F-Tile JESD204B Intel® FPGA IP User Guide

ID 723907
Date 4/23/2024
Public
Document Table of Contents

4. F-Tile JESD204B IP Functional Description

The F-Tile JESD204B IP implements a transmitter (TX) and receiver (RX) block. Each block has two layers and consists of the following components:

  • Soft IP (SIP)
    • The soft PCS consists of the 8b/10b encoder/decoder and word aligner.
    • Media access control (MAC)—a combination of Data link layer (DLL) block that contains the link layer (link state machine and character replacement), CSR, Control block for Subclass 1 and 2 deterministic latency, scrambler, descrambler and multiframe counter.
    • Custom cadence EFIFO for datapath crossing between native clock data rate divide 40 and system PLL clock.
    • Soft IP reset sequence.
  • Hard IP block that contains the PMA block for serializer and deserializer.

You can specify the datapath and wrapper for your design and generate them separately.

The TX and RX blocks in the DLL utilizes the Avalon® streaming interface to transmit or receive data and the Avalon® memory-mapped interface to access the CSRs. The TX and RX blocks operate on 32-bit data width per channel, where the frame assembly packs the data into four octets per channel. Multiple TX and RX blocks can share the clock and reset if the link rates are the same.

Figure 7. Overview of the F-Tile JESD204B IP Block DiagramThe 8B/10B and word aligner blocks are soft logic.
Figure 8.  F-Tile JESD204B IP TX and RX Datapath Block DiagramThe F-Tile JESD204B IP uses the Avalon® streaming source and sink interfaces, with unidirectional flow of data, to transmit and receive data on the FPGA fabric interface.

32-Bit Architecture

The F-Tile JESD204B IP consist of 32-bit internal datapath per lane. This means that F-Tile JESD204B IP expects the data samples to be assembled into 32-bit data (4 octets) per lane in the transport layer before sending the data to the Avalon® streaming data bus. The F-Tile JESD204B IP operates in the link clock domain. The link clock runs at (data rate/40) because it is operating in 32-bit data bus after 8B/10B encoding.

As the internal datapath of the core is 32 bits, the (F × K) value must be in the order of 4 to align the multiframe length on a 32-bit boundary. Apart from this, the deterministic latency counter values such as LMFC counter, RX Buffer Delay (RBD) counter, and Subclass 2 adjustment counter is the link clock count instead of frame clock count.

Avalon® Streaming Interface

The F-Tile JESD204B IP and transport layer in the design example use the Avalon® streaming source and sink interfaces. There is no backpressure mechanism implemented in this core. The F-Tile JESD204B IP expects continuous stream of data samples from the upstream device.

Avalon® Memory-Mapped Interface

The Avalon® memory-mapped slave interface provides access to internal CSRs. The read and write data width is 32 bits (DWORD access). The Avalon® memory-mapped slave is asynchronous to the txlink_clk, txframe_clk, rxlink_clk, and rxframe_clk clock domains. You are recommended to release the reset for the CSR configuration space first. All run-time F-Tile JESD204B configurations like L, F, M, N, N', CS, CF, and HD should be set before releasing the reset for link and frame clock domain.

Each write transfer has a writeWaitTime of 0 cycle while a read transfer has a readWaitTime of 1 cycle and readLatency of 1 cycle.