stacking height is 3.5 plus 2.8 = 6.3mm. The SEAF connector on the TX1 is the "-05.0" type. The SEAM type only carrier board can be from "-02.0" to "-11.0" resulting in a stacking height from 7 to 16mm. So even the smallest 7mm combination would be ok.

    I can move the SD card holder to the edge. This is a good idea. It is the push-pull type (Amphenol 101-00660-68-6).

    Regards, Jurgen

  • It really is a good implementation.
    There are still a couple of unknowns:
    - is there enough clearance for the thermal plate?
    Take al look at the pics. http://jetsonhacks.com/2015/11/13/nvidia-jetson-tx1-module-connector/
    - are you ok wtih the routing of the clock- data- i2c lines to the connectors
    - how will we adress the rpi cameras is the last remaing issue, it has to be easily implemented with v4l2

    Keep on the good work !!

    Best regards
This reply was deleted.