×

MATLAB环境下的高速转换器工具箱

消耗积分:2 | 格式:pdf | 大小:177.44KB | 2021-03-22

王超

分享资料个

This version (19 Feb 2021 22:53) was approved by Travis Collins.The Previously approved version (09 Jan 2021 00:57) is available.Diff

Table of Contents

ADI maintains a set of tools to model, interface, and target with ADI high-speed converter devices within MATLAB and Simulink. These are combined into single Toolbox which contains a set of Board Support Packages (BSP). The list of supported boards is provided below.

Quick Start with Toolbox

The current stable Toolbox can be downloaded from the High Speed Converter Toolbox GitHub Release Page. Download the latest mltbx file then open that file within MATLAB. Opening the file will automatically install the Toolbox, adding the necessary components to your MATLAB path. The “Analog Devices, Inc. High Speed Converter Toolbox” will appear in your Add-Ons Explore within MATLAB.

To interface and stream data with hardware will require installation of libiio and one of two Hardware Support Packages from MathWorks. The libiio library can be obtained on the Github page of the project.

Libiio Installers

Installation of either:

is required to use the streaming system objects or blocks. These support packages provide the necessary libIIO MATLAB bindings used by ADI's system objects.

Building the Toolbox Manually

The toolbox can only be built under Linux or with Cygwin on a Windows platform. Conveniently, the entire process is automated with a Makefile located in the CI/scripts folder of the repository. The following is required on the system before the build process can be run:

  • A supported MATLAB version installed in the default location (/usr/local/MATLAB)
  • A supported Vivado version installed in the default location (/opt/Xilinx)
  • Packages: git zip unzip tar make wget sed
You should only manually build the toolbox if you require a custom branch or no toolbox installer is available

First clone the repo and move into it: 

git clone https://github.com/analogdevicesinc/HighSpeedConverterToolbox.git
cd HighSpeedConverterToolbox

To build the toolbox run the following: 

make -C CI/scripts build

To create a installable tlbx file run: 

make -C CI/scripts gen_tlbx

Device Control and Data Streaming

Device interfaces which provide control and data streaming are implemented with MATLAB System Objects and Simulink Blocks. These System Objects can be access under the “adi” namespace in MATLAB and are followed by their part number or board name and finally Tx or Rx: 

adi..

For example, to instantiate an AD9144 object to control the high speed converter it can be created as follows: 

tx = adi.AD9144.Tx;

All supported boards are derived from low level objects based on their parts. For example, the DAQ2 Evaluation board actually contains an AD9680 and AD9144. Therefore, it simply uses AD9680 and AD9144 objects under the hood. However, to interact with the more familiar DAQ2 interface naming the Rx side can be instantiated like above as: 

rx = adi.DAQ2.Rx;

For example usage of certain objects, it can be useful to inspect their related test code which exercises initiations in different configurations. The available code is available in the GitHub repo folder here, where object tests have the naming convention Tests.m.

To get a list of currently available objects with the BSP installed simply run: 

help adi

To get more information on a given object run: 

help adi..

or 

doc adi..

Common Attributes

There are some common attributes that need to be set for system objects and parts.

  • uri Context address of IIO device. Possible options include:
    • IP with usage rx.uri = 'ip:192.168.2.1'
    • USB with usage rx.uri = 'usb:1.2.3'

Extending Interfaces

If a driver attribute or setting is not available in the standard objects it can be easily extended to cover more IIO attributes. See this guide.

HDL Targeting with HDL-Coder

The Analog Devices Toolbox for MathWorks HDL Workflow Advisor is a collection of board definitions and reference designs that provide to the MathWorks HDL Workflow Advisor support to:

  • Generate IP blocks compatible with Analog Devices HDL reference designs for various Analog Devices platforms
  • Automatically insert the generated IPs into the Analog Devices Vivado HDL reference designs

This Analog Devices Toolbox is based on the MathWorks Board and Reference Design Registration System.

Functionality

The MathWorks HDL Workflow Advisor enables users to automatically generate HDL code from a Simulink model. The user can choose from a selection of several different Target Workflows, including “ASIC/FPGA”, “FPGA-In-The-Loop”, and “IP Core Generation”. Target Platform selections include Xilinx Evaluation Boards and Altera Evaluation Boards as well as other custom evaluation boards.

The Analog Devices BSP for HDL Workflow Advisor extends the set of Target Workflows for IP Core Generation with the Analog Devices boards listed in the Supported Platforms section. The BSP consists of a set of board definitions that specify all the characteristics needed by the HDL Workflow Advisor to be able to incorporate a board in the code generation flow, as well as a set of Xilinx Vivado reference designs that are used by the Workflow Advisor to automatically insert the generated IPs into the Vivado designs. All the Analog Devices Vivado HDL reference designs have inside a ‘donut hole’ to accommodate custom IPs. Each design exposes a set of interface signals to which the IP can connect to. All these signals are specified in the board definition and are available in the Workflow Advisor GUI to connect to the generated IP’s ports.

When running the Workflow Advisor the first step if to select the Target Platform. The figure below shows some of the available Analog Devices target platforms.

Analog Devices Target Platforms

The next step is to configure the interfaces between the IP and the reference design. Each target platform has a set of interface signals that are accessible in the Target Platform Interfaces drop down boxes form step 1.2 (Set Target Interface) of the HDL Workflow Advisor. The figure below shows an example of how to configure the target interface for a specific model.

Target Interface Signals

All the Analog Devices platforms have similar interface signals and they are dependent on the type of flow that is selected – receive (Rx) or transmit (Tx). The table below describes the interface signals for the AD9361 based SDR platforms.

Receive flow (Rx)
Signal name Width Description
IP Data 0 OUT 16 Custom IP data output signal. This signal is connected to a DMA channel in the ADI reference design.
IP Data 1 OUT 16 Custom IP data output signal. This signal is connected to a DMA channel in the ADI reference design.
IP Data 2 OUT 16 Custom IP data output signal. This signal is connected to a DMA channel in the ADI reference design.
IP Data 3 OUT 16 Custom IP data output signal. This signal is connected to a DMA channel in the ADI reference design.
IP Data Valid OUT 1 Data valid signal from the custom IP. Used to signal to the rest of the design that the IP data out channels have valid data. The duration must be 1 clock cycle.
AD9361 ADC Data I0 16 AD9361 ADC I0 channel data.
AD9361 ADC Data Q0 16 AD9361 ADC Q0 channel data.
AD9361 ADC Data I1 16 AD9361 ADC I1 channel data.
AD9361 ADC Data Q1 16 AD9361 ADC Q1 channel data.
Transmit flow (Tx)
Interface signal name Width Description
IP Data 0 IN 16 Custom IP data input signal. This signal is connected to a DMA channel in the ADI reference design.
IP Data 1 IN 16 Custom IP data input signal. This signal is connected to a DMA channel in the ADI reference design.
IP Data 2 IN 16 Custom IP data input signal. This signal is connected to a DMA channel in the ADI reference design.
IP Data 3 IN 16 Custom IP data input signal. This signal is connected to a DMA channel in the ADI reference design.
IP Load Tx Data OUT 1 Custom IP output signal used to notify the design that the IP is ready to receive new input data. The duration must be 1 clock cycle.
AD9361 DAC Data I0 16 AD9361 DAC I0 channel data. To be used as input into the custom IP.
AD9361 DAC Data Q0 16 AD9361 DAC I0 channel data. To be used as input into the custom IP.
AD9361 DAC Data I1 16 AD9361 DAC I0 channel data. To be used as input into the custom IP.
AD9361 DAC Data Q1 16 AD9361 DAC I0 channel data. To be used as input into the custom IP.

The custom IP always runs at the sample clock and must be able to process / generate a sample every clock cycle.

Once the target interface has been defined, make sure to select the “Target Language” as Verilog (defaults to VHDL) in Step 3.1.1 of the HDL Workflow Advisor. All the other settings of steps 2 and 3 of the HDL Workflow Advisor can be left in their default state and the project generation process can be started by running step 4.1 (Create Project). The result of this step is a Vivado project which has the custom IP core integrated into the Analog Devices HDL reference design. The bitstream for the design can be generated either by running step 4.4 (Create bitstream) or by compiling the generated Vivado Project directly in Vivado. The project can be found in the hdl_prj/vivado_ip_prj folder.

Further Reading

Four Quick Steps to Production: Using Model-Based Design for Software-Defined Radio - Part 4

Creating a BOOT.BIN from HDL Workflow Advisor

Unlike the support package provided by MathWorks, the update process for the bitstreams requires the creation of a BOOT.BIN file which will be compatible with the ADI SD card.

First, make sure you have a valid SD card for your platform with the necessary device tree and kernel image selected. Follow this process here.

Once your SD card is ready, in step 4.3 “Build FPGA Bitstream” of HDL Workflow Advisor select a custom Tcl file for synthesis build. Utilize this adi_build.tcl file for Linux or this adi_build_win.tcl file for Windows located within the BSP as your custom tcl file.

Once the bitstream is built it will generate the BOOT.BIN necessary for booting your system. Place the generated BOOT.BIN in the root of your ADI SD card's BOOT partition.

Simulation Models of Hardware

The HSX toolbox has behavioral simulation models of the following parts.

Part Type Initial Supported Release
AD9081 Behavioral R2019b

These models are available under the adi.sim namespace in MATLAB and can be instantiated as follows: 

model = adi.sim..

Examples of using these models are located in the hsx_examples folder which ships with the toolbox, which is also available on GitHub. The tests folder will also contain usage examples that are primarily used for testing functionality but can be useful to learn different configurations.

Supported Boards

The following have device-specific implementations in MATLAB and Simulink. If a device has an IIO driver, MATLAB support is possible, but a device-specific MATLAB or Simulink interface may not exist yet.

Evaluation Card FPGA Board Streaming Support Targeting Variants and Minimum Supported Release
DAQ2 (AD9680/AD9144) ZC706 Yes No ADI (2019a)
ZCU102 Yes Yes ADI (2019a)
Arria10 SoC Yes No ADI (2019a)
AD9081/AD9082 ZCU102 Yes No ADI (2020a)
VCU118 Yes No ADI (2020a)
AD9988/AD9986 ZCU102 Yes No ADI (2020a)
VCU118 Yes No ADI (2020a)
AD9209/AD9209/AD9177 ZCU102 Yes No ADI (2020a)
VCU118 Yes No ADI (2020a)
QuadMxFE (AD9081 x4) VCU118 Yes No ADI (2020a)
AD9467 Zedboard Yes No ADI (2018b)

Examples

Examples for streaming data and targeting FPGAs are listed within the Toolbox documentation itself. To view run the following with MATLAB: 

doc adi

Help & Support

下载并关注上传者 低至0.43元/天 开通VIP 免费下载

下载资料需要登录，并消耗一定积分。

声明：本文内容及配图由入驻作者撰写或者入驻合作网站授权转载。文章观点仅代表作者本人，不代表电子发烧友网立场。文章及其配图仅供工程师学习之用，如有内容侵权或者其他违规问题，请联系本站处理。 举报投诉

评论(0)
发评论
  • 相关下载
  • 相关文章

下载排行榜

全部0条评论

快来发表一下你的评论吧 !

0