Background of Camera Link Protocol
In an era of high data rates and comprehensive data exchange, traditional cables no longer meet the demands. To ensure reliable communication at high data rates, it became necessary to develop a standard pin allocation and cable assembly method to ensure that compatible devices could easily connect. This standardization allows users to reduce development costs by purchasing cables in bulk. Furthermore, the existence of this standard interface should significantly save users time spent on obtaining technical expertise for integrating a set of compatible cameras and image acquisition cards. In this context, the Camera Link standard was born.
In 2000, driven by the Automation Industry Association (AIA) in the United States, several companies specializing in image cards and cameras jointly released the Camera Link standard protocol. This standard protocol is based on the technology development of National Semiconductor's Channel Link and defines matching standard industrial interface devices, standardizing signal lines to simplify signal transmission between cameras and image acquisition cards. It also defines three transmission modes: Base Configuration, Medium Configuration, and Full Configuration, along with corresponding signal pin specifications and data transfer rates, while providing four camera control signal lines simultaneously.
Variants of Camera Link
As a single Camera Link chip only provides 28 bits of data, multiple Camera Link chips are needed to improve the efficiency of data transmission. Depending on the requirements, video transmission modes are divided into three configurations: Base (single Camera Link chip, one cable), Medium (two Camera Link chips, one cable), and Full (two Camera Link chips, two cables).
In Base mode, a Channel Link chip and a Camera Link mechanical interface are needed. The transmitter sends 28 bits of data per pixel clock, including a 4-bit image enable signal and 24 bits of image data. The throughput of Base mode video data is 2.04 Gbit/s (255 MB/s).
Medium mode requires two Channel Link chips and two Camera Link mechanical interfaces. The transmitter sends 40 bits of data per pixel clock, including a 4-bit image enable signal and 36 bits of image data. The throughput of Medium mode video data is 4.08 Gbit/s (510 MB/s).
Full mode requires three Channel Link chips and two Camera Link mechanical interfaces. The transmitter sends 68 bits of data per pixel clock, including a 4-bit image enable signal and 64 bits of image data. The throughput of Full mode video data is 5.44 Gbit/s (680 MB/s).
Deca mode extends the width of the "Full" configuration by utilizing eight unused bits and reallocating eight redundant framing/enabling bits to generate up to an 80-bit data path width across two connectors/cables, further increasing video bandwidth. The throughput of Deca mode video data is 6.8 Gbit/s (850 MB/s).
The industry has reached a consensus on the 80-bit version, and compatible cameras and image acquisition cards enter the market under the term "Camera Link Deca." However, some manufacturers use "Extended Full" to refer to the Deca configuration, while others use "Full Deca."
Technical Features of Camera Link Standard
The Camera Link technology standard is based on National Semiconductor's Channel Link standard and is a multiplexed version of multiple Channel Links. It includes not only camera image data signals and clock signals but also camera control signals and serial communication signals. The Channel Link standard is a parallel LVDS transmission interface standard. LVDS is a low swing differential signal technology with a voltage swing of around 350 mV, characterized by low interference and fast transition rates. The theoretical maximum transmission rate in an unobstructed transmission medium is 1.923 Gbps.
Camera Link is a physical layer LVDS transmission method. The figure below shows the connection diagram of the Camera Link bus transmitter and receiver, which is also the basic mode of the bus. At the bus transmitter, 28-bit parallel data is converted into four pairs of LVDS serial differential data for transmission, and one pair of LVDS serial differential data lines is used to transmit the image data output synchronization clock. At the bus receiver, the serial differential data is converted into 28-bit parallel data, while the synchronization clock is also extracted. This not only reduces the use of transmission lines but also reduces electromagnetic interference during transmission due to the use of serial differential transmission.
Therefore, Camera Link features real-time, high-speed characteristics. The high bandwidth ensures rapid transmission of dense data without delay, while the use of serial differential transmission enhances interference resistance.
Continuous Upgrades and Evolution
As camera frame rates continue to increase, the Camera Link standard cannot easily extend to beyond 850 MB/s. Hence, another new interface is needed to meet the demand for obtaining higher bandwidth at the same or even lower costs. After extensive exploration of the best technological solutions by many leading manufacturers in the industry, Teledyne DALSA's HS-Link technology provided the first "proof of concept," achieving a maximum transfer speed of 300 MB/s with a single lane (or "lane" in Camera Link HS jargon) and up to 6 GB/s with 20 channels.
In the Camera Link HS standard, the cable can be directly connected to the FPGA without additional devices or equipment, resulting in lower costs. At the same time, Camera Link HS can use higher voltages and receiver automatic equalization to compensate for cable losses, thereby achieving long-distance signal transmission.
Compared to Camera Link cables, Camera Link HS can have a longer transmission distance beyond 15m; with less than 15m, a single Camera Link HS cable can achieve the same bandwidth as two Camera Link cables, but with lower deployment costs. With the same 3mm cable diameter, Camera Link HS can achieve a bandwidth of up to 3.4 Gbit/s on a single cable.
Conclusion
As cameras continue to evolve, high resolution and high frame rates remain important development directions. In the future, to meet higher production and research requirements, more cameras will be adapted to the Camera Link protocol. Applications using Camera Link cameras will occupy a stable market share, and the Camera Link protocol will provide a superior experience for your data transmission.
FALenses Technology specializes in providing machine vision core hardware. You can go to the official website of FALenses Technology at https://www.falenses.com/ for more information.