Microwave RF transceiver system development (L- through Ka-Band) is rapidly changing as new hardware becomes available. Significant advances in analog-to-digital converters (ADCs), digital-to-analog converters (DACs), system on chip (SoC) and system in package (SiP) technologies, requiring varying degrees of corresponding software, create a daunting set of variables for today’s engineers.2 In addition, expanding market applications requiring simultaneous processing of multiple microwave RF bands to provide multiple services, such as ground penetrating radar, navigation and satcom, while adhering to SWaP-C constraints, requires next-generation planning and partnering. Newly developed L- through Ka-Band ADCs, DACs and mixed signal systems in packages (MiXiP SiPs), including SoC field-programmable gate arrays (FPGAs), solve these problems.1 They are completely software controlled, enabling simultaneous sampling for multi-band and multi-service operations.
RF transceivers have traditionally employed a heterodyne architecture (see Figure 1). A transmitted (Tx) or received (Rx) analog RF signal is derived from a DAC or digitized by an ADC through a mixer (up-converted for Tx or down-converted for Rx) with a local oscillator (LO). The derived Tx and Rx signals are “In-Direct” to the RF and are called the intermediate frequency (IF) signals.
IF signals (Tx or Rx) are the sum or difference of the RF and LO signals (RF = IF + LO or IF = RF - LO).
The heterodyne architecture is generally designed for a single frequency band and single service. The advantages of the heterodyne architecture are low-cost, mature implementations, available narrowband components and hardware switchable reconfigurations. The disadvantages are that it is hardware constrained, single band (narrow), single service and limited to sequential sampling/operations.
Some microwave RF transceivers use a homodyne direct RF architecture (see Figure 1).5 Direct RF eliminates the need for mixing (or up/down frequency conversion) and processes the Tx and Rx signals directly in the appropriate RF band required for the service operation. Effectively, direct RF looks like a wire between the antenna from the DAC/Tx or to the ADC/Rx.
The evolution from heterodyne to homodyne transceiver development enabled by advancements in both hardware and software, is shown in Figure 2. Over this time, the greater the processing bandwidth of the hardware, along with software development, has incrementally reduced transceiver analog hardware to the point that the transceiver simply becomes a software-defined radio.
Teledyne e2v now offers a direct RF simultaneous sampling multi-band/service transceiver for conversion up to Ka-Band.1 This single chain transceiver can convert L- through Ka-Band simultaneously (not sequentially) for multi-service operations. This direct RF conversion approach requires that the Tx/Rx module be as close to the antenna as possible, along with all the required circuitry for digital frequency conversion, beam steering, modulation and demodulation functions.
Optimal noise and frequency performance is realized without up-/down-conversion and insertion losses are reduced with no combiner/splitters required. For frequency planning, once the hardware implementation is set (e.g., data converters and filters), the only design variable in the system (other than software reconfigurations) is the sample clock frequency. A completely software-defined system allows for continuous dynamic software reconfigurations, ultimately using AI technology.
TRANSCEIVER APPLICATIONS AND IMPLEMENTATIONS
Microwave transceivers include both Tx and Rx functions within the same module using either heterodyne or homodyne architectures. Transceivers either operate in half-duplex or full-duplex modes. Half-duplex means that Tx and Rx functions must alternate in time, while full-duplex allows the system to transmit and receive data simultaneously.
Figure 3 shows various satellite applications operating in different RF bands.3 Traditionally, transceivers were designed as separate systems per application per RF band. This resulted in one investment per single band per service (for Figure 3, this could represent 19 individual investments and developments/deployments). To maximize functionality (multi-band and multi-service) and minimize development costs, switchable transceiver channels can sequentially switch analog and digital channel functionality with delay times in-between.
Today’s direct RF transceiver developments are efficiently designed for multi-band performance per service (see Figure 4). Unfortunately, multi-band operation may be limited, such as L- and C- or X- and Ku-Bands, depending on the ADCs and DACs that are available. Whether multi-service operation can be fully realized generally depends on digital modulator/demodulator processing and computation speeds. Digital data routing can also be a significant impediment.
Enabling waveform management of multiple services simultaneously as shown in Figure 4 is accomplished with this new direct RF simultaneous sampling multi-band/service transceiver. It simultaneously (not sequentially) processes multiple bands (L- through Ka-Band) as well as multiple services. This is accomplished using Teledyne e2v’s Ka-Band capable ADCs and DACs in conjunction with AMD Xilinx’s advanced 7 nm FPGA XQRVC1902 digital engine packaged together using MiXiP SiP technology.
A multi-band/service microwave transceiver, besides requiring ultra-high performance components, requires cutting edge packaging and interconnect technology to partition the system in ways that minimize analog and digital interference.2 Figure 5 shows several transceiver partitioning techniques: 1) common single multi-chip module with centralized data around the FPGA that minimizes digital routing, 2) SoC and 3) SiPs interconnected with optical digital fiber that place the ADCs and DACs near the antenna while isolating the digital functions from the analog.
Technique 1 has been used for decades but requires complex routing of analog signals for both the ADCs and DACs. Of course, as operational system frequencies increase, analog signal routing becomes problematic.
Technique 2 uses an SoC, which requires monolithic process technology and device geometries that must accomplish both the microwave analog and digital functions necessary for the transceiver. This is very difficult, expensive and requires extreme development timeframes.
Technique 3 combines the best of both techniques 1 and 2 but requires optical data link drive capabilities and is not offered by Teledyne e2v. It lends itself particularly to the use of separate transmitters and receivers connected to separate antennas.
DIRECT RF SIMULTANEOUS SAMPLING MULTI-BAND/SERVICE TRANSCEIVER ARCHITECTURE
Figure 6 is the block diagram of a software-defined direct RF simultaneous sampling transceiver in a single, contained, transceiver antenna module (TAM) connected directly to the antennas for L- through Ka-Band service/operations. At the core of the TAM is the Tx/Rx MiXiP SiP which houses Teledyne e2v’s EV12DD700 (DAC), two EV10AS940 (ADCs) and AMD Xilinx’s XQRVC1902 7 nm FPGA digital engine. The TAM also includes auxiliary components such as lowpass filters (LPFs), bandpass filters/multi-band (n) pass filters (BPFns), low noise amplifiers (LNAs), high-power amplifiers (HPAs) and circulators.1,2,5
Packaging and partitioning of the TAM depends on the required service frequency bands, power transmission levels, physical dimensions and thermal requirements of each auxiliary component. For example, it may be optimal to have the MiXiP SiP, LNAs, LPFs and BPFns on a single PCB directly connected to the HPAs and circulators, which would in-turn connect to the antennas.