Troubleshooting connection issues with AMD SoC devices in MATLAB/Simulink

This guide provides troubleshooting steps for connection issues with AMD SoC (previously Xilinx Zynq, Xilinx Zynq Ultrascale+) devices when using MATLAB/Simulink workflows. It applies to supported AMD SoC and RFSoC evaluation boards, such as: Zedboard, ZC702, ZC706, ZCU102, ZCU208, ZCU111, ZCU216, ADI RF SOM, and PicoZed.

You may see error messages such as the following when you try to create a "zynq", "xilinxsoc" or "socHardwareBoard" object at the MATLAB command line, in step 4.4 of HDL Workflow Advisor ("Program Target Device"), in SoC Builder "Test Connection", or when you try to follow any of the Hardware/Software Co-Design examples when pressing "Build, Deploy & Start":

Error connecting to SSH server at 192.168.1.100
------
Cannot establish a connection to the board with device address "192.168.1.100". Wait 30 seconds and try again.
------
Could not connect to Zynq hardware board using IP address 192.168.1.101.
Check that your device is connected to a LAN or direct network connection accessible from the MATLAB development computer.
------
Cannot connect to Xilinx SoC device at IP Address "192.168.1.100".
Click here to view the troubleshooting guide and verify the IP Address settings of your hardware.
------
Cannot connect to "Zynq hardware". Details: Could not connect to the hardware over Serial Port.
Check that USB cable is plugged into 'USB-to-UART' port of the hardware and the power switch
is turned on. If this error persists, switch off the hardware, wait for 10 seconds, switch on
the hardware and then repeat this step.

Answer 1

[MathWorks Support Team]

1) Run the Hardware Setup app 

Before you can use any MATLAB/Simulink features with AMD SoC boards, you must install the correct Support Package for your target workflow and fully complete the Hardware Setup app.

Here is a list of current AMD SoC Devices support packages with links to the setup instructions and troubleshooting pages:

• SoC Blockset Support Package for AMD FPGA and SoC Devices*: Setup | Troubleshooting
• HDL Coder Support Package for AMD FPGA and SoC Devices**: Setup | Troubleshooting
• Embedded Coder Support Package for AMD SoC Devices***: Setup | Troubleshooting
• Deep Learning HDL Toolbox Support Package for AMD FPGA and SoC Devices****: Setup | Troubleshooting

Legacy Xilinx support packages available in R2023b and prior:

• HDL Coder Support Package for Xilinx RFSoC Devices: Setup | Troubleshooting
• Communications Toolbox Support Package for Xilinx Zynq-Based Radio: Setup | Troubleshooting
• Vision HDL Toolbox Support Package for Xilinx Zynq-Based Hardware*****: Setup | Troubleshooting

To launch the Hardware Setup app, go to the Home tab in MATLAB, select Add-Ons > Manage Add-Ons, find the support package for your target workflow, then click the Setup button as shown below:

The Hardware Setup app will perform the following tasks:

• It helps you make the correct network settings on the development computer and allows you to set/check the IP address of the board.
• It copies the correct Mathworks Linux OS image to the SD card, which you then insert into your board to boot from it. Make sure to turn off the board before inserting the SD card, then turn it back on.
• It helps you apply the correct jumper settings on the board.
• It performs a basic connection test.

2) Verify the Boot Switches & Jumpers

Ensure the boot switches and jumpers on your board are configured to boot from the SD card. If your board does not boot the MathWorks Linux image from SD card, the support package features will not work. The correct boot switches and jumper positions should be indicated during the Support Package Setup, but you can also consult your board manual or the Setup links above.

3) Check the Hardware

Try using a different (faster/newer) SD card. Make sure the Ethernet cable is in good condition, and try swapping any USB-to-Ethernet adaptors if used, as they are prone to failure. For Radio and Vision workflows, ensure the RF FMC or HDMI I/O cards are firmly mounted to the board.

4) Check the Status LEDs on your board

Turn the power switch of your board to ON. Your board should have one or several Status LEDs that indicate whether the boot was successful. The Status LED should first flicker green, then turn into steady green to indicate a boot success, or into steady red for a boot failure. See the manual to confirm how a boot success is indicated for your board.

5) Inspect the boot log with PuTTY

If the LEDs in 4) indicate a boot failure, plug a Micro-USB cable into the USB-UART port on the board. Note that your board may have multiple Micro-USB ports that serve other purposes, such as USB-JTAG, so make sure you don't accidentally select the wrong one. For more information, refer to your board manual.

Using the COM Port number shown for the device in Windows Device Manager, open a serial command-line session with terminal software, such as PuTTY. Detailed instructions can be found here: Command Line Session with AMD SoC Devices. If the device isn't shown in Windows, make sure your USB device drivers for the USB-UART Bridge are installed correctly.

You should be able to observe the Linux boot log in the PuTTY terminal window when you power-cycle the board. If the boot succeeds, enter ifconfig at the Linux command prompt to obtain the correct IP address of the board. If the boot fails, please contact MathWorks technical support with the output from the boot log.

Important: Close the PuTTY serial connection before trying to reconnect to the board via MATLAB/Simulink.

6) Verify the Network Settings on the Development Computer

With the board IP address that you obtained from PuTTY (for example, 192.168.1.101), open the Network Settings on your Development Computer and specify a compatible IP address and subnet mask for the development computer network connection. See the following page for detailed instructions: Configure Network Interface Card (NIC) on Development Computer

Finally, ping the board from the development computer command line:

ping 192.168.1.101

7) Configure the firewall to allow communication with the board

If the ping succeeds, but the connection from MATLAB/Simulink still fails, contact your local IT team to check the firewall settings on the development computer to not intercept the connection between MATLAB and the AMD SoC board. They need to make sure that:

• The firewall does not block the IP addresses and ports that the development computer and AMD SoC board use to communicate.
• The  <matlabroot>\bin\matlab.exe process is allowed through the firewall. IMPORTANT NOTE: A rule must be set up for every installed version of MATLAB that you intend to use with AMD SoC hardware.

See also: How do I configure Windows Defender Firewall for MATLAB to communicate with external hardware?

Footnotes regarding Xilinx SoC-related Support Package Name Changes:

* Formerly "SoC Blockset Support Package for Xilinx Devices". Since R2024a, it includes the functionality of "Communications Toolbox Support Package for Xilinx Zynq-Based Radio" and "Vision HDL Toolbox Support Package for Xilinx Zynq-Based Hardware".

** Formerly "HDL Coder Support Package for Xilinx FPGA and SoC Devices" and "HDL Coder Support Package for Xilinx Zynq Platform".*** Formerly "Embedded Coder Support Package for Xilinx Zynq Platform".**** Formerly "Deep Learning HDL Toolbox Support Package for Xilinx FPGA and SoC Devices".

***** Formerly "Computer Vision Toolbox Support Package for Xilinx Zynq-Based Hardware".