New Bridging Capabilities Address Needs of Mobile-Influenced Markets
Advances simplify integration of legacy, mobile components
A couple decades ago when embedded designers wanted to reduce cost, they turned to the high-volume architecture of their day—the PC. By moving to hardware and software originally developed for the PC architecture they could drive down development costs while using highly reliable components already proven in millions of consumer applications. The PC bus offered a highly attractive, lower cost alternative to specialized embedded applications. At the same time, it allowed developers to take advantage of an ecosystem of easy-to-use design tools and open source operating systems on a very familiar hardware architecture. By migrating to the PC architecture, they could pass cost savings onto their customers, while shortening their own development cycle.
Fast forward to the present day. The PC architecture is no longer the high-volume leader it once was. When today’s embedded system developers want to leverage the cost benefits of high-volume consumer markets, they look to the smartphone market. With billions of units of smartphones sold per year, designers in a wide range of fields are looking for ways to take advantage of the high performance and low cost of the key components in today’s mobile solutions. Much like the PC architecture twenty years ago, the mobile market offers a familiar hardware architecture and applications whose performance has been proven in the highly competitive consumer markets.
Figure 1. The MIPI Alliance has endorsed a core set of standards to simplify mobile system design and ensure interoperability. (Courtesy MIPI Alliance)
Core Set of Standards
Driving the design of current-generation portable devices are the interfaces and busses defined by the MIPI Alliance. Founded in 2003, the MIPI Alliance was formed to develop a comprehensive set of specifications for mobile and mobile-influenced devices. Its goal was to provide the hardware and software interface specifications that device vendors needed to create innovative mobile products while accelerating time-to-market and driving down costs. By developing a core set of standards, the MIPI Alliance enabled mobile device manufacturers to source components from different vendors and easily optimize the performance of their designs.
Today, the organization has grown from four to hundreds of member companies and its reputation in the smartphone industry is well established. All major chip vendors use MIPI Alliance specifications and all smartphones use at least one MIPI Alliance specification.
The impact of the MIPI Alliance, however, falls well beyond the mobile industry. As mobile connectivity becomes increasingly common across all aspects of life, a growing number of industries are attempting to leverage mobile technologies in their designs. However, they face a significant obstacle. The cameras and displays used in many embedded systems do not match the type or number of interfaces on today’s mobile Application Processors (APs).
The most common components embedded designers can leverage from the MIPI market are application processors, image sensors, and displays. Many mobile designs today use the MIPI Display Serial Interface (DSI) for their display and the MIPI Camera Serial Interface (CSI-2) for their image sensors. Both the DSI and CSI-2 interfaces are based on the D-PHY physical bus, a center-aligned, source synchronous interface featuring a differential clock and one to four differential data lines. The D-PHY is clocked on both the rising and falling edges. One of the unique characteristics of the D-PHY is its ability to change “on-the-fly” from differential to single-ended signaling.
Figure 2: The video bridge enables designers to integrate legacy displays with MIPI APs. (Courtesy Lattice Semiconductor)
Figure 3: In this example, a video bridge allows the developer to integrate a legacy image sensor with a MIPI AP.
How can designers in the embedded market leverage the numerous advantages of the MIPI-compliant market when, in most cases, their designs feature legacy or proprietary displays and image sensors? Take the industrial market for example; traditionally, designers of embedded applications have relied on displays featuring a LVDS, RGB, or SPI interface. Most embedded processors don’t feature a DSI interface. Industrial designers who want to take advantage of MIPI processors and applications but retain their traditional LVDS display will need a bridge from their LVDS display to a MIPI-compliant AP. However, it’s possible to construct a video bridge that converts from OpenDSI, LVDS, or proprietary interfaces to MIPI DSI (see Figure 2).
Similarly, many designers in the industrial market who want to take advantage of the latest generation of mobile processors and applications would like to retain the use of their CMOS cameras. Figure 3 illustrates how designers can now bridge between a legacy CMOS parallel output and the CSI-2 input of a MIPI AP.
In the automotive industry, the use of MIPI-compliant APs and components is also on the rise. As content and the number of cameras in automotive designs continue to climb for applications like Advanced Driver Assistance Systems (ADAS) and infotainment, the need for video bridging capabilities is increasing as well. Cameras are now being used not only to help the driver see behind the vehicle when backing up, but also to replace side mirrors, provide full 360-degree visibility, and to support applications like tracking lane changes and minimizing blind spots. In today’s automobiles, designers can aggregate video data from multiple image sensors, stitch it together, and deliver it to the AP via a single CSI-2 interface.
The same video bridging capability is proving highly useful in gaming applications when designers want to aggregate data from multiple cameras or divide that data up across multiple displays. For example, one of the emerging trends in the rapidly evolving virtual reality (VR) market is the migration from a single display to dual head mount displays, running at half-bandwidth, to deliver stereographic images. But how does the designer split the video feed if the AP only has a single DSI interface? Figure 4 illustrates how to use a video bridge to distribute video data coming from the AP across a single DSI interface. The block diagram shows that it’s rather straightforward to split a single DSI signal across two DSI interfaces, one for the left eye and a second for the right. The bridge, as shown in Figure 4, can support two HD displays or a single QHD display at I/O rates up to 1.5 Gbps/lane.
Figure 4: By splitting the video data across two DSI interfaces, a bridge allows developers to implement VR head mount displays with stereo vision. (Courtesy Lattice Semiconductor)
Another potential video bridging application is the aggregation of data from several sources into a single CSI-2 bridge. Figure 5 depicts how drone or VR developers are using new video bridging devices to combine image sensor video outputs from multiple sources into one stream to meet the interface requirements of the AP. Therefore, a video bridge can serve in applications where the AP does not offer enough interfaces to support the number of image sensor inputs or where there is a processing latency between the image sensors and imaging data. In this case, the processor must capture multiple CSI-2 outputs at the exact same time with minimal latency. The multiple merged video streams must also share a common clock and, in some cases, require individual power up routines.
Figure 5: As drone and other system manufacturers integrate multiple CSI-2 cameras, they may need to aggregate their video content for delivery to the AP.
Finding the Right Solution
As shown above, to address new video applications, today’s embedded video designers not only need bridging solutions, but solutions that can deliver high performance in a low-power, compact footprint. Ideally, they need a bridging solution that allows them to convert incompatible interfaces between cameras, displays, and APs, combine multiple video streams into a single output, or split a single output across multiple interfaces.
One way to solve this problem is to use a general purpose, multi-channel passive switch to route signals to multiple locations on a circuit board. However, most multiplexor/demultiplexor (mux/demux) solutions don’t offer the high performance or level of design flexibility that designers need. Another option is to invest in a bridging solution based on an Application Specific Standard Product (ASSP) or Application Specific Integrated Circuit (ASIC). However, most bridging applications are far too narrowly defined to justify the high non-recurring engineering (NRE) costs and long development cycles associated with that type of approach.
A third option is to use a product that’s designed to address the aforementioned bridging issues. A little more than a year ago, Lattice Semiconductor introduced its CrossLink family of FPGAs with the high performance, fast time-to-market, and level of design flexibility that other solutions cannot deliver. CrossLink is an integrated bridging solution that offers designers the industry’s fastest MIPI D-PHY bridging device, capable of delivering up to 4K UHD resolution at a 12 Gbps bandwidth. A ready-made solution that’s designed to solve interface mismatches between APs, image sensors, and displays, CrossLink supports a wide range of leading and legacy protocols in a low-cost and highly compact form factor.
Each CrossLink bridge is capable of supporting a wide range of video functions, including multiplexing, merging, de-multiplexing, arbitrating, splitting, and data conversion. Products like the CrossLink can support a wide range of interfaces including MIPI D-PHY, MIPI CSI-2, MIPI DSI, CMOS, RGB, SubLVDS, SLVS, LVDS, and Open LDI. The CrossLink bridge is configured with Lattice’s Diamond design software, which comes at no cost but includes a license.
There are many benefits to leveraging mobile products for other markets, but compatibility with interfacing video can keep even the best designers from moving forward rapidly, video interfacing being one of the more complex challenges. There are several solutions out there to accomplish video bridging. Designers can create a mux/demux circuit from scratch, find and apply just the right ASSP, or commission an ASIC, all of which can limit choices or slow down the development cycle. However, another alternative that effectively addresses rapid time-to-market demands is an FPGA. The CrossLink FPGA video bridging solution is extremely versatile, has a development flow cycle that’s engineered for a rapid design cycle, and allows a high level of design control for exacting results; all in an off-the-shelf device.
Two decades ago the dominant PC architecture drove development in a wide range of markets. Today, the high-flying mobile market is exerting a similar impact on fields as diverse as industrial, automotive and medical. As mobile technology continues to expand into new applications, look for designers of mobile-influenced products to utilize bridging techniques to take full advantage of the mobile market’s economies of scale, lower costs, and performance benefits.
Tom Watzka is the Technical Mobile Solutions Architect at Lattice Semiconductor with over 20 years of experience in developing embedded products, including seven years developing consumer mobile solutions. Watzka also serves as the Marketing Product Manager for the CrossLink video bridge product line, focused on mobile and mobile influenced markets. He received his BS degree from the Rochester Institute of Technology, MS degree from Pennsylvania State University, and conducted his Master’s Thesis on FFT algorithms.
Sign up to get tech news, product updates and store promos from Symmetry Electronics.Subscribe