USRP Hardware Driver and USRP Manual  Version:
UHD and USRP Manual
USRP N3xx Series

Comparative features list

  • Hardware Capabilities:
    • Dual SFP+ Transceivers (can be used with 1 GigE, 10 GigE, and Aurora)
    • External PPS input & output
    • External 10 MHz input & output (20 MHz and 25 MHz inputs also supported)
    • External White Rabbit time/frequency reference input support
    • Internal 25 MHz reference clock
    • Internal GPSDO for timing, location, and 20 MHz reference clock + PPS
    • External GPIO Connector with UHD API control
    • External USB Connection for built-in JTAG debugger and serial console
    • Xilinx Zynq SoC with dual-core ARM Cortex A9 (Speedgrade 2) and Kintex-7 FPGA (XC7Z100 or XC7Z035 depending on variant)
  • Software Capabilities:
  • FPGA Capabilities:
    • Timed commands in FPGA
    • Timed sampling in FPGA
    • RFNoC capability

The N3XX series of USRPs is designed as a platform. The following USRPs are variants of the N3XX series:

N310/N300 4-channel/2-channel ("Magnesium")

N310 Exploded View

The N310 is a 4-channel transmitter/receiver based on the AD9371 transceiver IC. It has two daughterboards with one AD9371 each; every daughterboard provides two RF channels. Note that the product code "N310" refers to the module consisting of mother- and daughterboard, the daughterboard itself is referred to by its codename, "Magnesium".

  • Supported master clock rates: 122.88 MHz, 125 MHz, 153.6 MHz
  • Tuning range: 10 MHz to 6 GHz (below 300 MHz, additional LOs and mixer stages are used to shift the signal into the frequency range of the AD9371)
  • Support for external LOs
  • 4 RX DDC chains in FPGA (2 for N300)
  • 4 TX DUC chain in FPGA (2 for N300)

The N300 is a subset of the N310. It has 2 TX/RX channels (on a single daughterboard; the daughterboard itself is the same as the N310) and a smaller FPGA (XCZ035). Also, it does not have connectors for external LOs.


The Zynq CPU/FPGA and host operating system

The main CPU of the N310 is a Xilinx Zynq SoC XC7Z100 (exception: The N300). It is both a dual-core ARM Cortex A9 CPU and Kintex-7 FPGA on a single die. The CPU is clocked at 800 MHz (speedgrade 2).

The programmable logic (PL, or FPGA) section of the SoC is responsible for handling all sampling data, the 10 GigE network connections, and any other high-speed utility such as custom RFNoC logic. The processing system (PS, or CPU) is running a custom-build OpenEmbedded-based Linux operating system. The OS is responsible for all the device and peripheral management, such as running MPM, configuring the network interfaces, running local UHD sessions, etc.

It is possible to connect to the host OS either via SSH or serial console (see sections SSH connection and Serial connection, respectively).

The STM32 microcontroller

The STM32 microcontroller controls various low-level features of the N3xx series motherboard: It controls the power sequencing, reads out fan speeds and some of the temperature sensors. It is connected to the Zynq via an I2C bus.

It is possible to log into the STM32 using the serial interface (see Connecting to the microcontroller). This will allow certain low-level controls, such as remote power cycling should the CPU have become unresponsive for whatever reason.

The SD card

The N3XX series uses a micro SD card as its main storage. The entire root file system (Linux kernel, libraries) and any user data are stored on this SD card.

The SD card is partitioned into four partitions:

  1. Boot partition (contains the bootloader). This partition usually does not require touching.
  2. A data partition, mounted in /data. This is the only partition that is not erased during file system updates.
  1. Two identical system partitions (root file systems). These contain the operating system and the home directory (anything mounted under / that is not the data or boot partition). The reason there are two of these is to enable remote updates: An update running on one partition can update the other one without any effect to the currently running system. Note that the system partitions are erased during updates and are thus unsuitable for permanently storing information.

Note: It is possible to access the currently inactive root file system by mounting it. After logging into the device using serial console or SSH (see the following two sections), run the following commands:

$ mkdir temp
$ mount /dev/mmcblk0p3 temp
$ ls temp # You are now accessing the idle partition:
bin   data  etc   lib         media  proc  sbin  tmp    usr
boot  dev   home  lost+found  mnt    run   sys   uboot  var

The device node in the mount command will likely differ, depending on which partition is currently already mounted.

Getting started

This will run you through the first steps relevant to getting your USRP N3XX series up and running.

Assembling the N3XX

Unlike the X300 or N200 series, there is no assembly of daughterboards required. Members of the N3XX product family, such as the N310, ship with daughterboards pre-installed.


  • Connect power and network
  • Read security settings
  • Connect clocking (if required)
  • Connect external LOs (if required)

Updating the file system

Before doing any major work with a newly acquired USRP N300/N310, it is recommended to update the file system. Updating the filesystem can be accomplished directly on the N300/N310 by using Mender or externally by manually writing an image onto a micro SD card and inserting it. While manual updating is faster, Mender requires no direct physical access to the device. For details on using Mender, see Section Mender: Remote update capability .

Manual updating is simply loading an image on the micro SD card. The first step in that process is to obtain an image.

To obtain the default micro SD card image for a specific version of UHD, install that version of UHD ( or later) on a host system with Internet access and run:

$ uhd_images_downloader -t n3xx_common_sdimg_default

The image will be downloaded to <UHD_INSTALL_DIR>/share/uhd/images/usrp_n3xx_fs.sdimg, where <UHD_INSTALL_DIR> is the UHD installation directory.

To load an image onto the micro SD card, connect the card to the host and run:

$ sudo dd if=<YOUR_IMAGE> of=/dev/<YOUR_SD_CARD> bs=1M

The <YOUR_IMAGE> is the path to the micro SD card image (i.e.<UHD_INSTALL_DIR>/share/uhd/images/usrp_n3xx_fs.sdimg).

The <YOUR_SD_CARD> device node depends on your operating system and which other devices are plugged in. Typical values are sdb or mmcblk0.
CAUTION: Operating on the wrong device can cause damage to that device.

The micro SD card used can be the original SD card shipped with the device or another one that is at least 16 GB in size.

Insert the updated micro SD card and power on the device.

Serial connection

It is possible to gain root access to the device using a serial terminal emulator. Most Linux, OSX, or other Unix flavours have a tool called 'screen' which can be used for this purpose, by running the following command:

$ sudo screen /dev/ttyUSB2 115200

In this command, we prepend 'sudo' to elevate user privileges (by default, accessing serial ports is not available to regular users), we specify the device node (in this case, /dev/ttyUSB2), and the baud rate (115200).

The exact device node depends on your operating system's driver and other USB devices that might be already connected. Modern Linux systems offer alternatives to simply trying device nodes; instead, the OS might have a directory of symlinks under /dev/serial/by-id:

$ ls /dev/serial/by-id

Note: Exact names depend on the host operating system version and may differ.

Every N3XX series device connected to USB will by default show up as four different devices. The devices labeled "USB_to_UART_Bridge_Controller" are the devices that offer a serial prompt. The first (with the if00 suffix) connects to Linux, whereas the second connects to the STM32 microcontroller. If you have multiple N3XX devices connect, you may have to try out multiple devices. In this case, to use this symlink instead of the raw device node address, modify the command above to:

$ sudo screen /dev/serial/by-id/usb-Silicon_Labs_CP2105_Dual_USB_to_UART_Bridge_Controller_007F6CB5-if00-port0 115200

You should be presented with a shell prompt similar to the following:


On this prompt, you can enter any Linux command available. Using the default configuration, the serial console will also show all kernel log messages (unlike when using SSH, for example), and give access to the boot loader (U-boot prompt). This can be used to debug kernel or bootloader issues more efficiently than when logged in via SSH.

Connecting to the microcontroller

The STM32 microcontroller (which controls the power sequencing, among other things) also has a serial console available. To connect to the microcontroller, use the other UART device. In the example above:

$ sudo screen /dev/serial/by-id/usb-Silicon_Labs_CP2105_Dual_USB_to_UART_Bridge_Controller_007F6CB5-if01-port0 115200

It provides a very simple prompt. The command 'help' will list all available commands. A direct connection to the microcontroller can be used to hard-reset the device without physically accessing it (i.e., emulating a power button press) and other low-level diagnostics.

SSH connection

The USRP N-Series devices have two network connections: The dual SFP ports, and an RJ-45 connector. The latter is by default configured by DHCP; by plugging it into into 1 Gigabit switch on a DHCP-capable network, it will get assigned an IP address and thus be accessible via ssh.

In case your network setup does not include a DHCP server, refer to the section Serial connection. A serial login can be used to assign an IP address manually.

After the device obtained an IP address you can log in from a Linux or OSX machine by typing:

$ ssh root@ni-n3xx-311FE00 # Replace with your actual device name!

Depending on your network setup, using a .local domain may work:

$ ssh root@ni-n3xx-311FE00.local

Of course, you can also connect to the IP address directly if you know it (or set it manually using the serial console).

Note: The device's hostname is derived from its serial number by default (ni-n3xx-$SERIAL). You can change the hostname by modifying the /etc/hostname file and rebooting.

On Microsoft Windows, the connection can be established using a tool such as Putty, by selecting a username of root without password.

Like with the serial console, you should be presented with a prompt like the following:


Network Connectivity

The RJ45 port (eth0) comes up with a default configuration of DHCP, that will request a network address from your DHCP server (if available on your network).

The SFP+ (sfp0, sfp1) ports are configured with static addresses and, respectively. Their default MTU value is 8000. These settings are independent of the image type (HG vs. XG), i.e., the defaults are the same for 1 GigE and 10 GigE (unlike the X310!).

The configuration for the sfpX port is stored in /etc/systemd/network/

For configuration please refer to the systemd-networkd manual pages

The factory settings are as follows:

eth0 (DHCP):




sfp0 (static):




sfp1 (static):




Additional notes on networking:

  • Care needs to be taken when editing these files on the device, since vi / vim sometimes generates undo files (e.g. /etc/systemd/network/, that systemd-networkd might accidentally pick up.
  • Temporarily setting the IP addresses or MTU sizes via ifconfig or other command line tools will only change the value until the next reboot or reload of the FPGA image.
  • If the MTU of the device and host computers differ, streaming issues can occur.

Security-related settings

The N3XX ships without a root password set. It is possible to ssh into the device by simply connecting as root, and thus gaining access to all subsystems. To set a password, run the command

$ passwd

on the device.

Updating the FPGA

Updating the FPGA follows the same procedure as other USRPs. Use the uhd_image_loader command line utility to upload a new FPGA image onto the device.

A common reason to update the FPGA image is in the case of a UHD/FPGA compat number mismatch (for example, if UHD has been updated, and now expects a newer version of the FPGA than is on the device). In this case, simply run

$ uhd_images_downloader

to update the local cache of FPGA images. Then, run

$ uhd_image_loader --args type=n3xx,addr=ni-n3xx-311fe00

to update the FPGA using the default settings. If a custom FPGA image is targeted for uploading, use the --fpga-path command line argument. Run

$ uhd_image_loader --help

to see a full list of command line options. Note that updating the FPGA image will force a reload of the FPGA, which will temporarily take down the SFP network interfaces (and temporary settings, such as applied via ifconfig on the command line, will be lost).

Using an N3XX USRP from UHD

Like any other USRP, all N3XX USRPs are controlled by the UHD software. To integrate a USRP N3XX into your C++ application, you would generate a UHD device in the same way you would for any other USRP:

auto usrp = uhd::usrp::multi_usrp::make("type=n3xx");

For a list of which arguments can be passed into make(), see Section Device arguments.

Device arguments

Key Description Supported Devices Example Value
addr IPv4 address of primary SFP+ port to connect to. All N3xx addr=
second_addr IPv4 address of secondary SFP+ port to connect to. All N3xx second_addr=
mgmt_addr IPv4 address or hostname which to connect the RPC client. Defaults to `addr'.All N3xx mgmt_addr=ni-sulfur-311FE00 (can also go to RJ45)
find_all When using broadcast, find all devices, even if unreachable via CHDR. All N3xx find_all=1
force_reinit Force full reinitialization of all subsystems. Will increase init time. N310 force_reinit=1
master_clock_rate Master Clock Rate in Hz N310 master_clock_rate=125e6
identify Causes front-panel LEDs to blink. The duration is variable. N310 identify=5 (will blink for about 5 seconds)
serialize_init Force serial initialization of daughterboards. All N3xx serialize_init=1
skip_dram Ignore DRAM FIFO block. Connect TX streamers straight into DUC or radio. All N3xx skip_dram=1
skip_ddc Ignore DDC block. Connect Rx streamers straight into radio. All N3xx skip_ddc=1
skip_duc Ignore DUC block. Connect Rx streamers or DRAM straight into radio. All N3xx skip_duc=1
skip_init Skip the initialization process for the device. All N3xx skip_init=1
time_source Specify the time (PPS) source. All N3xx time_source=internal
clock_source Specify the reference clock source. All N3xx clock_source=internal
ref_clk_freq Specify the external reference clock frequency, default is 10 MHz. N310 ref_clk_freq=20e6
init_cals Specify the bitmask for initial calibrations of the RFIC. N310 init_cals=BASIC
init_cals_timeout Timeout for initial calibrations in milliseconds. N310 init_cals_timeout=45000
discovery_port Override default value for MPM discovery port. All N3xx discovery_port=49700
rpc_port Override default value for MPM RPC port. All N3xx rpc_port=49701
tracking_cals Specify the bitmask for tracking calibrations of the RFIC. N310 tracking_cals=ALL
rx_lo_source Initialize the source for the RX LO. N310 rx_lo_source=external
tx_lo_source Initialize the source for the TX LO. N310 tx_lo_source=external
rfic_digital_loopback Digital data loopback inside the RFIC. N310 rfic_digital_loopback=1

Device Initialization

To maximally speed up UHD, an initialization sequence is run when the device (or more accurately, the MPM service) starts. This means even on the first run of UHD, the device will already be initialized into a usable state. Note that it will always come up in a default state, which can be changed by modifying the configuration file in /etc/uhd/mpm.conf (see also Configuration Files), such as this:

; Note: To boot into a fully initialized state, a clock reference must be
; connected before turning the device on if it set to external here:

If you prefer not to have the device initialize on boot, but rather have a fast boot time, add the line skip_boot_init=1 to your /etc/uhd/mpm.conf file.

For more details on the initialization sequence, see the corresponding section for the specific N3XX device:

Subdev Specifications

The four RF ports on the front panel correspond to the following subdev specifications:

Label Subdev Spec
RF0 A:0
RF1 A:1
RF2 B:0 (N310 only)
RF3 B:1 (N310 only)

Note: Before UHD, the subdev spec options were different (A:0, B:0, etc.). Make sure to update your application if you migrated from an earlier UHD version.

The following example will map RF0 onto channel 0 of a uhd::usrp::multi_usrp object, and RF3 onto channel 1:

auto usrp = uhd::usrp::multi_usrp("type=n3xx");
usrp->set_rx_subdev_spec("A:0 B:1");
// This line will now set the gain for RF3 to 20.0:
usrp->set_rx_gain(20.0, 1);
// And this will affect RF0:
usrp->set_rx_gain(20.0, 0);

See also uhd::usrp::subdev_spec_t.

The sensor API

Like other USRPs, the N3x0 series have daughterboard and motherboard sensors. When using uhd::usrp::multi_usrp, the following API calls are relevant to interact with the sensor API:

The following motherboard sensors are always available:

  • ref_locked: This will check that all the daughterboards have locked to the external reference.
  • temperature: The temperature of the die itself
  • gps_lock: GPS lock
  • gps_time: GPS time in seconds sin ce the epch
  • gps_tpv: A TPV report from GPSd serialized as JSON
  • gps_sky: A SKY report from GPSd serialized as JSON

Remote Management

Mender: Remote update capability

Mender is a third-party software that enables remote updating of the root file system without physically accessing the device (see also the Mender website). Mender can be executed locally on the device, or a Mender server can be set up which can be used to remotely update an arbitrary number of USRP devices. Mender servers can be self-hosted, or hosted by Mender (see for pricing and availability).

When updating the file system using Mender, the tool will overwrite the root file system partition that is not currently mounted (note: every SD card comes with two separate root file system partitions, only one is ever used at a single time). Any data stored on that partition will be permanently lost. After updating that partition, it will reboot into the newly updated partition. Only if the update is confirmed by the user, the update will be made permanent. This means that if an update fails, the device will be always able to reboot into the partition from which the update was originally launched (which presumably is in a working state). Another update can be launched now to correct the previous, failed update, until it works. See also Section The SD card.

To initiate an update from the device itself, download a Mender artifact containing the update itself. These are files with a .mender suffix.

Then run mender on the command line:

$ mender -rootfs /path/to/latest.mender

The artifact can also be stored on a remote server:

$ mender -rootfs

This procedure will take a while. After mender has logged a successful update, reboot the device:

$ reboot

If the reboot worked, and the device seems functional, commit the changes so the boot loader knows to permanently boot into this partition:

$ mender -commit

To identify the currently installed Mender artifact from the command line, the following file can be queried:

$ cat /etc/mender/artifact_info

If you are running a hosted server, the updates can be initiated from a web dashboard. From there, you can start the updates without having to log into the device, and can update groups of USRPs with a few clicks in a web GUI. The dashboard can also be used to inspect the state of USRPs. This is simple way to update groups of rack-mounted USRPs with custom file systems.

Salt: Remote configuration management and execution

Salt (also known as SaltStack, see Salt Website) is a Python-based tool for maintaining fleets of remote devices. It can be used to manage USRP N3XX series remotely for all types of settings that are not controlled by UHD. For example, if an operator would like to reset the root password on multiple devices, or install custom software, this tool might be a suitable choice.

Salt is a third-party project with its own documentation, which should be consulted for configuring it. However, the Salt minion is installed by default on every N3XX device. To start it, simply log on to the device and run:

$ systemctl start salt-minion

To permanently enable it at every boot, run (this won't by itself launch the salt-minion):

$ systemctl enable salt-minion

To make use of Salt, both the device needs to be configured (the "minion") and, typically, a server to act as the Salt master. Refer to the Salt documentation on how to configure the minion and the master. A typical sequence to get started will look like this:

  1. Install the salt-master package on the server (e.g. by running apt install salt-master if the server is an Ubuntu system), and make sure the Salt master is running.
  2. Add the network address / hostname of that server to the /etc/salt/minion file on the device by editing the master: line.
  3. Launch the Salt minion on the USRP by running the command systemctl start salt-minion.
  4. The minion will try to connect to the master. You need to authorize the minion by running salt-key -a $hostname where $hostname is the name of the minion.
  5. Once the device is authorized, you can try various commands to see if the communication was established:

    $ [sudo] salt '*' ni-n3xx-311FE00: True $ [sudo] salt '*' network.interfaces



    hwaddr: 02:00:03:11:fe:00 inet:


    address: broadcast: label: eth0 netmask: up: True


Clock/Time Synchronization

Internal references

The N3xx series has an onboard GPSDO as well as a 25 MHz reference oscillator, which can both be used as time- and clock references. The GPSDO will function as a reference even when there is no GPS reception.

Note that this does not enable the synchronization of multiple devices. Using an internal reference is the default.

External references

In order to synchronize multiple USRPs, an external reference, such as the CDA-2990 (OctoClock), is required. If only a clock reference is available, it is possible to derive an internal PPS signal from the reference (which will allow devices to share a frequency, but not a time reference). If both an external clock and time source are provided, devices will be synchronized in frequency and time.


White Rabbit

White Rabbit is an Ethernet-based synchronization procedure; it is an extension of the IEEE 1588 Precision Time Protocol (PTP). The N3xx device can be configured as a White Rabbit slave.

To use White Rabbit, it is necessary to provide an appropriate reference via Ethernet. This reference must be connected to SFP0. Finally, a White Rabbit-compatible FPGA must be loaded. SFP0 will not be available for data transport in this mode.

The White Rabbit image is provided as a default image. To obtain the default images, simply run:

$ uhd_images_downloader -t n3xx -t fpga

Then, you can install the WX (or WA) image using uhd_image_loader:

$ uhd_image_loader \
    --args type=n3xx,addr=ni-n3xx-<DEVICE_SERIAL>,WX

Once the image is loaded, select internal as the clock source and sfp0 as the time source (note: this will fail if the WX or WA image is not currently loaded):

// Or if you want to change it to White Rabbit after initialization:
// Using the older time/clock source APIs is also possible:
// The 2nd call can technically be skipped because the device implementations
// will coerce, but for consistency with other code and for being explicit this
// is the preferred way. The 2nd call will immediately return in this case.

For more information, refer to the White Rabbit Homepage, or the Ettus Research Knowledge Base.


Errors while streaming

If you are getting sequence or other errors while streaming, make sure the MTU settings of the network devices match up. UHD will try and do an automatic MTU discovery, but there are cases when the automatic MTU discovery will yield incorrect values. Often, the host computer MTU is set smaller than the device MTU, but the MTU discovery will detect a larger MTU than the host computer MTU in this error case.

The default MTU for the N3x0 series is 8000. The simplest solution is often to set the host computer MTU to 8000 as well:

$ [sudo] ifconfig eth0 mtu 8000 # Replace eth0 with the device you're using

Of course, you can also reduce the MTU on the device to match your host computer, see Section Network Connectivity.

Theory of Operation

The N3xx-series are devices based on the MPM architecture (see also: The Module Peripheral Manager (MPM) Architecture). Inside the Linux operating system running on the ARM cores, there is hardware daemon which needs to be active in order for the device to function as a USRP (it is enabled to run by default).

A large portion of hardware-specific setup is handled by the daemon.

Building custom filesystems and SD card images

Ettus Research provides SD card images at regular intervals, but there can be good reasons to build custom SD cards, e.g., to test the very latest UHD or MPM for which there has not been an SD card release, to add own applications to the SD card, or to run a modified version of UHD.

Note that building SD cards is very disk space and RAM intensive.

Using Docker to build filesystems

Ettus Research provides a Docker containers to facilitate building filesystems. Using Docker hub, the container can be downloaded by running

$ docker pull ettusresearch/oe-build

Then, navigate to a location with enough disk space:


Create a world-writable directory called 'build':

$ mkdir build && chmod 777 build

Then run the Docker container:

$ docker run -i -t -v $(pwd)/build:/home/oe-builder/build:rw,z ettusresearch/oe-build /bin/bash

Note the order of the naming above might vary by docker version, sometimes it might need to be:

$ docker run -i -t ettusresearch/oe-build -v $(pwd)/build:/home/oe-builder/build:rw,z /bin/bash

After running the previous command, you will be inside the container. First, configure your build environment:

$ TEMPLATECONF=pwd/meta-ettus/conf/sulfur source ./oe-core/oe-init-build-env ./build ./bitbake

Then, you can invoke bitbake to build the image:

$ bitbake $image_name

If you keep the build directory, bitbake will reuse it on consecutive runs, which will speed up builds significantly between runs.

This step will build the SDK, the SD card image, and the Mender artefact.

Modifying and compiling UHD and MPM for the N3XX

N3xx devices ship with all relevant software installed on the SD card. Updating UHD and/or MPM on the SD card is typically easiest done by updating the filesystem image (see Section Mender: Remote update capability). However, it is certainly possible to compile UHD and MPM by hand, e.g., in order to modify and try out changes without having to build entire filesystems in between. At Ettus R&D, this mode of operation is often used for rapid iteration cycles.

Compiling MPM natively

In general, compiling natively is not a recommended way of compiling code for the ARM processors. However, in the case of MPM, the amount of C++ code that needs to be compiled is very little, and a full compile of MPM will take a few minutes even on the N3xx. First, you need to get a copy of the MPM source code onto your device. If you have an internet connection, you can use git to pull it directly from the Ettus repository (all commands are run on the device itself, inside the home directory):

$ git clone

You can also SSHFS it from another computer:

$ mkdir uhd # Create a new, empty directory called uhd
$ sshfs user@yourcomputer:src/uhd uhd # This will mount ~/src/uhd from the remote machine to ~/uhd on the N3xx

Now, create a build directory and use the regular cmake/make procedure to kick off a build. It can be advantageous (especially for slow network connections) to create the build directory outside of the repository directory:

$ mkdir build_mpm
$ cd build_mpm # You are now in /home/root/build_mpm
$ cmake ../uhd/mpm
$ make -j2 install # This will take several minutes

Note that this overwrites your system MPM. You can install MPM to another location by specifying -DCMAKE_INSTALL_PREFIX, but make sure to update all of your paths appropriately.

If you prefer cross-compiling MPM the same way as UHD, refer to the following sections and adapt the instructions for UHD appropriately.

Obtaining an SDK

The recommended way to develop software for the N3xx is to cross-compile. By running the compiles on a desktop or laptop computer, you will be able to speed up compile times considerably (compiling UHD natively for the N3xx would take many hours).

SDKs are distributed along with other binaries. They contain a cross-compiler, a cross-linker, a cross-debugger, and all the libraries available on the device to mirror its environment.

The SDK is shipped in the same way as the other binaries, and you can download the correct version using uhd_images_downloader

$ uhd_images_downloader -t sdk -t n3xx

To unpack and install the SDK, simply execute it after downloading it:

$ cd /usr/local/share/uhd/images # Change this to where your images are stored
$ ./

This will prompt you for an installation path. Please ensure you have sufficient disk space, as each of the SDKs may require several gigabytes of disk space (depending on the image flavor selected).

This will allow you to compile UHD as well as (depending on the image flavor) other software, such as GNU Radio.

Please note, that while several toolchains can be installed in parallel, they have to be installed to different directories.

SDK Usage

Having installed the toolchain in the last step, in order to build software for your device open a new shell and type:

$ . $SDKPATH/environment-setup-armv7ahf-vfp-neon-oe-linux-gnueabi

This will modify the PATH, CC, CXX etc, environment variables and allow you to compile software for your USRP N3xx device. To verify all went well you can try:

$ $CC -dumpmachine

which should return 'arm-oe-linux-gnueabi'.

Building UHD

  1. Obtain the UHD source code via git or tarball
  2. Set up your environment as described in SDK Usage
  3. Type the following in the build directory (assuming a build in host/build):
     $ cmake -DCMAKE_TOOLCHAIN_FILE=../host/cmake/Toolchains/oe-sdk_cross.cmake -DCMAKE_INSTALL_PREFIX=/usr .. # Add any CMake options you desire
     $ make # You can run make -j12 to compile on 12 processes at once

Note: The UHD you are cross-compiling will not run on your host computer (the one where you're doing the development). Compiling UHD regularly on your host computer (with MPMD enabled) will allow you to talk to your N3xx.

Building GNU Radio

  1. Obtain the GNU Radio source code via git or tarball
  2. Set up your environment as described in SDK Usage
  3. Use the following commands to create a build directory, configure and compile gnuradio. You only need create the build directory once.
$ mkdir build-arm
$ cd build-arm
$ cmake -Wno-dev -DCMAKE_TOOLCHAIN_FILE=../cmake/Toolchains/oe-sdk_cross.cmake \-DCMAKE_INSTALL_PREFIX=/usr -DENABLE_GR_VOCODER=OFF -DENABLE_GR_ATSC=OFF \
-DENABLE_GR_DTV=OFF -DENABLE_DOXYGEN=OFF ../ # Append any CMake options you desire

Several GNU Radio components depend on running binaries built for the build machine during compile. These binaries can be built and used for cross compiling, but this is an advanced topic.

N310-specific Features

Front and Rear Panel

Like the USRP X300 series, the N310 has connectors on both the front and back panel. The back panel holds the power connector, all network connections, USB connections for serial console (see Serial connection), JTAG, peripherals, SMA connectors for GPS antenna input, 10 MHz clock reference, PPS time reference input and output (TRIG in/out), the slot for the SD card (see also The SD card), and indicator LEDs.

The following indicator LEDs are used:

  • LINK: This LED will be lit when this USRP has been claimed by a UHD session.
  • REF: Indicates a lock to the reference clock. In particular, when using an external reference clock, this LED is useful to see if the LMK04828 PLLs are locking to this reference clock. Note that some software interaction is necessary to enable the LMK04828, and thus this LED may be off even if a valid reference clock signal is connected.
  • GPS: Indicates a GPS lock (i.e., GPS satellites are in view of the GPS antenna and signal levels are sufficient)
  • PPS: This LED will blink once every second to indicate a valid PPS signal.
N310 Rear Panel

The front panel is used for all RF connections (including the external LO inputs, see External LOs) and all TX/RX connections, as well as the front-panel GPIO.

The connectors labeled RF0 and RF1 are also referred to as slot A, and the connectors labeled RF2 and RF3 are also referred as slot B (matching the internal connections to the daughterboard. Every slot is powered by a single AD9371 RFIC).

N310 Front Panel

Device Initialization (Fast and Slow)

When a UHD session is created, an initialization sequence is started. As part of the initialization sequence, the following steps are performed:

  • All clocking is initialized
  • The JESD links are trained and brought up (between the FPGA and the AD9371)
  • The AD9371 is reset, its firmware is uploaded, and calibrations are initialized (See also RF Calibrations)
  • N310 only: The multi-chip synchronization is performed to align all the RFICs to the common time and clock reference

This sequence can take a while, depending on the master clock rate and the calibration sequence. To speed things up, the device will retain a state between sessions, but only if no relevant settings were touched. In particular, changing the master clock rate, the clock source, or the calibration masks will force a full re-initialization which is very slow compared to the fast re-initialization. By setting the log level to DEBUG you will be able to observe the exact settings that cause fast vs. slow re-initialization. If you require a full re-initialization every time a UHD session is spawned, specify the force_reinit flag as a device arg. Specifying it will always do the full, slow initialization, but will guarantee a full reset of the RFIC.

To maximally speed up UHD, an initialization sequence is run when the device (or more accurately, the MPM service) starts. This means even on the first run of UHD, the device will already be initialized into a usable state. Note that it will always come up in a default state, which can be changed by modifying the configuration file in /etc/uhd/mpm.conf (see also Configuration Files), such as this:

; Note: To boot into a fully initialized state, a clock reference must be
; connected before turning the device on if it set to external here:

If you prefer not to have the device initialize on boot, but rather have a fast boot time, add the line skip_boot_init=1 to your /etc/uhd/mpm.conf file.

RF Calibrations

The onboard RFIC (AD9371) has built-in calibrations which can be enabled from UHD. A more detailed description of the calibrations can be found in the AD9371 user guide, see chapter "Quadrature Error Correction, Calibration, and ARM configuration".

Not all calibrations available on the AD9371 are applicable to the USRP N310. However, those calibrations that are applicable can be enabled/disabled at initialization time using the tracking_cals and init_cals device args (see also Device arguments). These device can be set to the precise bit mask the chip uses to set those calibrations (e.g., init_cals=0x4DFF,tracking_cals=0xC3) or they can use the following descriptive keys provided by UHD (e.g.init_cals=DEFAULT,tracking_cals=TX_QEC|RX_QEC). The | symbol can be used to combine keys (equivalent to a bitwise OR).

Calibrations can significantly delay the initialization of a session. By only picking relevant calibrations, sessions can be initialized faster.

Key (init_cal) Function
TX_BB_FILTER Tx baseband filter calibration
ADC_TUNER ADC tuner calibration
TIA_3DB_CORNER Rx TIA filter calibration
DC_OFFSET Rx DC offset calibration
TX_ATTENUATION_DELAY Tx attenuation delay
RX_GAIN_DELAY Rx gain delay calibration
FLASH_CAL ADC flash calibration
PATH_DELAY Path delay calibration
TX_LO_LEAKAGE_INTERNAL Tx LO leakage internal initial calibration
TX_LO_LEAKAGE_EXTERNAL Tx LO leakage external initial calibration (requires external LO)
TX_QEC_INIT Tx QEC initial
LOOPBACK_RX_LO_DELAY Loopback ORx LO delay (ORx not connected by default!)
LOOPBACK_RX_RX_QEC_INIT Loopback Rx QEC initial calibration
RX_QEC_INIT Rx QEC initial calibration
BASIC Preset for minimal calibrations (TX_BB_FILTER, ADC_TUNER, TIA_3DB_CORNER, DC_OFFSET and FLASH_CAL)
OFF Preset for disabling all initial calibrations
ALL Enable all applicable calibrations
Key (tracking_cal) Function
TRACK_RX1_QEC Rx1 QEC tracking
TRACK_RX2_QEC Rx2 QEC tracking
TRACK_ORX1_QEC ORx1 QEC tracking
TRACK_ORX2_QEC ORx1 QEC tracking
TRACK_TX1_LOL Tx1 LO leakage tracking
TRACK_TX2_LOL Tx2 LO leakage tracking
TRACK_TX1_QEC Tx1 QEC tracking
TRACK_TX2_QEC Tx2 QEC tracking
OFF Disable all tracking
RX_QEC Enable all RX QEC tracking
TX_QEC Enable all TX QEC tracking
TX_LOL Enable all TX LO leakage tracking
DEFAULT Enable all QEC tracking
ALL Enable all tracking (except ORx)

External LOs

The N310 has inputs for external local oscillators. For every daughterboard, there is one input for TX and RX, respectively, resulting in 4 LO inputs total per N310.

Reasons to use an external LO include:

  • Improving phase alignment: The N310 itself has no way of aligning phase between channels, and phase will be random between runs. By applying an external LO, the phase ambiguity is reduced to 180 degrees, produced by a by-2 divider in the AD9371 transceiver IC.
  • Improving phase noise: The quality of the onboard LO depends on the external reference clock, among other things. By providing a custom LO signal, it is possible to more accurately tune, assuming the externally generated LO signal is coming from a high-quality oscillator.

Storing user data in the EEPROM

The N310 daughterboard has an EEPROM which is primarily used for storing the serial number, product ID, and other product-specific information. However, it can also be used to store user data, such as calibration information.

Note that EEPROMs have a limited number of write cycles, and storing user data should happen only when necessary. Writes should be kept at a minimum.

Storing data on the EEPROM is done by loading a uhd::eeprom_map_t object into the property tree. On writing this property, the driver code will serialize the map into a binary representation that can be stored on the EEPROM.

FPGA Register Map

The following tables describe how FPGA registers are mapped into the PS. This is for reference only, most users will not even have to know about this table.

AXI Slave Address Range UIO Label Description
Slave 0 4000_0000 - 4000_3fff - Ethernet DMA SFP0
Slave 1 4000_4000 - 4000_4fff misc-enet-regs0 Ethernet registers SFP0
Slave 2 4000_8000 - 4000_bfff - Ethernet DMA SFP1
Slave 3 4000_c000 - 4000_cfff misc-enet-regs1 Ethernet registers SFP1
Slave 4 4001_0000 - 4001_3fff mboard-regs Motherboard control
Slave 5 4001_4000 - 4001_41ff dboard-regs0 Daughterboard control, slot A
Slave 6 4001_8000 - 4001_bfff dboard-regs1 Daughterboard control, slot B
N310 Register Map
AXI Slave Module Address Name Read/Write Description
Slave 0 axi_eth_dma0 4000_0000 - 4000_4fff Ethernet DMA RW See Linux Driver
Slave 1 n3xx_mgt_io_core 4000_4000 PORT_INFO RO SFP port information
[31:24] COMPAT_NUM RO -
[23:18] 6'h0 RO -
[17] activity RO -
[16] link_up RO -
[15:8] mgt_protocol RO 0 - None, 1 - 1G, 2 - XG, 3 - Aurora
[7:0] PORTNUM RO -
n3xx_mgt_io_core 4000_4004 MAC_CTRL_STATUS RW Control 10gE and Aurora mac
[0] ctrl_tx_enable (PROTOCOL = "10GbE")RW-
[0] bist_checker_en (PROTOCOL = "Aurora")RW-
[1] bist_gen_en RW -
[2] bist_loopback_enRW -
[8:3] bist_gen_rate RW -
[9] phy_areset RW -
[10] mac_clear RW -
n3xx_mgt_io_core 4000_4008 PHY_CTRL_STATUS RW Phy reset control
n3xx_mgt_io_core 4000_400C MAC_LED_CTL RW Used by ethtool to indicate port
[1] identify_enable RW -
[0] identify_value RW -
mdio_master 4000_4010 MDIO_DATA RW -
4000_4014 MDIO_ADDR RW -
4000_4018 MDIO_OP RW -
n3xx_mgt_io_core 4000_4020 AURORA_OVERUNS RO -
eth_switch 4000_5000 MAC_LSB RW Device MAC LSB
4000_5004 MAC_MSB RW Device MAC MSB
4000_6000 IP RW Device IP
4000_6004 PORT1, PORT0 RW Device UDP port
eth_dispatch 4000_6008 [1] ndest, [0] bcastRW Enable Crossover
4000_600c [1] my_icmp_type, [0] my_icmp_code
eth_switch 4000_6010 BRIDGE_MAC_LSB Bridge SFP ports in ARM
4000_6014 BRIDGE_MAC_MSB -
4000_6018 BRIDGE_IP -
4000_6020 BRIDGE_EN -
chdr_eth_framer 4000_6108 onwards LOCAL_DST_IP W Destination IP, MAC, UDP for Outgoing Packet for 256 SIDs
4000_6208 onwards LOCAL_DST_UDP_MAC_MSBW Destination MAC for outgoing packets (MSB)
4000_6308 onwards LOCAL_DST_MAC_LSBW Destination MAC for outgoing packets (LSB)
4000_7000 onwards REMOTE_DST_IP W Destination IP, MAC, UDP for Outgoing Packet for 16 local addrs
4000_7400 onwards REMOTE_DST_UDP_MAC_HIW Destination MAC (MSB)
4000_7800 onwards REMOTE_DST_MAC_LOW

Destination MAC (LSB)

Slave 2 axi_eth_dma1 4000_8000 -

Same as Slave 0, different base address

Slave 3 n3xx_mgt_io_core 4000_c001 - 4000_cfff - - Same as Slave 1, different base address
eth_dispatch 4000_d000 - 4000_dfff - - Same as Slave 1, different base address
eth_switch 4000_e000 - 4000_efff - -

Same as Slave 1, different base address

Slave 4 n310_core 4001_0000 COMPAT_NUM R FPGA Compat Number
[31:16] Major RO -
[15:0] Minor RO -
4001_0004 DATESTAMP RO -
4001_0008 GIT_HASH RO -
4001_000C SCRATCH RO -
4001_0010 NUM_CE RO Number of Computation Engines (RFNoC Blocks)
4001_0014 NUM_IO_CE RO Number of fixed IO CEs - Radios + DMA Fifo
4001_0018 CLOCK_CTRL
[0] pps select (internal 10 MHz)RWOne-hot encoded pps_select to use the external PPS input.
[1] pps select (internal 25 MHz)RWOne-hot encoded pps_select to use the internally generated PPS with a 10 MHz ref_clk.
[2] pps select (external)RW One-hot encoded pps_select to use the internally generated PPS with a 25 MHz ref_clk.
[3] pps select (GPSDO)RW One-hot encoded pps_select to use the PPS from the GPSDO input to the FPGA.
[4] pps output enableRW
[8] ref clk mmcm resetWO -
[9] ref clk mmcm lockedRO -
[12] meas clk mmcm resetWO -
[13] meas clk mmcm lockedRO -
[11:0] FPGA temperatureRO
4001_0020 BUS_CLK_RATE RO -
4001_0024 BUS_CLK_COUNT RO -
axi_crossbar 4001_1010 XBAR_VERSION RO See crossbar kernel driver
4001_1014 XBAR_NUM_PORTS RO See crossbar kernel driver
4001_1018 LOCAL_ADDR RW See crossbar kernel driver
4001_1020 remote_offset WO XBAR settings reg
4001_1420 local_offset WO XBAR settings reg
n3xx_mgt_io_core (NPIO0) 4001_0200 PORT_INFO RO
n3xx_mgt_io_core (NPIO1) 4001_0240 PORT_INFO RO
n3xx_mgt_io_core (QSFP0) 4001_0280 PORT_INFORO
n3xx_mgt_io_core (QSFP1) 4001_02c0 PORT_INFORO
n3xx_mgt_io_core (QSFP2) 4001_0300 PORT_INFORO
n3xx_mgt_io_core (QSFP3) 4001_0340 PORT_INFORO

Slave 5 4001_40004001_41FFClockingsee Clocking regmap
4001_42004001_43FFSyncsee Sync regmap
4001_46004001_47FFDaughterboard see Daughterboard regmap (EISCAT)
4001_60004001_6FFFJESD Core 0see JESD regmap (EISCAT)
4001_70004001_7FFFJESD Core 1see JESD regmap (EISCAT)
Slave 6 4001_8000 - 4001_bfff see above -same as Slave 5