The use of multi-channel phased array radars goes back before World War II. The basic idea of a phased array system is relatively simple; an electromagnetic wave transmits from multiple antennas with precisely controlled time delays to each element. In doing so, it is possible to create regions where signals constructively add and regions where they subtract—creating the effect of a "beam" in one direction and not another.
On top of this basic idea of signal directivity, engineers have layered on advanced techniques to increase capabilities using more channels for narrow or multiple beams, polarization and pulse modulation. One thing is clear: given more channels and more bandwidth, it is possible to apply advanced digital signal processing to extract more advanced information for many applications from military, aerospace, astronomy, or several sensor applications.
Wider bandwidth translates to faster sampling rates—producing more data to transfer and process. Adding more channels multiplies the data management challenge by creating multiple streams. These combined trends have put considerable strain on digital bus interfaces to move data and digital signal processing fabrics to process and consume that data.
The Data Processing Challenge
The field programmable gate array (FPGA) is a key technology heavily used by multi-channel sensor arrays for its ability to handle constant streams of high speed data. Thanks to innovations from companies like Xilinx and Intel, we have enjoyed a constant stream of evolving chipsets with more and more resources. With devices like the Virtex/Kintex Ultrascale, engineers have been able to produce truly amazing technological capabilities in multi-channel sensor systems. Today, FPGAs are not only a massive fabric of configurable processing resources, but also a complete "system-on-chip" (SoC) in the truest sense of the word. The recent release of the Abaco VP880 is one such example of a product that includes both traditional FPGA and SoC technology that enables engineers to process data in harsh environments where multi-channel phased arrays are often deployed.
However: the data processing is only one piece of the puzzle.
The Data Movement Challenge
When PCI evolved to PCIe this was an acknowledgement that creating wider buses had its limitations, and going to a serial lane architecture was preferred. Today, analog conversion devices are going through a similar transition with the emergence of high speed serial interfaces on the fastest devices. To accommodate this, FPGAs include many multi-Gigabit transceiver connections that are capable of interface with protocols like JESD204B.
The VME International Trade Association (VITA) released a dedicated bus architecture for FPGA mezzanine cards (FMC) under VITA 57.1. FMC buses included 10 high speed serial lanes and 160 pins for traditional bus architectures. FMC has served the COTS industry well and is evolving to FMC+ under VITA 57.4 which now has up to 32 high speed serial lanes to a single mezzanine. This is particularly relevant for the needs of data movement for multi-channel receivers as this opens the door for even more wideband channels in this small modular form factor.
The recent release of the Abaco FMC134 is an example of a COTS product that has both the wide bandwidth—with four channels at 3.2 GSPS—and the high performance JESD204B interface to move data into the highest performance FPGAs. Aggregating multiple data streams into a large FPGA simplifies system design complexity and ultimately results in a lower cost and faster time to deployment. The FMC134 is the first FMC+ module from Abaco: however, we already have multiple carriers that can host this technology including the PC820, PC821, VP868, and the recent VP880.
Our customers demand the highest performance in their most demanding radar applications. To help us create the FMC134 FMC+ Direct RF Conversion Module, we turned to Texas Instruments and their ADC12DJ3200 technology. It uniquely gives us the support we need for a combination of high sample rates above 3 GHz and high 12-bit resolution. Beyond that, it delivers consistent input impedance over frequency which helps ensure better performance across a wide range of possible applications. Lastly, it has the ability to run in two operational modes, giving our customers greater flexibility in implementation. For us, it delivers the optimum contribution to the overall capabilities of the FMC134.
The FMC134 Direct RF Conversion FMC+ is a step up in what’s possible with modular COTS technology and we look forward to seeing how this will enable our customers to meet the challenges of next generation multi-channel wide-band systems.